MWJ 2002.03.09 (March 9, 2002)
==============================

   Copyright 2002, GCSF Incorporated.  All rights reserved.


Top of the Week
---------------

* MacCyclopedia_ has the full story on Mac OS X Printing, complete
  in _one_ issue with all three parts combined. This is no puff
  piece, either - to understand why the new architecture works as it
  does, you first must understand why Mac OS 9 printing is so very
  bad. We start there, work though Apple's first and failed attempt
  to revise the printing architecture, and then explain how Mac OS X
  printing works. We cover all the kinds of printing modules, how
  they fit together, what they do, how they're designed, and then
  provide useful tips like how printing is organized on disk and all
  the pitfalls of adding your own PostScript Printer Description
  files. Printing won't be a mystery after _MacCyclopedia_ explains
  it!

* This Week's News is crowded: lots of follow-ups on older issue,
  plus iMac ads, PowerPC directions, and HP's merger travails. We
  dissent from the prevalent assumption that iMacs are delayed due
  to flat-panel shortages because it doesn't make much sense. We
  separate Microsoft antitrust news into two tracks and look at
  articles from James Coates, Neil McIntosh, Stewart Alsop, Charles
  Haddad, and Stephan Somogyi. The Mac Show blinks, Mac OS X Web
  vendors are touting their products, and two technical issues
  (Finder icons and Internet Config) raise our eyebrows - and
  there's new sample code. All that plus nearly five dozen briefs
  for pointers to later weekend reading.

* Product Showcase provides the top 30 of the week, with full
  looks at Flash MX, Quicken Deluxe updaters, OpenSSH 3.1 (build it
  yourself!), Synchronize Pro X 1.0.2, Location X, Freeway 3.5,
  Apple's "White" input devices, DivX 5.0 alpha, Sputnix 0.6b,
  unsatisfying Mutsu applications, OSXMacPerl 0.2, and iPod
  mismanagement tools. Those are interspersed with 17 briefs for new
  stuff that's important but not hard to explain.

* First notice: per our announced vacation policy, there will be
  no MWJ_ 2002.03.23 in two weeks as we'll be recharging our
  batteries during the prior week. We'll be here next week, of
  course, and back after that with MWJ_ 2002.03.30. We'll repeat
  this next week so that if you ask us for a missing issue we can
  smirk. We don't get to smirk enough.


This Week's News
----------------

**Follow-Ups**

* Facing pushback on high pricing for Lasso 5, a significant
  upgrade to its database-to-Web layer that runs only on Mac OS X,
  Blue World Communications has dropped the price [1] for the
  "Developer Edition," a single-user version intended for
  programmers to use for development and testing instead of
  deploying (or with the limited Standard edition of FileMaker Pro)
  to US$349 for a download version and US$449 for a boxed copy. If
  you're thinking that's expensive enough for a _testing_ version,
  you're not alone: when announced last week, Developer Edition cost
  US$699 for a downloadable version and US$799 for a boxed edition.
  If you purchased Developer Edition at the higher price since its
  2002.02.26 release, Blue World will credit you the difference. No
  upgrades from earlier Lasso products are available. (MWJ_
  2001.06.09, 2002.02.28)

  [1] <http://www.blueworld.com/Lasso5/LassoPro/#Pricing>

* As expected, Apple updated its KnowledgeBase article [2] on
  reinstalling Mac OS X to warn not to reinstall any older version
  after installing the new Mac OS X 10.1.3. (MWJ_ 2001.12.01,
  2002.02.23, 2002.03.03)

  [2] <http://docs.info.apple.com/article.html?artnum=106605>

* The first reports have come in (this time to MacNN [3]) that the
  iMac (Flat Panel) mid-range model - 700MHz PowerPC G4 processor,
  40GB hard drive, 256MB of RAM in the factory slot, Apple Pro
  Speakers, and DVD-ROM/CD-RW "Combo" drive - is now shipping, at
  least in limited quantities. Apple originally said this model
  would be available in February. (MWJ_ 2002.01.16)

  [3] <http://www.macnn.com/>

* In Maine, Gov. Angus King announced that he has raised US$2
  million to help fund the iBook laptop program for middle-school
  students that he has championed, plus US$95,000 in seed money.
  That's the good news. The bad news is that he was supposed to
  raise US$15 million. He has until January 2003 to raise the money.
  Last week, King announced he was restoring some cuts to state
  social services proposed in January, and that took pressure off
  the US$25 million the legislature had set aside for the iBook
  purchases. US$1 million of the money comes from MBNA, but the
  _Lewiston_Sun-Journal_ does not mention MBNA's affiliation with
  Apple Computer. MBNA provides the "Apple instant loan" credit line
  available at the online and US retail Apple stores, and is the
  exclusive provider of such credit. (MWJ_ 2002.02.04, 2002.02.23)

* Still want more info on Adobe Photoshop 7? Company
  representatives will demonstrate the software in five of Apple's
  27 US retail stores [4] over the coming two months, with more
  demonstration of Adobe Illustrator 10 happening in Silicon Valley
  stores. Mac Help Radio also spoke [5] with product manager Karen
  Gauthier earlier this week, but the archive doesn't seem to be
  available online. (MWJ_ 2002.02.28, 2002.03.03)

  [4] <http://maccentral.macworld.com/news/0203/04.adobe.php>
  [5] <http://www.mactvnetwork.com/articles/story595.shtml>

* We found a fix to our difficulty with URL Manager Pro 3.0fc4
  registration not sticking - believe it or not, the solution is
  replacing the preferences file (at path "~/Library/Preferences/URL
  Manager Preferences"). After that, entering a serial number works
  both for that session and for future sessions. We've only seen two
  instances of the massive logout crash since resolving the URL
  Manager Pro difficulties, so we seem to be on the right track.
  (MWJ_ 2002.03.04)

**Business News**

* We take it back. When Apple first showed iMac (Flat Panel)
  animations [6] created by Pixar, we said they'd never be real ads
  in the US. We said Apple couldn't air them because the company
  would be "swamped by complaints from new buyers whose iMacs didn't
  dance like the ones on TV did." (MWJ_ 2002.02.18) Clearly, we're
  just way too cynical. On Thursday night, Apple premiered a new
  iMac television ad featuring Pixar-style animation of the little
  beastie, but in a different situation - a gentleman walking along
  a sidewalk sees an iMac in a store window and notices it mimicking
  him. When he moves his head, the iMac swivels his display. When he
  moves, it hops. When he sticks out his tongue, it opens its drive
  tray. It appears to be a much higher-quality rendition of the same
  style of Pixar animation. Apple put the ad online [7] late Friday.

  [6] <http://www.apple.com/hardware/ads/newimac.html>
  [7] <http://www.apple.com/hardware/ads/imac_window.html>

  We also wonder if Apple would be spending good money to advertise
  the machine if it wasn't about to seriously dent the backlog. It
  was announced over two months ago and despite a raft of publicity,
  this is Apple's first television ad for the computer. Why drum up
  demand for a computer you can't ship? Hmm. Thursday also brought
  news that Apple has purchased an ad spot on the upcoming
  2002.03.24 Oscar[tm] telecast [8] for an estimated US$1.2 million -
  perhaps a good venue for showing a shipping computer.

  [8] <http://www.usatoday.com/money/advertising/2002-03-07-oscars.htm>

* IBM says it will invest heavily in moving its PowerPC
  microprocessors to lower-power mobile devices, according to
  EETimes [9]. Since 95 out of 100 PCs use an Intel or AMD
  microprocessor, the PowerPC line never became the mass-production
  hit IBM and Motorola had intended for the line when it was
  conceived in the early 1990s - and since Apple revoked all Mac OS
  cloning licenses, neither company has had a huge self-interest in
  matching R&D investments seen at AMD and Intel. On the other hand,
  it's a sweet chip - it's powerful, it's fast, and it uses much
  less power than other chips capable of powering desktop PCs. IBM
  sees a huge potential market for processors that deliver big
  computing power with little electrical power, like cell phone and
  palmtop combinations or wireless handheld computers. Can you
  imagine a cell phone with as much computer processing power as the
  original Power Macintosh 8100? Quite a picture. IBM already makes
  lower-power PowerPC chips for embedded systems, but any
  significant gains the company makes could filter back into G3, G4,
  and "G5" processors, a potential benefit for Apple's portable
  computers. That's the good news. The bad news is that IBM
  obviously wants to spend more money on the low-power aspect of
  PowerPCs than the desktop aspect, leaving the development burden
  for future desktop systems still mainly on Motorola.

  [9] <http://www.eet.com/semi/news/OEG20020228S0058>

* With Hewlett-Packard shareholders voting on the merger with
  Compaq later this month, more of the "undecided" shareholders and
  consultants are staking out positions. It seemed like trouble for
  HP when former Lewis Platt - the HP veteran and ten-year CEO whose
  retirement brought current CEO Carly Fiorina to the helm -
  announced [10] his opposition to the merger. Platt is on the board
  of the Packard Foundation, the single largest HP shareholder, that
  had previously joined one of the Hewlett foundations in announcing
  it would vote its shares against the merger. Platt has reportedly
  been approached to return to the CEO position if the merger fails,
  as that would surely mean Fiorina's days of leading HP were over.

  [10] <http://www.nytimes.com/2002/03/04/technology/04HEWL.html>

  Platt also presented his views to Institutional Shareholders
  Service, an independent firm whose recommendations on mergers and
  acquisitions carry great influence with institutional investors -
  pension funds and mutual funds and banks that hold onto stock
  long-term. Some such investors almost always follow ISS
  recommendations, while others just use them as a data point (as
  ISS says it intends).

  Tuesday, after the markets closed, ISS recommended [11] the merger
  to its clients, saying in part that the cost savings predicted by
  HP and Compaq management "are achievable," that revenue loss
  projections "are reasonable," and that the record shows "no
  support" for many of dissident board member Walter Hewlett's
  claims of unexplored alternatives.

  [11] <http://zdnet.com.com/2100-11-852652.html>

  Hewlett responded [12] that ISS "clearly has a predisposition to
  support management and makes a general presumption that boards do
  the right thing." Most experts still believe the vote is too close
  to call [13], but on Wednesday, the FTC unanimously passed [14] on
  any chance to restrict the merger. That was the last regulatory
  hurdle; HP shareholders now vote on 2002.03.19. If they approve,
  Compaq shareholders get to have the same fun clearing a higher bar -
  but one step at a time, please.

  [12] <http://biz.yahoo.com/prnews/020305/nytu127_1.html>
  [13] <http://news.com.com/2100-1001-853480.html>
  [14] <http://news.com.com/2100-1001-853709.html>

* Joe Wilcox of CNet News can pull a dark cloud out of any good
  Apple news, so it's no surprise that his latest article (with
  Michael Kanellos) says the iMac (Flat Panel) is in such short
  supply it will keep Apple in trouble for months to come. It's easy
  to dismiss the doom as more Wilcox shadowing, but there seem to be
  some misconceptions here. Wilcox and Kanellos are correct about
  the backlog - it's real and tangible, and the duo reports that
  Ingram Micro is quoting delivery dates of four and a half _months_
  from now for new orders, with 6500 machines pending.

  The reason for the delay, though, seems specious. All the analysts
  immediately jumped on flat-panel displays as the cause of Apple's
  delayed shipments. Sales of LCDs are way up year-over-year, and
  the factories can't crank them out fast enough to fulfill all
  orders. Therefore, the usual suspects say this _must_ be the
  trouble, and is all Apple's fault. "This is a big _faux_pas_ on
  Apple's part, not having this all ironed out before they brought
  the new iMac to market, particularly if they want to wow Windows
  users and bring them over to what they call the 'light side' from
  the Windows dark side," ARS analyst Toni Duboise told the
  reporters.

  Ignoring for the moment ARS's extremely spotty record in tracking
  Macintosh numbers, this just doesn't make sense. Apple had clearly
  been planning the iMac (Flat Panel) computer for six to twelve
  months before introducing it this year, and in 1999 made a major
  equity investment in Samsung precisely because Samsung makes
  flat-panels. Apple sold its Samsung shares on 2001.04.17 (MWJ_
  2001.05.19), but it seems quite odd that Apple wouldn't have
  locked in enough flat-panel orders in 2001 to meet iMac production
  needs.

  Apple's operations chief Tim Cook gave no public indication over
  the past several quarters that Apple would have _any_ difficulty
  getting flat-panel parts in the next several quarters, including
  this one. Nor do any of the analysts mention that Apple's stand-
  alone flat-panel displays are all in stock and shipping in "1-3
  days," compared to "3-5 weeks" for iMacs. Another ARS analyst, Sam
  Bhavnani, positively gloated over Apple's alleged misstep: "Apple
  should have checked their supply beforehand," he told Wilcox and
  Kanellos. But Apple is the company that has kept amazingly
  aggressive inventory levels for the past several years. There's
  _no_ indication the company would have entered into iMac
  production without enough displays lined up.

  The following is our hunch and you should value it accordingly: we
  think it's an engineering problem. A MacInTouch Reader Report [15]
  from mid-February has stayed with us. Search for the name "Jean-
  Daniel" and you'll find a report of an iMac (Flat Panel) power
  cable with loose contacts. "My boss looked at the power cable and
  noticed it wasn't grounding correctly; the circular power contacts
  were loose. He used a small tool to tighten them, plugged it in,
  and the machine turned on instantly, and worked flawlessly from
  that point on."

  [15] <http://www.macintouch.com/flatpanelimac6.html>

  It continues: "Apple tech support, still on the phone, was amazed
  ... There were actually people in the background clapping, and
  [they] said something to the effect that he had just solved a
  known problem that they had no idea how to fix, and implied the
  'dead' iMacs [were] the reason shipments had been delayed, and
  that they were going to get engineering to redesign the power
  cable."

  Apple's public statements about iMac ramps and shipments had no
  implication supplies would be constrained due to _parts_, just due
  to ramp. The implication that the announcement should have been
  delayed is ridiculous - both press and customers were _demanding_
  a new iMac, and if Apple believed it would have one ready to ship
  in three weeks, failing to announce it at Macworld Expo would have
  been criminal incompetence. There have been several reports of
  power problems with early machines, too. As noted in "Follow-Ups,"
  the mid-level iMac is now shipping, so maybe the floodgates are
  about to open. We have no idea if the MacInTouch report is
  anecdotal or right on the money, but an engineering problem seems
  the data far more snugly than a theoretical flat-panel shortage.

**Business News In Brief**

* Elcomsoft, employer of once-arrested programmer Dmitry Sklyarov,
  argues that the Digital Millennium Copyright Act can't apply to a
  Russian company for technology not developed in the US; the US
  Government counters that Elcomsoft sold the software from a US
  address and marketed it to US customers
  <http://news.com.com/2100-1001-851418.html>

* By the end of next week, Be Inc. will no longer be
  <http://maccentral.macworld.com/news/0203/05.be.php>

* BrandChannel.com declares Apple the "2001 brand of the year,"
  one of the "quirky" brands at the top including Volkswagen,
  Target, Krispy Kreme, and Google.com
  <http://www.brandchannel.com/start.asp>

* Freeverse Software signs a deal to market Airburst, the arcade-
  style game by Aaron and Adam Fothergill of Strange Flavour; no
  word on new theme music by Jared, Bringer of Songs
  <http://maccentral.macworld.com/news/0203/01.strange.php>

* Roxio and Audible partner to make Audible's premium protected
  audio content (like current news and audio books) burnable on
  optical media; it's not clear if the technology gets added to
  Audible's software or if Toast and Easy CD Creator get Audible
  burning privileges
  <http://www.roxio.com/en/company/news/archive/prelease030204.html>

* Intel shipped the Pentium 4-M this week, the first mobile
  Pentium 4 chip, starting at 1.7GHz, a mere 15 months after
  shipping the desktop Pentium 4 and 13 months after Apple shipped
  the first PowerBook G4 (the desktop Pentium 4 uses too much power
  to use in portable computers)
  <http://money.cnn.com/2002/03/04/technology/wires/intel_re/>

* First Albany Companies initiated analysis of Apple Computer
  stock with a "neutral" rating, though no information is available
  on the company's Web site
  <http://www.fac.com/fac_corp_web/equity.html>

* Fund manager David Decker of Janus is a "big fan" of Apple,
  saying it's easier to find Apple products in stores
  <http://biz.yahoo.com/ms/020305/13484.html>

* Apple confirms its first retail store opening of 2002 on
  Saturday in Raleigh, NC (with a press release not posted to the
  company's Web site)
  <http://globalarchive.ft.com/globalarchive/article.html?id=020305005030>

* Apple has not yet confirmed a posted 2002.05.11 opening date for
  a store in Atlanta, GA at Lenox Square
  <http://www.macnn.com/news.php?id=12921>

* The _Orlando_Sentinel_ reports Apple will have a retail presence
  in the upscale Mall at Millenia in that city, opening in October
  <http://www.orlandosentinel.com/business/orl-bizmall07030702mar07.story>

* It's a _Wired_ Weekend at four Apple retail stores (Cambridge,
  MA; Tyson's Corner, VA; Glendale, CA; San Jose, CA) as the
  magazine pairs with Apple to promote digital photography - though
  it's not clear what _Wired_ has to do with any of it
 * Echoing strains of Apple Computer coverage from five years ago,
  _Forbes_'s Lisa DiCarlo says Gateway Computer is running out of
  options to turn around its money-losing and market-share-losing
  operations
  <http://www.apple.com/retail/theater/wired/>
  <http://www.forbes.com/home/2002/03/04/0304gateway.html>

* If you're confused about why companies _want_ to sell low-margin
  PCs, Sam Diaz explains it in the _San_Jose_Mercury_News_,
  including Apple in his overview
  <http://www.siliconvalley.com/mld/siliconvalley/2792783.htm>

**US v. Microsoft**

* We'll try to keep all the news sorted into two tracks. Track 1
  are the hearings that began Wednesday on whether or not the Second
  Revised Proposed Final Judgement, agreed upon by Microsoft, the US
  Government, and nine states, is in the public interest. The US
  Justice Department is still under pressure from several quarters
  for the perception of a weak settlement designed to appease
  Microsoft. The department therefore chose a common tactic after
  the shifting of presidential administrations: blaming the
  predecessors.

  Lead attorney Philip Beck told [16] Judge Colleen Kollar-Kotelly
  that the government had not been able to sustain anything more
  serious against Microsoft than monopoly maintenance, and that
  going for a remedy more sweeping than the settlement "would have
  been an uphill battle that would likely have been resolved against
  us." Microsoft agreed, saying that competitor briefs talking about
  unbridled monopoly power were not facts sustained by the trial
  court or the Court of Appeals. Microsoft's John Warden told the
  court, "Without causation, there's nothing to remedy. One doesn't
  get two bites of the apple," a somewhat disturbing reference.

  [16] <http://news.com.com/2100-1001-852709.html>

  Industry trade groups disagree [17] and told Kollar-Kotelly so in
  brief presentations. The main issue the dissenters raised was
  "commingling code." The trial court and appeals court both found
  that Microsoft merging browser code with Windows to prevent anyone
  from removing it was an anticompetitive act, and representatives
  argued that the SRPFJ does nothing to address that issue.
  Microsoft says the settlement proposal fully addresses the issue
  through requiring Windows to allow third-party middleware.

  [17] <http://zdnet.com.com/2100-1104-854059.html>

  Microsoft also faced questions on whether it has fully disclosed
  contacts with the Justice Department (required by the Tunney Act
  for all antitrust settlements). A _New_York_Times_ story reports
  on the strange case of a New York attorney [18], a lead counsel
  during the original case, who is still battling Microsoft even
  though New York is one of the nine settling states. Don't worry -
  there's still room for the case to get much weirder than this.

  [18] <http://www.nytimes.com/aponline/technology/AP-Microsoft-Battling-
  Lawyers.html>

  By the way, the Justice Department has now made _all_ of the
  comments submitted about the settlement available online [19], in
  either text or PDF format (the PDF file is 43MB when decompressed,
  the text file 56MB), or available free on CD-ROM if you so
  request.

  [19] <http://www.usdoj.gov/atr/cases/ms-comment.htm>

* Track 2 is the separate proceedings for the nine states that had
  sued Microsoft but didn't sign onto the proposed settlement. The
  consensus in the press is that these are the "litigating states"
  as opposed to the "settling states," though the phrase "unsettling
  states" must certainly be more descriptive in Redmond. This week,
  Microsoft revealed [20] that Bill Gates himself will testify at
  the hearing to determine what remedy the nine litigating states
  may have against Microsoft.

  [20] <http://seattletimes.nwsource.com/html/businesstechnology/134413794_microsoft02.html>

  The litigating states want to require Microsoft to make versions
  of Windows available without middleware like the browser or media
  player. Microsoft intends to argue [21] that's not technically
  possible, even though in Track 1 the company has already agreed to
  let third-party middleware replace its own versions built into
  Windows. In a deposition [22], CEO Steve Ballmer said that the
  litigating states' request would require Microsoft to withdraw
  Windows from the market because it couldn't meet the demands.

  [21] <http://www.washingtonpost.com/wp-dyn/articles/A33198-2002Mar3.html>
  [22] <http://zdnet.com.com/2100-1104-850976.html>

  As we had warned, the difference is over language. Ballmer
  interprets the word "middleware" in the litigating states'
  proposal as meaning any part of Windows. Microsoft therefore
  argues [23] that requiring it to remove any piece of Windows for
  anyone who asks would mean "8,000" separate versions of Windows,
  and that the company simply couldn't comply. So, on Friday, the
  litigating states revised their proposal [24] to require Microsoft
  to make a single, modular version of Windows from which
  technologies like browsers and media players could be removed or
  added at will. To counter this, Microsoft must argue that it can't
  make Windows modular, and that any expert who says so is just
  wrong. It's a tougher fight.

  [23] <http://money.cnn.com/2002/03/05/technology/wires/microsoft_ap/>
  [24] <http://zdnet.com.com/2100-1104-851205.html>

**Press Watch**

* The consensus grows: reviewers love iMac (Flat Panel) but don't
  like the ports being in the back. Despite our early insistence
  that the machine's open design diminished that complaint, it
  continues to show up in reviews from technology journalists who've
  obtained review units. Neil McIntosh in UK's _The_Guardian_ lists
  [25] it as his only complaint, noting it "is a minor complaint for
  a machine that looks good on your desk, performs admirably and -
  unusually for Apple - represents cracking value for money,
  especially at the top end of the range."

  [25] <http://www.guardian.co.uk/online/story/0,3605,658954,00.html>

  In the _Chicago_Tribune_, James Coates, as he usually does, finds
  something to twist so he can complain about Apple's products. This
  time, he complains [26] that Apple, with its _5%_ market share (or
  less), is a "monopoly" because it makes all the computers that run
  Mac OS X. Surrounding that is effusive iMac praise, including his
  conclusion that Apple "flat-own owns" DVD product, a "new paradigm
  of personal computing." The iMac, he writes, is "a low-cost
  machine with huge innovations that runs as smooth as a luge track
  and looks good enough to frame" that "also delivers as a
  computer." Coates feels compelled to warn you that "almost all of
  the new stuff" starts in the Windows world "well before it ever -
  if ever - moves into the Mac world," but says the iMac and Mac OS
  X is almost enough to close his perceptual gap.

  [26] <http://www.chicagotribune.com/technology/chi-0203030122mar03.column>

* Kudos to Coates, though, for doing something rare - correcting
  an Apple mistake. In a 2002.02.17 column [27], Coates said the
  "biggest negative" about iPhoto was that "when one uses the image
  tweaking tools to crop or change a photo to black and white, there
  is no option - beyond a short-lived undo command - for making a
  backup in case the tweaks go bad and you change your mind." (MWJ_
  2002.02.23). In fact, as we pointed out, iPhoto retains your
  originals forever untouched unless [28] you crop the photo in
  iPhoto before handing it to an external editor.

  [27] <http://www.chicagotribune.com/business/printedition/chi-0202170338feb17.story?coll=chi%2Dprintbusiness%2Dhed>
  [28] <http://docs.info.apple.com/article.html?artnum=61036>

  Now, in another new story [29] praising iDVD 2 (not the same one
  as in the previous item), Coates corrects himself: "A recent
  review of Apple's iPhoto software inaccurately said that once a
  photo is cropped or tweaked in iPhoto there is no going back.
  Users can click on File and Revert to Original to restore cropped
  pictures." Coates likes iDVD 2 very much, especially the
  acceleration from Velocity Engine, which he less-than-optimally
  calls "special circuitry," but it doesn't affect the facts of his
  review.

  [29] <http://www.chicagotribune.com/technology/chi-0203030124mar03.story>

* Is the iMac (Flat Panel) useful in business? Yes and no,
  according to two major business news magazines and their frequent
  Macintosh pundits. In _Fortune_, Stewart Alsop, a writer whose
  history shows him wrong about the Macintosh more often than right
  (MWJ_ 2001.06.16), says [30] "real office workers" might start
  buying Macintosh computers again because of the iMac (Flat Panel).
  He says they're more fun, that the iApplications strategy is
  really good, that lots of companies "still [have] a business in
  Macintosh software," and that Internet standards make platform
  choice irrelevant in most cases.

  [30] <http://www.fortune.com/indexw.jhtml?channel=artcol.jhtml&doc_id=206734>

  Alsop says Apple "once had something in excess of 10% of the
  corporate computer market," a statistic we've never seen before
  and that has no sourcing (According to Jim Carlton, who never
  tried to give Apple much credit, the company never held more than
  16% of the overall computer market and even that was before the
  Macintosh was introduced.) The real obstacle to Alsop's vision,
  though, is betrayed in his own words: he "hides" that he has a
  Macintosh, pretends to get one for his wife when he wants one, and
  patronizes the platform. "Even though I use a Windows-based
  notebook computer for all of my productive and 90% of my
  nonproductive activities, I try to find excuses to use my
  Macintosh," he writes. How about being productive on one, like
  most customers are? Then Alsop wouldn't need "excuses." His
  stories of laughter at the idea of the Macintosh as a corporate
  tool and dismissive nature of it "still" working reveal more than
  his backwards praise.

  On the other hand, Charles Haddad of _BusinessWeek_ thinks Apple
  is trying [31] for the corporate market despite most indications,
  but is doomed to fail. Haddad bases his belief on the highly-
  questionable _Financial_Express_ articles of late February
  asserting that Apple has hired software developers in India to
  port Unix and Windows applications to Mac OS X (MWJ_ 2002.02.28).
  The articles got most of the details about Apple wrong, but Haddad
  missed or ignored that, preferring to believe that Apple has just
  "kept quiet about this plan." He also praises the new version of
  Microsoft Outlook, but doesn't mention it's only for Mac OS 9,
  despite lots of praise for how Mac OS X would help on corporate
  desktops.

  [31] <http://www.businessweek.com/bwdaily/dnflash/mar2002/nf2002036_9682.htm>

  In the end, though, Haddad dismisses Apple's chances because its
  products cost more than PCs: "Macs, I'm afraid, are never going to
  undersell these IBM clones. Nor do the Mac versions of Outlook,
  Word, or Excel offer much that's different from their PC
  equivalents. So why should corporate IT managers switch platforms,
  given their priorities?" Haddad says Apple can hold on "to the few
  corporate customers it has," but that it should ignore the larger
  corporate market. Haddad hasn't been spot-on in his analysis over
  the past few months (especially when straying from technical
  matters), but if we have to pick either Haddad or Alsop to get the
  Mac market right, we'll take Haddad.

* Stephan Somogyi is just as annoyed [32] as we were earlier at
  Network Associates's bumbling management and non-sale of PGP
  Desktop Security products. We suggested that an Aladdin Systems or
  Power On Software might make a good cross-platform vendor for a
  solid utility application (if NAI didn't price it through the
  stratosphere - having paid US$36 million for PGP over five years
  ago, you can bet NAI wants at least that kind of figure to sell
  pieces of it now), but Somogyi thinks Apple would be a better
  candidate. "Acquiring PGP Desktop would allow Apple to make
  standards-based e-mail, file, and disk encryption, as well as VPN
  technology, an integral part of its new OS. Now would be a
  particularly good time for Apple to make such a move, given
  Microsoft's troubles with security of late."

  [32] <http://zdnet.com.com/2100-1107-851515.html>

  Unfortunately, Apple is not interested in solid standards-based
  encryption. Apple's chief scientist, Dr. Richard Crandall,
  invented an encryption algorithm called "Fast Elliptical
  Encryption" when he was working at NeXT. Thanks to Apple's
  purchase of NeXT, Apple now owns the algorithm. It's the one Mac
  OS 9 uses to encrypt files, and Apple's management has made it
  clear that Crandall and other former NeXT management (up to and
  including Steve Jobs) see it as the only algorithm Apple needs,
  especially when others like RSA might have required royalty
  payments. Apple has had the chance over and over again to include
  PGP-style encryption in the operating system or security
  architecture, and exhibiting the classic "not invented at NeXT"
  syndrome, has passed each and every time. Somogyi has a good idea,
  but Apple's own arrogance will prevent it this time just as it has
  many times before.

* How come Australia's IT Network [33] has comments from Steve
  Jobs criticizing the recording industry after accepting a
  technical GRAMMY[tm] award for Apple Computer, but the US press
  didn't (it's a _Wall_Street_Journal_ article, but it didn't get
  much coverage in the states)? Jobs told reporters that "no one" is
  going to use the music industry's online services that restrict
  what you can do with music you've purchased, adding, "If you
  legally acquire music, you need to have the right to manage it on
  all other devices that you own."

  [33] <http://it.mycareer.com.au/news/2002/03/05/FFX2E13V8YC.html>

  Jobs said the iPod's single-computer-synchronization restriction
  means "you can't use it as a piracy shuttle," which isn't _quite_
  true (you can find the hidden MP3 files on an iPod rather easily,
  and software to make them visible abounds), but at least Apple
  doesn't make moving songs around easy. Jobs also said that the
  GRAMMY[tm] award may have been influenced by the "piracy-limiting"
  features of the iPod, a novel interpretation. Also worth reading
  on the subject: Steven Levy's _Newsweek_ article, "The Customer Is
  Always Wrong," [34], where the author says potentially-mandated
  copy-protection [35] shows the recording industry sees you "not as
  potential consumers, but as candidates for prison denim."

  [34] <http://www.msnbc.com/news/718454.asp>
  [35] <http://www.wired.com/news/politics/0,1283,50784,00.html>

**Press Watch In Brief**

* The squeaky wheel gets the grease: after extensively documenting
  a GeForce4 MX problem with his new Power Macintosh G4 (QuickSilver
  2002) machine and Apple's retail and support refusals to fix it,
  MacNET publisher John Manzione wrote about it for MacNET, got
  several important sites to link to it - and got a new computer
  from Apple
  <http://www.macnet2.com/opinion/oped/index.shtml>

* Lisa DiCarlo's _Forbes_ article does a good job of asking key
  questions about Apple's supply lines in times of shortage - do
  resellers get scarce machines or do they go to Apple's own retail
  stores? - until the last paragraph, where she speculates that
  Apple's retail stores were behind the end of the Circuit City deal
  or that Apple might completely drop all non-Apple US retail, both
  completely without foundation
  <http://www.forbes.com/home/2002/03/07/0307apple.html>

* _Consumer_Reports_, a magazine that has rated Macintosh
  computers poorly in the past for not meeting PC standards, says it
  rated a Mac OS laptop in its new look at notebook computers and
  that all eight tested machines rated "Excellent" or "Very Good"
  <http://www.businesswire.com/cgi-bin/f_headline.cgi?bw.030102/220604517>

* The _Chicago_Sun-Times_ finally prints Andy Ihnatko's iMac (Flat
  Panel) review - no surprise that he loves the machine, but who
  else could deliver thoughts like "If Bill Gates figured out that
  Implications Inform Implementations, bad karma wouldn't cling to
  the guy like sweat to an Amazon rainforest field researcher"
  <http://www.suntimes.com/output/worktech/cst-fin-andy051.html>

* Concluding his two-part series comparing iMac (Flat Panel) and
  Gateway 500SE system running Windows XP, Jim Heid of the
  _Los_Angeles_Times_ finds Windows XP improved but frustrating:
  "the Mac is less frustrating, less commercially intrusive and more
  elegant. Quite simply, it's a better computer."
  <http://www.latimes.com/business/la-000016167mar04.story?coll=la%2Dheadlines%2Dbusiness>

* Dennis Sellers uses an iMac (Flat Panel) machine for two weeks
  and says "this could be the best Mac Apple has ever made, at least
  on a 'bang for the buck' basis"
  <http://maccentral.macworld.com/news/0203/06.imac.php>

* _Salon_'s Damien Cover explores the MPEG-4 licensing brouhaha
  and how it's driving companies to consider open-source
  alternatives
  <http://www.salon.com/tech/feature/2002/03/06/mpeg/>

* Following up on an earlier column with good and bad points about
  Mac OS X installer technology (MWJ_ 2001.12.31), John C. Welch
  lists examples of good and bad installers in his latest
  Workingmac.com column (the first original content on the site in
  nearly three weeks)
  <http://www.workingmac.com/inetd/142.wm>

* David Pogue says in his _New_York_Times_ column that both iPhoto
  and Windows XP's photo handling tools are "welcome developments
  that may spare you the trouble and cost of choosing and buying
  someone else's program"
  <http://www.nytimes.com/2002/02/28/technology/circuits/28STAT.html>

* Dan Gillmor, writing for _ComputerWorld_ for a change, seems to
  think good Mac OS X applications "could be ported back to Linux,
  giving developers a larger target market than they have today,"
  but these would apparently be "good" applications that don't use
  Carbon, Cocoa, or Aqua
  <http://computerworld.com/cwi/story/0,1199,NAV47-74_STO68372,00.html>

* Back at his own Weblog, Gillmor says Virtual PC 5 for Mac OS X
  "should [never] have been put on the market," and that "Connectix
  has damaged its good name with this not-ready-for-prime-time
  release"
  <http://www.siliconvalley.com/mld/siliconvalley/business/columnists/dan_gillmor/ejournal/2789811.htm>

* _PC_Magazine_ rates Apple's AirPort 2.0 products just three
  circles out of five, partially because the magazine _deliberately_
  installed an older version over a newer one and had trouble with
  it, calling it only "not a typical situation" instead of something
  Apple explicitly warns against doing
  <http://www.pcmag.com/article/0,2997,s=1470&a=23456,00.asp>

* Sky.com's iMac (Flat Panel) review points out some of the
  problem Alsop doesn't solve - two people in reviewer Mike Taylor's
  office said that Office wasn't available for the Macintosh
  <http://www.sky.com/skynews/article/0,,31500-1045417,00.html>

* Mike Wendland of the _Detroit_Free_Press_ exhibits some of the
  psychotic letters he got "defending" Apple after his iMac (Flat
  Panel) review, showing once again how a few nut cases tarnish the
  image of all Macintosh users
  <http://www.freep.com/money/tech/mwend5_20020305.htm>

* Bob LeVitus says you shouldn't rush to install each new Mac OS X
  version when released, especially since there's no easy way to
  revert to older versions
  <http://www.osxfaq.com/dailytips/03-2002/03-02.ws>

* _Macworld_UK_ covers the advent of higher-powered lasers that
  should enable 4X DVD writing drives (DVD-R, DVD-RW, DVD-RAM,
  DVD+RW, DVD-SyQuest, DVD-BFD, etc.) within the next year or year
  and a half
  <http://www.macworld.co.uk/news/main_news.cfm?NewsID=4313>

* Apple CEO Steve Jobs ranks #413 on _Forbes_'s list of the
  world's 497 billionaires with an estimated net worth of US$1.1
  billion, most of it in Pixar stock (#1 is still Bill Gates, with
  an estimated net worth of US$52.8 billion)
  <http://www.forbes.com/home/2002/02/28/billionaires.html>

* The _New_York_Times_, in another widely-linked article, joins
  the months-long controversy over copy-protected audio discs that
  don't play in computers (they work more often in PCs than in Macs)
  <http://www.nytimes.com/2002/03/01/technology/01PROT.html>

* David Morgenstern moderates the fallout of David Coursey's
  recent attempt to get Linux geeks to consider Mac OS X
  <http://www.zdnet.com/anchordesk/stories/story/0,10738,2852462,00.html>

**Internet News**

* In a time-tested procedure known as "blinking first," The Mac
  Show will no longer try to compete with its former host and co-
  founder. Shawn King left in January in a dispute over subscription
  plans - currently, The Mac Show streams only one hour to the
  public, with the second hour reserved for "MSL Club" members who
  pay US$30 for six months or, in a new and inexplicable alternate
  price, US$4 per month (US$4 times 6 months is still US$24, isn't
  it? Why would anyone pay US$30 for six months now?). King started
  a new program, Your Mac Life [36], broadcasting from Nashville,
  TN, in a free two-hour format at the same time - Wednesday nights
  at 9PM ET (Thursday at 0300 UTC).

  [36] <http://www.yourmaclife.com/>

  The Mac Show blinked. It announced [37] this week that it was
  moving its broadcast to Tuesday nights at 7PM PT - that's 10PM ET,
  or Tuesday mornings at 0400 UTC. This "more convenient" time puts
  the members-only hour after prime time in half the US and makes
  the show end at midnight on the US east coast. There are also
  segmented downloadable archives so you can pick the segments you
  want to hear, and more discounts from vendors for club members.
  Meanwhile, Your Mac Life is now "out of beta" and started non-test
  shows on Wednesday night. The slogan? "Tune in to hear the
  difference."

  [37] <http://www.macshowlive.com/pressrelease.html>

* What's the difference between two "white papers" posted Thursday
  proposing that Mac OS-specific programs are faster than more
  commonly used open-source competitors? One company actually
  remembered to post its paper and the other didn't. Blue World
  Communications released a 56-page white paper [38] trumpeting its
  Lasso technology over PHP scripting, showing Lasso "over 200%"
  faster than PHP "in select benchmarks" and playing up Lasso's
  security and administrative functions that PHP scripts have
  difficulty duplicating.

  [38] <ftp://ftp.blueworld.com/reports/Lasso_vs_PHP.pdf>

  Meanwhile, 4D said it had a new white paper proving that WebSTAR
  V, the expensive Mac OS X-only Web server, is up to three times
  faster than Apache. To prove it wasn't rigged, 4D said it would
  "provide the actual test suite so you can replicate the results
  yourself," from a configuration that is "simple and 'real life.'"
  However, 4D said the information would be available under the
  "Documentation" link on the WebSTAR V page [39], and as of a full
  day after the announcement, it was not. It's available now [40],
  but all the people who looked for it Thursday based on the
  announcement left without seeing it. D-oh.

  [39] <http://www.webstar.com/51>
  [40] <http://www.webstar.com/downloads/WS51_PDF_and_Test_Files.hqx>

**Internet News In Brief**

* Disney CEO Michael Eisner complained that Apple's "Rip. Mix.
  Burn." slogan promised users "they can create a theft if they buy
  this computer," ignoring that ripping, mixing, and burning CDs
  from CDs you've purchased is not illegal if you don't give away or
  sell the results
  <http://www.cinescape.com/0/editorial.asp?aff_id=0&this_cat=Music+%26+Audio&action=page&type_id=&cat_id=270466&obj_id=33173>

* Playnet.com is seeking beta testers for the Mac OS X version of
  World War II Online, a massively multiplayer role playing game
  <http://www7.wwiionline.com/macbetaapplication.htm>

* Microsoft will pay US$100 million in penalties and offer refunds
  to customers in southern California who were billed for MSN
  Internet services they never used
  <http://news.com.com/2100-1023-852434.html>

* Although Microsoft does not support Macintosh computers with MSN
  services, the company does have a KnowledgeBase article explaining
  how to connect from Mac OS 9 (it's not difficult)
  <http://support.microsoft.com/default.aspx?scid=kb;en-us;Q221462>

* Nearly half a billion people worldwide now access the Internet
  from their homes, says Nielsen//NetRatings
  <http://www.cnn.com/2002/TECH/internet/03/07/internet.reut/>

* And 0.2% of them have iPhoto, apparently, as Apple announces the
  free photo management application has been downloaded more than
  1,000,000 times since its release two months ago
  <http://www.apple.com/pr/library/2002/mar/07iphoto.html>

* If you'd like to add the "djbdns" domain name resolver and
  utilities to your Mac OS X system (perhaps to address the single-
  threaded resolver that's built in), Matt Simerson has some
  advanced instructions and pointers
  <http://matt.simerson.net/computing/dns/djbdns-macosx.shtml>

* _Salon_'s report on "Waiting for Wi-Fi" describes how few people
  are using public wireless networks and how private ones you might
  want to use are off-limits
  <http://www.salon.com/tech/feature/2002/03/05/wi_fi_nation/>

* Some former Adobe Systems executives have started Bellamax, a
  new Net-based photo enhancement and retouching service that gets
  your photos enhanced without owning custom software for about US$3
  each
  <http://maccentral.macworld.com/news/0203/05.photoservice.php>

* Derrick Story of the O'Reilly Network explores comments he's
  received over a six month period from Macintosh users on the
  transition to Mac OS X, the direction of the platform, and more
  <http://www.oreillynet.com/pub/a/mac/2002/03/05/mac_community.html>

* Andy Pastuszak has started a Yahoo! Group for Quicken for
  Macintosh users
  <http://groups.yahoo.com/group/quickenmac/>

**Other Macintosh News**

* At OSXFAQ.com, Bob LeVitus recommends [41] a simple solution for
  misplaced Finder icons in Mac OS X: remove the Finder's preference
  files. "Now restart, and then set the global icon preferences the
  way you like them and restart again. They should stick." This is
  probably the correct answer to the question about global icon
  positions, but there's another question that's come up: how come
  the Mac OS X Finder doesn't remember specific window icon
  settings? We're not the only ones to have noticed that even though
  you view a folder by list, the next time it may view by icon, or
  with different icon sizes, or otherwise.

  [41] <http://www.osxfaq.com/dailytips/03-2002/03-04.ws>

  It has to do with a characteristic of the Mac OS X Finder that,
  frankly, we could usually live without - the ability to view the
  same icon in multiple views. The traditional Mac OS Finder works
  quite hard to make sure the same icon isn't visible more than
  once. You can never open your Desktop Folder in a window. If you
  double-click a folder that's already expanded in list view, the
  list view version collapses as the folder opens in its own window
  (conversely, if you turn the triangle to view an open folder in
  list view, the folder's window closes). Mac OS X is not so
  restrictive - you can have the same folder open in twenty windows
  if you want. It's a necessary side effect of "columns" view -
  since one window can show so much of a volume, closing every other
  window that might show an icon - or taking hard disks off the
  desktop - just doesn't work.

  The problem is that Macintosh users are accustomed to a single
  window for a single folder. When a given folder always opens up in
  its one and only window, it can open in the same view every time.
  But what happens when you start with the top level of your hard
  drive and drill down to a familiar folder? Does the view stay the
  same as for the hard drive, or change as you drill down? What if
  you've never visited the folder before - does the child inherit
  the parent view or use global defaults? Do options like "keep
  arranged by name" apply to children? The Mac OS X Finder does a
  reasonable job of showing views how you probably want them. It
  doesn't always succeed (on our production system, the main hard
  drive view should stay arranged by name but it doesn't stick), but
  it's right more often than not. Just in case you were wondering
  why folders open as they do.

* Programmer Brad Oliver (MacMAME) reported an Internet Config bug
  in Mac OS X 10.1.3 to MacFixIt [42], and the MacFixIt staff
  confirmed it, but it's not an Internet Config bug. It's a
  Microsoft Entourage bug. Open an application that can change
  Internet Config settings and open Entourage X. Change some
  Internet Config settings in the first application, quit the
  program, then quit Entourage. Then relaunch the first program and
  see all your changes lost.

  [42] <http://www.macfixit.com/>

  As Oliver says, "What appears to happen is that, when Entourage
  quits, it writes out settings to com.apple.internetconfig.plist,
  completely oblivious to the fact that an app (in this case
  Internet Explorer) changed settings in that file since Entourage
  was launched." The problem is that applications shouldn't be
  writing to that file at all - that's what Internet Config is for.
  Any applications sharing a single data file would have this
  problem. That's why they're supposed to send any changes to
  Internet Config and let _it_ write the changes to the file. Even
  if Entourage is doing that, if it incorrectly tries to write _all_
  settings when quitting instead of any that changed, you'd see the
  same effect.

  The only way it could be an Internet Config bug is if Internet
  Config is rewriting the entire file when not asked to store any
  changed preferences. That's unlikely, but possible. The posted
  workaround is workable, but another one would be not to make any
  Internet Config changes while Entourage is open. If Entourage is
  the culprit, that would solve it more easily.

**Other Macintosh News In Brief**

* IDG World Expo confirms that Apple CEO Steve Jobs will deliver a
  keynote address to open Macworld Expo in Tokyo later this month
  (2002.03.21), though the company has already made clear there will
  be no new computer products at the show
  <http://maccentral.macworld.com/news/0203/06.jobstokyo.php>

* Mark Hoffman demonstrates in a protected PDF document (password
  "macosx") how Mac OS X's "Preview" application can view and print
  PDF files that supposedly have printing locked out - and if you
  can print to a PostScript file, you can recreate the PDF without
  any security protections
  <http://www.kearney.net/%7emhoffman/acro/pdf_security.pdf>

* Crooner and actor Harry Connick, Jr. receives a US patent for
  synchronizing music display on blue-and-white Power Macintosh G3
  computers for all members of a band or orchestra using Coda's
  Finale; technology created with help of "neighbor" and former
  Broadway conductor David Pogue
  <http://www.nytimes.com/2002/03/04/business/04PATE.html>

* Agfa Monotype finally gets around to announcing that nineteen of
  its fonts have been shipping with Mac OS X for, oh, about a year
  <http://www.businesswire.com/cgi-bin/f_headline.cgi?bw.030402/220632121>

**Apple's KnowledgeBase**

* Apple has confirmed [43] a known problem in Mac OS X 10.1.2: if
  the sound hardware on late-model machines doesn't play any sounds
  within 30 seconds, the system puts the sound system to sleep. It's
  supposed to save on battery power, and the issue was first noted
  on the second-generation iBook models (Dual USB, Late 2001, 14.1
  LCD) and the PowerBook G4 (Gigabit Ethernet).

  [43] <http://docs.info.apple.com/article.html?artnum=106766>

  It's an "issue" because the sound hardware takes time to warm up
  again, so output is delayed if the sound hardware was asleep when
  the sound was supposed to start. This three-to-four-second delay
  torpedoes synchronized sound and video - maybe not QuickTime, but
  definitely PowerPoint presentations and perhaps Flash animations.
  You also get the Rainbow Spinning Beach Ball of Death for your
  cursor during the warm-up period.

  Saving battery power is a noble goal, but it's the kind of thing
  that should be under user control, and like so much of Mac OS X's
  power management (all of which has been glued on since Apple
  bought Next, as OpenStep had no real power management), it's not.
  What's worse is that the problem also affects this year's desktop
  models - iMac (Flat Panel) and Power Macintosh G4 (QuickSilver
  2002) machines - even though they're not portables and don't need
  to save battery power. You just get the inconvenience of
  inappropriately delayed sound with no tangible benefits.

  Jonathan Feinberg has a solution for you, his KeepSoundAwake [44]
  freeware application. Every 20 seconds, KeepSoundAwake plays a
  "silent" sound - it tells the sound hardware to play a sample
  that's complete silence - so the hardware stays awake and the
  delays never happen. The Web page is full of warnings from
  bystanders that it could drain battery power or make recharging
  your batteries take longer, but that seems inapplicable to the
  desktop systems. Then again, the "feature" would seem not to apply
  either, and yet there it is. Feinberg had hoped the issue would be
  addressed, or at least more controllable, in Mac OS X 10.1.3, but
  that didn't happen either. It's pretty silly to waste CPU cycles
  playing silent sounds to keep regular sound working, but that's
  what happens with motherboards, the operating system, or both fall
  victim to the "one size fits all" engineering economy syndrome. If
  you don't like it, file bugs.

  [44] <http://mrfeinberg.com/KeepSoundAwake/>

**Apple's KnowledgeBase In Brief**

* Apple provides full FireWire specifications for the iMac (Flat
  Panel) machine, including the obvious but rare admission that it
  supports IEEE 1394a (and not just IEEE 1394)
  <http://docs.info.apple.com/article.html?artnum=34589>

* Want to see an empty KnowledgeBase article?
  <http://docs.info.apple.com/article.html?artnum=36700>

* Apple documents over 100 free scripts for QuickTime 5 on Mac OS
  X 10.1 and later, even though the "documentation" refers to non-X
  entities like OSA Menu
  <http://docs.info.apple.com/article.html?artnum=42649>

* If you can get a sound into AIFF format (in a file with name
  ending in ".aiff", of course, since Mac OS X is stupid about file
  types), you can add it as an alert sound even though the OS still
  has no built-in way to record sounds
  <http://docs.info.apple.com/article.html?artnum=106409>

* If your Mac OS 9 machine can't connect via File Sharing to a Mac
  OS X machine, telling you "The connection to this server has been
  unexpectedly broken," make sure the Mac OS 9 client has an
  assigned IP address
  <http://docs.info.apple.com/article.html?artnum=106779>

* The Mac OS X 10.1 installer may cause minor damage to hard disks
  initialized or driven by InTech's Hard Disk SpeedTools software,
  but there's an easy fix
  <http://docs.info.apple.com/article.html?artnum=106781>

* Power Macintosh G3 computers that stop with "File could not be
  written" during Mac OS X 10.1 installation from CD-ROM probably
  have a badly-configured ATA bus (such as a solo hard drive marked
  as the "slave" instead of the "master")
  <http://docs.info.apple.com/article.html?artnum=106782>

**Developer News**

* Lots of programmers watch for new Technical Notes and Q&A
  entries [45] from Apple's Developer Technical Services group, but
  sometimes Sample Code [46] gets less attention. Unlike the other
  two, Apple doesn't seem to maintain a current list of the newest
  sample code, but the ADC News lists new samples from time to time.
  Sample code and documentation go hand-in-hand: samples show how to
  get things done and documentation guarantees the methods will work
  in future OS releases.

  [45] <http://developer.apple.com/qa/>
  [46] <http://developer.apple.com/samplecode/>

  Recent samples show implementing an IODataQueue [47] in a kernel
  extension so it can communicate with a regular application, using
  nib files [48] made by Interface Builder in Carbon applications,
  and an entire "MoreSCF" framework [49] that simplifies creating
  locations and configuring network ports (Ethernet, AirPort, PPP).

  [47] <http://developer.apple.com/samplecode/Sample_Code/Devices_and_Hardware/IOKit/SimpleDataQueue.htm>
  [48] <http://developer.apple.com/samplecode/Sample_Code/Human_Interface_Toolbox/Converter_nib.htm>
  [49] <http://developer.apple.com/samplecode/Sample_Code/Networking/MoreSCF.htm>

  Also of note: a full contextual menu plug-in [50] for Mac OS X, in
  both Project Builder and CodeWarrior 5 and later. We've bemoaned
  the lack of useful contextual menu items in Mac OS X. There have
  been more in recent months since Mac OS X 10.1 made third-party
  plug-ins possible, but there could and should be more, and this
  should help. Sample code is cool.

  [50] <http://developer.apple.com/samplecode/Sample_Code/Human_Interface_Toolbox/SampleCMPlugIn.htm>

**Developer News In Brief**

* Xplain Custom Services, part of the _MacTech_ organization, is
  hosting an exhibitor fair at WWDC 2002 as in past years;
  reservations for the "cost-effective" exhibits are available via
  phone or E-mail
  <http://www.mactech.com/news/mt-news-detail.mgi?id=0000050eb8a4f3a6>

* Daniel Steinberg explains how to use AppleScript Studio to
  compile and run Java code at the O'Reilly Network - the
  application passes parameters to the javac compiler and runs the
  output
  <http://www.oreillynet.com/pub/a/mac/2002/03/01/osx_java.html>


Product Showcase: The Week's Top 30
-----------------------------------

**1. Flash MX**

  If you're an animator who targets the Web, Macromedia Flash is on
  your shelf. Some people prefer Adobe's recent LiveMotion 2.0 [51]
  for its advanced scripting, integration with other Adobe software,
  and better frame rate management. Some people like going further
  and using full-blown QuickTime movies authored with Totally Hip's
  LiveStage Professional [52]. But the simple animations you see on
  most sites - the big ads at CNet, the introductions to new
  products, the cartoons - those are all Flash, baby. The program
  had a storied history [53] even before Macromedia purchased it and
  changed it from "FutureSplash Animator."

  [51] <http://www.adobe.com/products/livemotion/main.html>
  [52] <http://www.totallyhip.com/lo/products/lsp/>
  [53] <http://untoldhistory.weblogs.com/stories/storyReader$4>

  Flash MX [54] is coming in a couple of weeks, says Macromedia
  [55], and it's Mac OS X native, matching LiveMotion 2.0's major
  platform feature just weeks after LiveMotion 2.0 shipped. In Flash
  5, Macromedia made a raft of interface changes and added
  "ActionScript" (better known as JavaScript) as a primary control
  language to the product. Some of these interface changes were
  well-received, and many were not.

  [54] <http://www.macromedia.com/software/flash/>
  [55] <http://www.macromedia.com/macromedia/proom/pr/2002/flash_mx.html>

  In Flash MX, the first in a line of beyond-the-Web "MX" products
  from Macromedia, most of the unpopular interface changes are
  rolled back to Flash 5 versions. All the tabbed panels are gone
  (perhaps part of the Adobe-Macromedia lawsuit over dockable tabbed
  palettes), replaced by dockable "collapsing" palettes - a title
  next to a disclosure triangle. Turn the triangle and see the full
  palette, or turn it back to drop it to just a single line. Dock as
  many of these palettes in one floating window as you like; Flash
  MX automatically collapses other palettes to save screen real
  estate. If you want more visible at once, put them in separate
  floating windows. It's the same functionality as dockable tabbed
  palettes but without the potential patent dispute.

  Even better is a "Property Inspector," a Dreamweaver and
  QuarkXPress-like palette that displays editable attributes for the
  currently selected object, whatever it may be: text options for
  text, animation options for frames, and so on. The timeline is
  improved as well - first by undoing what Flash 5 did, then by
  adding usability changes like more intuitive double-click actions,
  "Delete" removing only objects on the Stage, and more. We're
  getting most of this from Webmonkey's review of Flash MX [56], by
  the way, the most comprehensive look we've seen so far and worth
  your time if you're a Flashaholic.

  [56] <http://hotwired.lycos.com/webmonkey/02/09/index4a.html>

  But we're ignoring the major new features. Flash MX imports video
  into its projects, so you can have video in your Flash animations.
  On the one hand, Macromedia says the feature is not designed to
  replace dedicated media players like QuickTime or RealPlayer, just
  to integrate short videos into existing Flash projects. On the
  other hand, Macromedia's people are quick to point out that
  playing your video in Flash MX makes it simpler.

  "A lot of sites require the user to choose which video player
  they'll use, and creators have had to produce their content in
  multiple formats," Flash product Manager Jeremy Clark told
  MacCentral [57]. "Now you can now create just one video stream.
  You can incorporate video right inside your Web site, add it to
  existing Flash content, and make it interactive." Even so, Clark
  also said - in the _same_ article - that "our focus with
  Macromedia Flash MX is allowing people to add video capabilities
  to the already existing Macromedia Flash interactive capabilities.
  Also, to differentiate further, you can view video in Flash
  immersive in one client as opposed to launching external media
  clients." Dennis Sellers is normally on the ball, but we're unsure
  how he let Clark get away with saying that before Flash MX,
  "there's been no good way to incorporate video right inside a Web
  site."

  [57] <http://maccentral.macworld.com/news/0203/04.flashmx.php>

  The video in Flash MX uses the Sorenson Spark codec, as both
  companies announced [58]. Following the familiar QuickTime
  pattern, Sorenson has a professional version of Spark available
  with its new Sorenson Squeeze fro Flash MX [59] product, a US$299
  bundle that offers two-pass variable bitrate encoding for better
  and smaller files. There's nothing on the Web site about an
  upgrade for existing Squeeze owners, and since existing versions
  of Cleaner or other tools wouldn't output Flash MX files, it looks
  like Flash MX and Squeeze are the only games in town for now. Toon
  Boom Studio is also offering a free importer [60] so you can make
  projects in that program and then pour them directly into Flash
  MX.

  [58] <http://www.sorenson.com/company/press_room/press_releases.asp>
  [59] <http://www.sorenson.com/products/squeeze.asp>
  [60] <http://www.toonboomstudio.com/product/tbs_importer.ch2?prodinfoid=2>

  Not all new features are so provincial or usurping. Flash MX lets
  you separate common graphics and audio files (like JPEG pictures
  or MP3 sounds) from the actual Flash file, making them easier to
  update and making the movie faster to download. A "Free Transform"
  tool replaces separate rotate and scale options, with a new
  ability to distort an object's "envelope" (the Bezier curves that
  comprise its frame). You can split all objects on a layer into
  separate layers, make sure objects are an exact number of pixels
  wide, and more. On windows, Flash MX files can conform to Section
  508 disability standards, but the implementation depends on Active
  Accessibility in Internet Explorer for Windows. You can save any
  movie in Flash 5 format, making it QuickTime 6-compatible and
  available to colleagues who haven't yet upgraded.

  There's lots more in there - more server-side tools imported from
  Macromedia's purchase of ColdFusion, the elimination of the "Flash
  Generator" product due to new Flash MX built-in features,
  libraries of common interface widgets, and a built-in capability
  to target handheld devices or other non-Web outlets. There's
  source-level debugging in Flash or in a browser, an ActionScript
  editor that Webmonkey doesn't necessarily believe is improved, and
  "much more," as the ads say.

  The free Flash 6 player is to be available on 2002.03.14 - it's
  called "Flash 6" instead of "Flash MX" so it's easier to see it's
  later than the Flash 5 plug-in - and the full package on
  2002.03.15. New purchases are US$499, with upgrades [61] from any
  previous Flash version (including Flash FreeHand Studio or Flash
  Generator Studio) going for US$199. Boxed versions won't be
  available until after the downloadable upgrades, so plan
  accordingly.

  [61] <http://www.macromedia.com/software/flash/productinfo/upgrade/#how>

**2. Virtual PC 5.0.2**

  US$79 Windows emulator for Macintosh gets a maintenance update for
  much better Mac OS X performance, adjusting the processor time it
  takes for most tasks (including video), fixes issue for sound
  initialization causing a crash at launch, and adds fix for rare
  crash if saved states were restored on older machines, plus much
  more
  <http://www.connectix.com/support/vpcm_online.html>

**3. Quicken Deluxe R3 Updates**

  As much as Intuit deserves slams for its anti-Macintosh decisions
  and their lame justifications (like discontinuing QuickBooks for
  Macintosh because the company couldn't make any profit controlling
  only 60% of the Macintosh small business market where smaller
  players thrive), it deserves praise for recent updates to not only
  Quicken Deluxe 2002 [62] but also Quicken Deluxe 2001 [63], the
  older version that the company no longer sells.

  [62] <http://www.intuit.com/support/quicken/updates/qkn_updates_mac2002.html>
  [63] <http://www.intuit.com/support/quicken/updates/qkn_updates_mac2001.html>

  Both updates include a new secure certificate so you can keep
  using online features over SSL - in other words, securely. The
  2001 update also fixes problems with "Save a Copy," with crashing
  when switching between mutual fund and portfolio accounts, with
  reports containing splits of more than 14 lines, and more. The
  2002 update fixes incorrect splits in the "Security Detail"
  window, a missing "Lots" button in the "Investment Actions"
  window, the same report calculation error, plus problems with
  "QuickMath" (the built-in calculator) and cosmetic problems with
  downloading transactions in Mac OS X. Both updates are free and
  strongly recommended, especially if you use Quicken's online
  transactions.

**4. LiveMotion EPS Support 2.0**

  Rather than delay the recent LiveMotion 2.0 shipment by a few
  weeks, Adobe released the program and is updating its EPS support
  online via a new plug-in; the company recommends that all users
  download and install it though there are no notes about why
  <http://www.adobe.com/support/downloads/detail.jsp?ftpID=1434>

**5. OpenSSH 3.1**

  Since version 10.0.1, Mac OS X's remote login capabilities (for a
  command-line prompt, not for file sharing or Web server or
  anything else) have been provided by OpenSSH, an open-source
  secure shell protocol. OpenSSH replaces telnet because telnet uses
  clear text - your passwords and all your commands are visible to
  anyone who can snoop TCP/IP packets. OpenSSH uses SSL so all
  communications are encrypted, just as if you were viewing a secure
  Web page. Mac OS X 10.1.3 includes OpenSSH 3.0.2p1, where the "p"
  means the "portable" version compilable on Mac OS X and many other
  operating systems (MWJ_ 2002.02.23).

  That version has a security bug [64] - a single missing "=" in the
  source code that could, according to OpenSSH's developers, allow
  an authenticated user to hack a vulnerable OpenSSH host or vice-
  versa. The bug affects all versions between 2.0 and 3.0.2, and is
  corrected in OpenSSH 3.1, released Thursday.

  [64] <http://marc.theaimsgroup.com/?l=openssh-unix-dev&m=101555552910337&w=2>

  Unfortunately, it's not as easy to update as a simple download and
  decompression. Like most open-source software, you first have to
  fetch the source code, then compile it (so you must have Apple's
  Mac OS X Developer Tools installed), and then put it in the right
  place. You'll need root-level access to do it, but if you can get
  that far, there's a thorough step-by-step tutorial on the process
  from Scott Anguish at Stepwise [65]. The instructions first build
  the tcpwrappers utilities that OpenSSH may use, then build OpenSSL
  and finally OpenSSH itself, each time grabbing the appropriate
  source files from Stepwise where they have been updated.

  [65] <http://www.stepwise.com/Articles/Workbench/2001-12-17.01.html>

  To gain root access, you can either log in as "root" (after
  enabling the root user in NetInfo Manager) or use the command
  "sudo -s". That opens a new instance of the shell with root
  privileges after you authenticate. After that, just execute the
  commands in Anguish's tutorial one by one. We used BBEdit 6.5's
  shell window to execute them in groups - you _can_ authenticate
  with sudo in a BBEdit shell window, but you have to remember to
  enter your password on a blank line and press "Enter." Don't type
  it on the same line as the "Password" prompt or, when you press
  "Enter," BBEdit sends "Password:yourpasswordhere" instead of
  "yourpasswordhere" to the shell.

  Be warned - these commands put out a lot of output and some
  warnings as well, including one that OpenSSL isn't configured for
  multithreaded applications in the default installation, but that
  seems OK (your Web browser uses its own SSL implementation
  anyway). Test your security when done installing, and save the
  shell output so you can examine it later for difficulties. We
  tried it and found no problems, but this is an advanced operation -
  if you don't use remote login much (for example, if you don't
  have it turned on in System Preference's "Sharing" panel), you can
  probably wait for Apple's next Mac OS X update to fix this for
  you. If you want the latest and greatest, it's available, Stepwise
  shows how to install it, and it works.

**6. TinkerTool 2.11**

  Free Mac OS X utility to customize hidden options of Mac OS X
  fixes an issue with multi-display computers running Mac OS X 10.1.
  <http://www.bresink.de/osx/TinkerTool2.html>

**7. Apple DVD Player 3.1.1**

  Available only through Software Update at press time (as was the
  prior 3.1 update), the new Mac OS X DVD video player adds features
  for DVD authoring like the ability to select and play contents of
  a "Video_TS" folder from a local volume, plus AppleScript support
  for playback and gets a new "wide" layout for info window; no news
  on changes in Friday's 3.1.1 update

**8. Synchronize Pro X 1.0.2**

  We know we've looked at Qdea's Synchronize Pro X [66] several
  times recently, but this first "bug fix" update to the US$100 file
  synchronization and backup utility gains an interesting new
  feature - it can make _bootable_ backups of Mac OS X startup
  disks. There are catches - you must first log in as root and use
  Synchronize Pro to copy the entire bootable volume to a _blank_
  disk - but it works. The update also fixes problems with including
  or excluding folders by name and with proper files not being
  correctly ignored. Few utilities can make bootable Mac OS X disks,
  and to have the ability in a program you may already own is quite
  nice.

  [66] <http://www.qdea.com/pages/pages-sprox/sprox1.html>

**9. Font Reserve 3.0.1**

  US$90 font management tool (MWJ_ 2002.02.23) fixes "a problem with
  DVDs and other external devices" and fixes an installation problem
  on Mac OS X systems without Classic installed
  <http://www.fontreserve.com/products/frmac.html>

**10. Tattoo 1.0.3**

  US$23 Mac OS X utility creates skins for QuickTime Player with
  custom control buttons for wired actions (MWJ_ 2002.02.18) adds
  four more button actions (mute, half volume, full volume, close
  window), lets buttons be invisible, lets you change the
  compression setting on the skin, adds cursors for buttons, and
  puts video on top of other layers
  <http://www.feelorium.com/Tattoo/>

**11. Location X 1.0.3**

  Among the many useful features of Mac OS 9 that Apple's
  engineering executives have not seen fit to provide in Mac OS X is
  the Location Manager. Code-named "Walkabout" and originally only
  for PowerBooks, the Location Manager lets you define system
  settings for a specific _place_ where you use your computer - the
  QuickTime settings, network settings, energy saver choices, and
  more. When you change "locations," as easy as clicking in the
  Control Strip, all the settings you have configured change. Set up
  "home" and "work" locations and change your printer, network
  connection, and more just by clicking once. The genius of the
  Location Manager is that Apple opened it up to third-party
  developers, so in addition to standard Control Panel settings, a
  change in location can also flip your Conflict Catcher extension
  sets, some mail account settings, Internet Config settings, and
  more.

  And yet this entirely useful technology is not available for Mac
  OS X, so in the best Macintosh tradition, shareware developers
  have filled the gap. Location X [67], from Alex Keresztes and Greg
  Novick, brings the same idea to Mac OS X. Don't get confused - the
  Network pane in System Preferences uses the term "location" for a
  set of network settings, what Open Transport would have called a
  "configuration." It's even more confusing because these are
  available in the standard Apple menu, but only affect network
  settings.

  [67] <http://homepage.mac.com/locationmanager/>

  Location X provides Walkabout-style locations, starting with four
  modules: QuickTime settings, network settings (what Mac OS X calls
  "locations"), the time zone, and Mail.app settings. All modules
  feature a "Capture" option that sets location parameters from your
  current settings. The program includes a system menu component
  (like the sound, displays, or scripts menu), and a plug-in SDK for
  developers to add their own modules is coming soon. Here's hoping
  that Apple buys technology like this to encourage shareware
  developers instead of just putting them out of business when it
  finally gets its act in gear. Location X costs US$10; version
  1.0.3 is a quick update to "resolve an issue that could cause
  potential data loss in certain user scenarios," or in English, "it
  doesn't forget your settings sometimes."

**12. Freeway 3.5**

  SoftPress Systems's Freeway has always been an odd duck among Web
  design applications. Freeway is more like a page layout program,
  and inherited a great deal from UniQorn, SoftPress's page layout
  application for QuickDraw GX. Using GX Graphics and Typography,
  Freeway allowed laying out a Web page like a printed page, with
  Freeway automatically generating GIF and JPEG images for all the
  fonts and graphics you'd laid out, keeping the "plain" text in
  regular HTML. Over the years, it's been upgraded to support other
  image types (like TIFF and PNG), JavaScript, Flash support, and
  more.

  Only now, however, has it finally ditched its tie to the long-
  unsupported GXGraphics extension. Freeway 3.5 [68] updates the
  application to Mac OS X native status, using ATSUI typography as
  the successor to QuickDraw GX. The new version has a full Aqua
  interface and adds transparency for TIFF and 24-bit PNG images,
  allowing image compositing directly within Freeway. Advanced
  drag-and-drop support for everything from Illustrator files to
  Flash movies helps assemble pages more quickly, too. Best of all,
  it's a free upgrade to the US$250 program. If you want to develop
  artistic sites without using a ton of separate programs, give
  Freeway a free trial.

  [68] <http://php.softpress.com/softpress/macosx/>

**13. ChronoSync 1.0.1**

  US$20 Mac OS X file synchronizer update fixes a bug with symbolic
  link and alias resolution
  <http://www.econtechnologies.com/site/chronooverview.htm>

**14. Apple's White Input Devices**

  Since introducing the Apple Pro Keyboard and Apple Pro Mouse in
  July 2000 with the Power Macintosh G4 (Gigabit Ethernet) models,
  Apple has made the devices available as peripherals for US$59
  each; we purchased a few to update our computers last year because
  the original half-size Apple USB Keyboard didn't have full-size
  function keys (it's not a _general_ complaint, but one that
  affected us). Now Apple is selling a minor revision to each with:
  the Apple Pro Keyboard (White keys) [69] and the Apple Pro Mouse
  (White color) [70]. They're the same electronics as previous
  versions, but the keys and the mouse internals are white instead
  of black to better complement the iMac (Flat Panel) machine. This
  keyboard and mouse ships with the iMac (Flat Panel); the black
  models still ship with the Power Macintosh G4 (QuickSilver 2002)
  and CRT-based iMac models (PowerBook and iBook computers don't
  come with keyboards or mice); each is priced identically at US$59.

  [69] <http://store.apple.com/1-800-MY-APPLE/WebObjects/AppleStore.woa?productLearnMore=M8691LL/A>
  [70] <http://store.apple.com/1-800-MY-APPLE/WebObjects/AppleStore.woa?productLearnMore=M8690G/A>

  A warning from an MDJ_ reader: the original "black color" Apple
  Pro Keyboard has a slightly narrow USB connector so it can fit
  into the USB cable extender that comes with all of Apple's desktop
  computers. The new Apple Pro Keyboard (White keys) does _not_, so
  it won't fit into those existing keyboard extenders. They are,
  however, easier to read in low light. It's always a trade-off,
  isn't it?

**15. Fire 0.30.b**

  Free Mac OS X Cocoa-based instant messaging client for AIM, ICQ,
  MSN, IRC, and Yahoo! Messenger sees fairly major update: passwords
  now stored in the keychain (you'll have to reenter all account
  settings anyway), completely new MSN library support requiring all
  buddies to use complete Passport IDs, automatic update checks, new
  accounts controller, and tons of bug fixes
  <http://www.epicware.com/fire.html>

**16. Ready Set Go 7.2.8**

  US$150 desktop publishing system popular in many parts of the
  world now at least works under Classic in Mac OS X (previous
  versions did not); update also fixes printing error from
  unexpected characters in text block, uses default color screening,
  and gets a new "Drop Caps" menu
  <http://www.diwan.com/english/pages/inside/updates/728upd.htm>

**17. Codec Calculator 2.0**

  Free utility calculates data rates for any QuickTime 4 video or
  audio codec, calculates maximum file size for videos, and converts
  from between kilobytes per second and kilobits per second; update
  gets support for QuickTime 5 codecs (adds eleven new codecs in
  all), gets better MPEG prediction, and a "Clear" button for byte
  converter
  <http://raytraced.cjb.net/>

**18. DivX 5.0 Alpha**

  If you've felt left out of the DivX revolution so far, cheer up.
  Mac OS 8 and above, including Mac OS X, are full citizens in the
  new DivX 5.0 Alpha [71] release. The new release is much faster,
  and although PowerPC G4 optimizations aren't ready yet, the
  Windows version has AMD optimizations that allow _triple-speed_
  real-time encoding on dual-processor systems. The Mac version only
  has some G3 optimizations, but is faster than older versions and
  will get faster yet. It reads and writes an MPEG-4 file format and
  is fully MPEG-4 video compatible, and it's compatible with all
  versions of DivX video, stretching back to version 3.11 alpha. If
  there's DivX video you want to see, DivX 5.0 will view it. It has
  full support for I-VOP, P-VOP, B-VOP, MPEG-4 Simple and Advanced
  profiles, and H.263 video, too. This release also includes a
  workaround for a QuickTime bug that keeps MP3 sound in AVI movies
  from playing properly. There's no Mac version of Divx Pro, but
  it's a better pre-release than the version 4 one and it's still
  free. If you like DivX movies, you want this update now.

  [71] <http://www.divx.com/divx/divx_mac_versions.php>

**19. NotifyMail 4.1.4**

  US$18 Mac OS 9 and X utility notifies you when new E-mail arrives;
  update fixes sound problem, adds support for QuickMail Pro, and
  adds a new preference so you do not get notified too often
  <http://www.notifymail.com/>

**20. BTV Pro 5.4**

  US$40 video capture, editing, and stop-motion animation tool
  features time-laps capture, motion detection, and DV video input
  and output; update can always add new frames to the end of movies
  if animating, gains an option to preview video during DV stream
  capture, gets a "Deinterlace" filter for captured and exported
  images, plus minor bug fixes
  <http://www.bensoftware.com/>

**21. Cypher 1.0.2**

  New US$5 "donationware" encryption and decryption tool for Mac OS
  X is a front end to the open-source "ccrypt" command-line program
  that implements the US Government's Advanced Encryption Standard
  (AES)
  <http://homepage.mac.com/jhammer/cy/>

**22. Sputnix 0.6b**

  Sputnix [72] is a Mac OS X-only unofficial client for the
  AudioGalaxy music-swapping service that has no bona fide Mac
  client yet available. We noted the like-minded MacSatellite back
  in September (MWJ_ 2001.09.26), but it's become harder to find.
  Sputnix is a spare, simple approach and seems to work well. The
  current version boasts better protocol compatibility, owing to
  some pointers from the company itself, and it's free. Bob "Dr.
  Mac" LeVitus [73] suggests we point you to OpenAG X [74], an
  open-source client based on Eric Seidel's decoding of the
  AudioGalaxy protocols. LeVitus suggests it may be even more
  polished than Sputnix, and it's also free, but was not revised
  this week.

  [72] <http://www.biggerplanet.com/unsupported/sputnix.html>
  [73] <http://www.boblevitus.com/>
  [74] <http://homepage.mac.com/macdomeeu/dev/current/openag/>

**23. EIMS Server 3.1.1**

  Free update to EIMS 3.1 (US$60 from version 3.0) fixes problems
  with "Incoming Mail" folder used by AutoShare, fixes problem with
  LDAP authentication, and fixes NTLM authenticating problems with
  Outlook
  <http://www.eudora.co.nz/updates.html>

**24. Kerberos 4.0**

  Free Kerberos authentication utility provides Kerberos 4 and 5
  protocols; update expands native Mac OS X support, native KfM
  Kerberos libraries, fixes a memory leak, and adds support for
  Kerberos during remote connections
  <http://web.mit.edu/macdev/Development/MITKerberos/Common/Documentation/release-4.0.html>

**25. FutureBASIC 3 Version 6**

  US$169 integrated development environment for application
  development in BASIC - around longer than REALbasic but getting
  far less publicity because it's been behind; the new version
  finally adds access to all Appearance Manager controls, generates
  Carbon-compatible code, preference files, contextual menus, and
  other long-time necessities
  <http://www.stazsoftware.com/futurebasic/latest.php>

**26. Mutsu Repository Applications**

  At first glance, these open-source AppleScript Studio projects
  from Mutsu Repository [75] look like alternatives to for-sale
  utilities, but they're lacking the amenities of the non-
  AppleScript versions. Simpleshade 1.2 does collapse windows to
  their title bars, just like WindowShade X, but it only works on
  Finder windows, it only works on _all_ Finder windows, and it
  resizes the windows to the smallest height possible instead of
  truly collapsing them. When you uncollapse them, they're not the
  same size as they were before. A "Smartshade" variant comes closer
  but is slower. RSSX 1.0b2 is a primitive RSS headline reader, but
  it only works on sites that are located on a hard-coded Web page.
  There are others, but you get the idea. On the bright side,
  they're free, the source code is included, and they're written in
  AppleScript Studio so you may have a better chance at modifying
  them for your needs.

  [75] <http://sourceforge.net/projects/mutsu>

**27. DeskTastic 1.1**

  Free Mac OS X tool with amazing technology (MWJ_ 2002.02.28) to
  draw directly on the desktop now has new pen colors and an erase
  function
  <http://www.panic.com/~stevenf/index.php?node=DeskTastic>

**28. OSXMacPerl 0.2**

  Perl on the Macintosh has evolved interestingly. Unlike most
  operating systems, Mac OS has no built-in command-line
  functionality and no text-oriented shell, so porting the popular
  scripting language to Mac OS raised some challenges. MacPerl [76]
  got around that by making Perl a full OSA scripting language,
  accessible from any application and a full peer with AppleScript.
  Now Mac OS X is here, and it's a shell-based operating system with
  the latest Perl distribution built-in - but now scripts that use
  the special features of MacPerl are left out in the cold.
  Integrating the two tracks is Gero Hermann, who has created
  OSXMacPerl [77]. Now at version 0.2, the new module adds support
  for "DoAppleScript" and other specialized MacPerl functions. We
  love it when a plan comes together.

  [76] <http://www.macperl.org/>
  [77] <http://news2.ils.uec.ac.jp/~herr/>

**29. iPod Free File Sync 2.3**

  Free utility for Mac OS 9 or X synchronizes music from different
  iPods or any other source to your computer; update shows correct
  message if target disk is full, works better with localized iTunes
  versions, and no longer skips "duplicate" songs, plus more
  <http://www.cooc.de/softies/index.html>

**30. iPod mismanagement tools**

  It's hard to believe that when this week started, iPod owners had
  no good solutions for turning their MP3 players into fake PDAs,
  and now there are _two_ of them. Talk about the free market
  overfilling a need. Panorama iPod Organizer [78] was shown at
  Macworld Expo a few months ago, and it's still a goofy idea: it
  creates zero-length MP3 files named for the people in your address
  book, so you can treat your iPod like a read-only PDA in disguise.
  It's US$20 and shipping now if your silliness tolerance is higher
  than ours. If you're both silly _and_ cheap, ipoAdress [79]
  (that's how it's spelled) does the same thing in freeware, but is
  a more generic database - you determine what fields each MP3 tag
  means so you control what the iPod displays. Since a real PDA is
  much less expensive than an iPod and all this clutters up iTunes
  and your real playlists, it's hard to imagine why anyone would go
  for either of these programs, and yet they're both out within a
  week.

  [78] <http://http://www.ipodorganizer.com/>
  [79] <http://www.kohlenbach.de/prod_ipoadress_engl.htm>


MacCyclopedia: The Tao of X Printing
------------------------------------

**Turning an accident into an architecture**

  For a computer that achieved prominence through desktop publishing
  and the "what you see is what you get" (WYSIWYG) principle, it's
  always been something of a miracle that any application can print
  a close representation of anything displayed. Classic Mac OS
  printing has been a boondoggle from the start because the system
  was never comprehensively designed. It merely evolved from what
  Apple shipped first - a driver for a single printer, the
  ImageWriter dot matrix printer. Background printing, 2800 DPI
  printers that cost under US$60, imagesetters, and PDF weren't even
  on the horizon.

  Mac OS X finally adds what the OS has needed for at least a
  decade: a comprehensive, modular, modern printing architecture.
  Pieces of the Mac OS X printing world are well-defined and perform
  small tasks, are interchangeable with each other, and contribute
  to a common human and application interface that demystifies each
  particular part of the process.

  You may be thinking, "Wait - Mac OS 9 printing works just fine for
  me, and Mac OS X printing is slow and confusing." Mac OS 9
  printing only works well for printers from a limited number of
  manufacturers - Apple, HP, Canon, Epson, Brother - and only
  because they or programmers they hired spent anywhere from three
  to ten years learning how to make Mac OS printer drivers that work
  with most applications. Ever seen those little receipt printers or
  label printers and thought, "Wow, I could get that and hook it up
  to a FileMaker Pro database and have something useful?"

  Wrong - most printers other than the big desktop models have no
  real Mac OS printer drivers. They're nearly impossible to create
  so they work with all printers. Understanding the train wreck that
  has been Mac OS printing is the first step to seeing why the new
  architecture is both necessary and welcome. It's been a few years
  since MWJ_ tackled this subject (MDJ_ covered it extensively in
  late 1996 and early 1997 when discussing QuickDraw GX, the system
  Apple at the time promised was the future of printing). QuickDraw
  GX and GX printing have gone away, but some of the concepts and
  architectural lessons Apple learned survive in Mac OS X, to
  everyone's benefit. In this special feature, we walk the road from
  old to new, showing how Mac OS X arrived where it did.

**Accidental ink**

  The classic Mac OS is built around a drawing technology called
  _QuickDraw_. It combines some resolution-independent objects like
  lines, rectangles, arcs, ovals, and polygons, with resolution-
  dependent bitmaps and irregularly-shaped regions. Just as painting
  in Photoshop depends on the current settings (such as brush size
  and shape, foreground color, and layers), QuickDraw's drawing
  depends on the graphics environment settings such as foreground
  and background colors, pen size, pen pattern, fill color, and
  more. These settings, along with a description of the width and
  height of the drawing area and regions for clipping, comprise a
  _graphics_port_, or grafPort in programming terms.

  QuickDraw always draws into a graphics port - on the desktop, in
  every window, and in most off-screen buffers (programmers manage
  some buffers themselves, but if they want QuickDraw to draw into
  their buffers, there must be a graphics port tied to them). When
  it came time to print, the solution seemed simple: make a special
  "printing graphics port" and tell applications to draw into it.
  Instead of real drawing - that is, changing actual pixels in
  memory - QuickDraw would instead call upon the printer driver to
  change the drawing commands into codes that make the printer put
  ink on paper. Programs would just "draw to the printer" the way
  they drew to the screen and everything would work out fine with
  the same code. It was supposed to be incredibly easy.

  It was more like a nightmare. Printing requires data that drawing
  on the screen does not - the range of pages, the number of copies,
  the paper size - so Apple created a _print_record_ to define a
  printing job. It contained all this data as well as a reference to
  the printing graphics port. That's where the problems started. The
  original print record was defined by the ImageWriter printer
  driver because until the ImageWriter driver, there was no printing
  system. That driver used 120 bytes of private and public data,
  sharing some with applications (like rectangles defining both the
  size of the page and the size of the printable area, since the
  ImageWriter couldn't print edge-to-edge) and keeping some private.

  The original _Inside_Macintosh_ books documented almost every part
  of the print record under the philosophy that should rule -
  telling developers how it works and trusting them not to use the
  parts they shouldn't use except for debugging. That philosophy
  hasn't held up well over time. Programmers have repeatedly shown
  that if they're told how something works but told not to use it,
  they'll use it anyway and try to fix it if it ever breaks. Of
  course, it usually _does_ break eventually, and if the company
  doesn't feel like spending the money to fix it, they just tell
  customers not to upgrade to the latest hardware or software.

  To be fair, developers really did need to know more than the
  ImageWriter driver wanted to tell them. Did you choose to print in
  portrait (normal) or landscape (sideways) mode? There was no way
  to tell, except to see if the page was wider than it was tall, and
  no way to force one mode over the other. Did you want to know if
  background printing was turned on? There originally wasn't any
  background printing, so there was no way to mark it as active or
  inactive. How about getting the scaling value, like 50% or 200%?
  Sorry, only the driver needs to know that - it just adjusts the
  rectangles for paper and page size and fools the application into
  printing normally, taking care of the math itself.

  Programmers are a hardy lot, though, and they were undeterred by a
  lack of ways to do it. Apple's developer support groups helped,
  showing how to look at a print record before and after the "Page
  Setup" and "Print" dialog boxes to see what had changed and how
  values were stored in the print record. In this way, developers
  learned how the ImageWriter driver did its work, and relied upon
  it to get their own work done.

**From bad to worse** -- Naturally, the arrival of the _second_
  printer driver - the LaserWriter driver - broke almost everything.
  The LaserWriter driver didn't work like the ImageWriter driver at
  all. The ImageWriter driver recorded QuickDraw commands into a
  picture and played them back over and over again, drawing the page
  one small horizontal band at a time. It changed that narrow band
  of pixels into ImageWriter commands to print dots, then repeated
  the process until the entire page was printed. That was pretty
  much the only way to go - the printer couldn't print more than one
  band at a time, and the system didn't have enough memory to hold a
  pixel map for an entire page at once anyway.

  The LaserWriter driver, on the other hand, captured QuickDraw
  commands and changed them into PostScript language code on-the-
  fly. There was no actual drawing, no pixels to send (unless the
  application was printing a pixel map, of course), no banding.
  There were, however, several new concepts - error reporting, font
  downloading, bitmap smoothing, precision bitmap alignment, and
  PostScript files, just to name a few.

  At this point, everyone realized that the ImageWriter driver had
  defined a 120-byte print record for its own use and left very
  little room for any other printer driver to reinterpret those
  fields. The LaserWriter driver had to keep reasonable values in
  all the fields that applications were examining, while finding new
  places to store its own private values. The print record was the
  only storage guaranteed to stay with a print job, and the
  ImageWriter driver had limited it to 120 bytes because that's all
  the ImageWriter needed.

  The LaserWriter driver authors also discovered lots of small
  problems that programs had created through reliance on the
  ImageWriter driver. A favorite was in the Finder, which checked a
  certain address in low memory (one for the use of printer drivers)
  and, if it was zero, refused to print icons. Finder's low-level
  icon drawing routines avoided QuickDraw for performance reasons
  unless printing, and this specific, bizarre method is how Finder
  knew it was printing, based on how the ImageWriter driver happened
  to work. The LaserWriter driver authors had all the source code to
  examine for help, and it still took years to get the driver in
  great shape.

  It wasn't just the Finder, though, and it wasn't just the
  LaserWriter driver. As the Macintosh started to sell well and
  printer manufacturers wanted to support it, they all found
  themselves in the same boat. If you wanted to write a printer
  driver, it meant about six months of writing the basic code to
  make well-behaved applications print, and another twelve to
  eighteen months of tracing through popular applications to figure
  out what they were doing, why they were doing it, and what they
  expected to happen.

  When first released, most of these printer drivers had subtle
  problems in most programs. It couldn't have been any other way -
  the vast majority of printer driver dependencies between the
  system and applications were completely undocumented, and remain
  so to this day. It took years of evolution to make basic printer
  driver toolkits that worked with most applications. The companies
  that had enough money to develop them or contract to get them -
  HP, Epson, Canon - got good printer drivers. Other manufacturers
  did without.

**Printer-specific parts** -- As printers grew more capable,
  programmers and printer manufacturers wanted to take advantage of
  new advanced features. The printing system that evolved from the
  ImageWriter driver didn't allow that very easily or, in some
  cases, at all. Developers have been frustrated for years with the
  "Printing Manager," the set of Mac OS APIs that allow drawing into
  a printing graphics port and eventually put ink on paper. They
  imagine that the Mac OS implements a common set of printing
  routines like recording pictures, presenting "Page Setup" and
  "Print" dialog boxes, handling errors, providing USB and AppleTalk
  communications with printers, and so on. After all, that only
  makes sense, right?

  That's not how it works. Not to invoke images of the red pill, but
  there is no Printing Manager. _All_ those APIs do is load the
  printer driver you selected in the Chooser and dispatch to it. The
  Mac OS does _nothing_ else in printing - everything is performed
  by the printer driver. That's another reason it takes so long to
  develop a printer driver. Applications tend to assume unsupported
  facts, like that the page number field in the "Print" dialog box
  is dialog item #3, and nothing restricts any printer driver like
  that. Over the years, applications have grown more conservative
  and tend to behave themselves more often - at least when adding
  new code - but it's been a huge compatibility problem over the
  years.

  It's also been a problem for printer driver writers trying to add
  support for their printers features. PostScript has a more capable
  imaging model than QuickDraw (though not perfect), and one of the
  first things the early desktop publishing programs wanted to do
  was send their own PostScript code to a LaserWriter printer. There
  was no API for that - the system was designed around the
  ImageWriter, and it had no concept of accepting PostScript code.
  On the other hand, the original printing API includes low-level
  calls to stream text directly to a printer or to advance the paper
  1/6th of an inch - requests laser printers shrug at.

  Over the years, as these capabilities became more common, Apple
  added support for an API called PrGeneral that changed general
  printer settings instead of starting or stopping a job or page.
  The advent of the LaserWriter brought lots of PostScript-like
  demands to printing. In addition to sending custom PostScript
  code, developers wanted to take advantage of PostScript's ability
  to rotate graphics without losing resolution, so the LaserWriter
  driver adds a PrGeneral selector that rotates the printing
  graphics port by a specified amount. Developers can turn the
  LaserWriter driver's line layout routines on or off (the ones that
  try to make PostScript line spacing approximate QuickDraw's on-
  screen text spacing), rotate graphics or text, and download
  partial fonts - a useful feature for large Japanese or Chinese
  fonts that don't fit completely in some PostScript printers.

  Unfortunately, the barn door was already open - the ImageWriter
  printer driver was already out there without support for these
  features. Apple revised it to return "I can't do that" for
  applications that asked the ImageWriter driver to rotate text or
  accept PostScript code, but Apple couldn't make the ImageWriter
  match the LaserWriter's capabilities. Also, all the drivers had
  been out for a few years before it became clear that the system
  needed a way to let applications know what PrGeneral selectors
  were and were not supported. An application that tries to rotate
  graphics can ask the driver to do it, but if it doesn't work it
  may fail silently such that the printout includes unrotated
  graphics. That's not WYSIWYG. There is a complicated system to
  work around this, involving drivers "hiding" some drawing if they
  support certain operations, but it's a true boondoggle.

  Even supporting a printer's native resolution is fraught with
  pitfalls. A PrGeneral selector lets applications find out what
  resolutions a printer supports, so graphics applications can crank
  up the resolution and print the best possible quality. If your
  printer is capable of 2800 DPI, why print a 72 DPI downsampled
  version of a photo? But drivers don't have to give true resolution -
  or any resolution at all. The LaserWriter driver returns that it
  supports one resolution - 300 DPI - even though it's the driver
  for all PostScript printers. As it turns out, some applications
  didn't behave very well when the driver returned anything _other_
  than 300 DPI, and since companies like Quark were actively
  encouraging customers not to upgrade LaserWriter drivers because
  it broke their printing hacks (you know how slow Quark is with
  upgrades), it stayed at 300 DPI to preserve compatibility.

  All of these features, by the way, are dependent upon the printer
  driver in effect when an application prints. Since there was no OS
  support for spooling files to disk or even implementing background
  printing, every printer driver records drawing in the printing
  graphics port its own way. Since the format is private to each
  driver, the jobs can't be moved from driver to driver. It wouldn't
  work even if you could. Options like a printer's resolution are
  queried and answered while the application is printing (when you
  see the status dialog box on the screen), so if you printed while
  an Epson printer was selected, the application might find a
  resolution of 2800 DPI. If you then redirected the job to a 360
  DPI StyleWriter, it would be completely out of whack. That's not
  to differences between monochrome and color printers or any number
  of other device-specific problems.

**That's why** -- You might think a slam on Mac OS 9 printing is
  counterintuitive, since it seems to work so well. The part most
  people don't know is that they're looking at the back end of an
  18-year battle. It took untold thousands of programmer-years for
  programmers to get their applications printing something close to
  the same ink on a wide variety of printers and to take advantage
  of your printer's features without printing horrid ink on _other_
  printers.

  This kind of accidental system is what happens when its designed
  for one use - an ImageWriter printer - and expanded far beyond its
  original capabilities. It's why printer companies often update all
  their drivers at once, too. Having spent dozens of man-years on a
  working printer driver framework, they use it for all their
  drivers, so a bug fix to the framework affects all the current
  drivers. Every printer driver has to handle every task - dialog
  boxes, status reports, even talking to the printer. That's right,
  every printer driver has its own serial or AppleTalk or USB code
  built-in, unusable by any other printer driver. It's why every
  printer's dialog boxes look different, every background printing
  utility is different, and why "WYSIWYG" printouts on similar
  printers using different drivers have such noticeable differences.

  It's too late to change it significantly, either. Apple finally
  released enough information a few years ago to let third-party
  printer manufacturers use desktop printing, but no major
  manufacturer (to our knowledge) has yet adopted it, preferring to
  stick with its own background printing. Why? Because it _works_.
  Changing printing code opens the possibility of months of
  debugging, just like cleaning up code to remove old hacks runs the
  risk that dozens of current applications will stop printing
  properly. It's a great mystery and entirely too fragile.

  It's also not as useful for you as it should be. Printer drivers
  quickly needed storage beyond the ImageWriter driver's 120-byte
  print record, but the printing API design doesn't allow for
  expanding the print record. Drivers therefore starting allocating
  their own storage in global variables, and it worked because the
  Macintosh was a single-program system - you could only print one
  document at a time. When MultiFinder and System 7 rolled around,
  printer drivers depended on not being reentrant. They can't handle
  trying to print a document while they're already printing a
  document. Therefore, even on today's dual-processor 1GHz machines,
  Mac OS 9 can print only one document at a time.

  The sad part is that it didn't have to be this way. There could
  have been more printers, more driers, more features common to all
  printing - if only there had been a printing architecture.
  Instead, every printer driver was responsible for every part of
  the printing process, so developers reinvented the wheel over and
  over and over again, all without documentation.

  A real printing architecture would have made it all _so_ much
  easier. It could present standard dialog boxes and provide well-
  defined ways for both applications and printer drivers to extend
  them, instead of requiring printer drivers to use a specific kind
  of dialog box and requiring applications to patch into it. It
  could segment the printing process so each piece of code has a
  specific small task, rather than forcing every printer driver to
  handle everything from human interface to network communications.
  It could work with the system's graphics architecture to minimize
  every printer driver's work, provide consistent storage for
  graphics that flow through printing, make PostScript just another
  part of the graphics system and not some mysterious private
  component of the LaserWriter driver.

  In fact, Apple tried all that once before and failed. That also
  helps explain why Mac OS X printing works the way it does.

**QuickDraw GX Printing**

  Apple spent over six years developing a "new printing
  architecture" that eventually became QuickDraw GX Printing.
  Originally planned as part of System 7 in 1991, the printing
  changes were delayed because Apple decided to merge them with its
  new graphics and typography architecture. By building all three
  together, Apple hoped for and got optimizations that made sure
  every QuickDraw GX graphic would print perfectly and faster than
  otherwise.

  GX Printing added a lot of features that had been missing from the
  printing experience. It changed printing dialog boxes from modal
  to movable modal, so you could switch applications while changing
  print settings. It offered well-defined printing extensions, so
  applications or true extensions could add printing features with a
  real human interface. It provided full support for converting any
  GX graphics to bitmaps, PostScript code, or even into a set of
  short straight lines for vector devices like plotters. Printing
  extensions could transform any part of the printing process - GX
  Printing shipped with sample extensions that provided watermarks,
  2-up and 4-up printing, changing page order, typography
  transformations, and more.

  Since GX Printing provided all of the systemic mechanisms for
  spooling, it also imposed a universal spool file format. In other
  words, all printers stored printing jobs in the same way. You
  could open a desktop printer (a GX invention later appropriated
  for non-GX use), double-click any job waiting to be printed, and
  view it in SimpleText. Want to print it on another printer? No
  problem - drag it to another desktop printer and it printed there
  instead. A special printer driver converted print jobs into
  _portable_digital_documents_. PDDs were viewable on any system
  with QuickDraw GX installed, and contained all the glyphs from any
  font necessary to exactly reproduce the document.

  QuickDraw GX Printing provided just about everything that
  application developers had wanted from the printing process since
  1984. So why didn't it succeed? The primary cause was politics.
  Apple had promised developers up, down, and sideways that
  QuickDraw GX was _the_ future of graphics and printing on the
  Macintosh. GX Printing was to be the _only_ supported printing
  method in "Mac OS 8," the original name for Copland. When Copland
  was scuttled in late 1996, QuickDraw GX was still the wave of the
  future.

  But the NeXT purchase in December 1996 was the last straw. It
  didn't take the NeXT engineers more than a month to insist that
  they would not revise their PostScript-based frameworks to use the
  "proprietary" QuickDraw GX graphics and printing, despite a
  superior imaging model. GX Printing was inexorably tied to the
  graphics model, though. Older applications that used GX Printing
  internally converted their QuickDraw commands to GX graphics to
  make it work. When QuickDraw GX bit the dust, printing was the
  first part to go - in late 1997, Apple issued a "GXGraphics"
  extension with QuickDraw GX graphics and typography support but no
  printing.

  GX Printing had been unpopular in many quarters because, as the
  first major revision to printing in ten years, it bore the brunt
  of compatibility problems. The GX Printing developers - correctly -
  chose not to implement every tiny trick and undocumented feature
  from the old world in GX's compatibility layer. Since GX would be
  _required_ in a new version of the OS, they saw its earlier
  release as a fine notice to developers to fix their code not to
  use such undocumented and unsupportable behaviors. Publishing
  developers, often a surly lot not inclined to change, instead told
  their customers not to install QuickDraw GX and refused support to
  those that did. They lobbied Apple hard to drop GX rather than fix
  their code to work correctly, and Apple's new management team
  listened.

**The problems** -- In fairness, full GX Printing support was not
  a trivial commitment. In response to years of problems caused by
  no easy way to modify the printing process (other than praying the
  printer driver did what you wanted), GX Printing was in a sense
  over-engineered. To make sure any program could get any control it
  wanted, GX Printing defined hundreds of _messages_ that the
  architecture sent back and forth. When it was time to despool a
  page from disk and image it, for example, the "GXDespoolPage"
  message got sent from QuickDraw GX to all printing extensions.
  They could intercept the message and change it, or pass it along
  to the other extensions. It eventually got to the printer driver,
  whose job was to implement the message or return it back up the
  chain to GX for the default implementation.

  It was all very flexible, but it was hard to get the big GX
  picture. A programmer could do almost anything he wanted, _if_ he
  could find the right message to override and intercept it at the
  right time. All the messages didn't make GX a speed demon on the
  machines of the day - 68040 machines or the very first Power
  Macintosh machines that topped out at 80MHz PowerPC 601
  processors.

  Basic GX printing was very simple for a GX application, and basic
  customization wasn't too complicated in concept. But when you
  start throwing dozens of messages around, the chance for
  misunderstanding among programmers or bugs in the process increase
  too much. Everything from allocating memory to preparing to
  rasterize a portion of a page generated a new message. It was too
  much for developers who weren't completely focused on printing.

  Despite the short-sighted failure of a huge and worthwhile
  project, GX's legacy is still all over the system. A stripped-down
  version of QuickDraw GX graphics found its way into QuickTime 3
  and later as QuickTime Vector graphics [80], still part of
  QuickTime but rarely mentioned. Cocoa could have learned a lot
  from it, though - GX graphics are object-oriented, just like Cocoa
  itself, and the two would have made a good match.

  [80] <http://developer.apple.com/techpubs/quicktime/qtdevdocs/REF/refVectors.2.htm>

  The NeXT system used Display PostScript, a horribly inefficient
  way to draw on the screen. Instead of a simple API call like "draw
  a rectangle with these corners," NeXT applications had to generate
  PostScript code in text format and send it to the Display
  PostScript server to be interpreted and, eventually, drawn. Mac OS
  X's graphics system, Quartz2D, uses a PostScript imaging model
  (Apple says "PDF", but PDF is a PostScript imaging model) but is
  API based instead of the text-based Display PostScript
  abomination. It has most of the same graphics features of GX,
  including transparency and full printing support, but it's not
  quite as programmatically powerful. It is, however, more standard.

  Apple Type Services for Unicode Imaging (ATSUI) is a reworked
  version of QuickDraw GX typography, but without all the graphics
  capabilities (like an easy way to turn any line of text into the
  curves for each font glyph). It's still head-and-shoulders above
  the typography provided by any other operating system on the
  planet, and it's finally getting the recognition it has always
  deserved in Mac OS X. Although Cocoa applications don't use ATSUI
  as much as they should, the OS itself does in Aqua - all the menu
  items, dialog boxes, and other text in the human interface is
  handled by ATSUI. So is all text layout in the built-in Unicode
  text editors that programmers use. There are still speed
  improvements to be found, but ATSUI is holding its own.

  Mac OS X printing, our core subject, clearly learned lessons from
  the failure of GX Printing. The design is flexible but not open at
  every possible step because that was overkill. The most popular
  features of GX Printing survived, but in a new and more limited
  format that should still serve most printing needs.

**The X printing philosophy**

  In QuickDraw GX, the printing philosophy was, "We'll provide every
  service a driver or extension wants if it wants to use them, but
  it will still do most of the work." In Mac OS X printing, that's
  reversed: "The system does every possible task, letting drivers
  and extensions only do the work the system can't possibly
  perform." For example, every printer driver provided its own "Page
  Setup" and "Print" dialog box layout in QuickDraw GX, just as in
  the old world. Mac OS X provides standard "Page Setup" and "Print"
  dialog boxes, offering printer drivers or other code a way to
  _extend_ them if necessary.

**Dialog features** -- Printing dialog boxes in Mac OS X are
  standardized. The "Page Setup" dialog box always includes the
  "Page Attributes" panel where you choose a printer's settings to
  use for formatting, the paper size, the scaling percentage, and
  the orientation - portrait, landscape, or reverse landscape (in
  other words, rotated 90[deg] either clockwise or counter-
  clockwise). It also has a "Summary" panel that provides a text-
  only summary of all your "Page Setup" options. These panels are
  the same for every application. If a program needs more choices,
  it can add its own panels (selected through the "Settings" pop-up
  menu at the top of the dialog box), but it cannot change these
  panels.

  The "Print" dialog box is more complex but similar. Every panel
  starts with your choice of the printer and the "Presets." Some
  printer drivers in Mac OS 8 and 9 allowed you to save default
  settings by pressing the "Option" key when clicking "OK", or
  through some other mechanism. Mac OS X makes it explicit by
  letting you save _one_ collection of settings for any printer,
  name them, and recall them in any "Print" dialog box as the
  "Custom" presets. The option to save custom settings is at the
  bottom of the panel selection pop-up menu.

  "Copies & Pages" provides an interface for choosing how many
  copies you want printed, whether they should be collated, and what
  page range you want printed. The "Summary" pane again provides a
  text-only summary of the chosen print settings that would be saved
  as presets if you wanted. Mac OS X also provides a "Layout" panel
  for N-Up printing (up to 16 pages per sheet with your choice of
  borders and page ordering on the sheet). If your printer says that
  it supports _duplex_ printing (putting ink on both sides of a
  sheet of paper), Mac OS X provides a standard "Duplex" pane for
  turning the feature on and choosing a "binding" (so pages would
  look correctly in a notebook, either bound horizontally or
  vertically).

  "Output Options," available in Mac OS X 10.1 and later, allows
  saving a job to a PDF file. If your currently selected formatting
  printer is a PostScript printer, you may also save the print job
  as a PostScript file. There is no "Encapsulated PostScript" option
  since nearly every program that's been revised in the past few
  years can take PDF as easily as EPS. Of course, you can always
  choose the "Preview" button instead of "Print" to get a PDF
  version of the print job anyway, but the "Output Options" pane
  lets you save that setting for later use, perhaps in scripted
  printing.

**PostScript extras** -- Mac OS X Printing provides other standard
  panes, but only for PostScript printers. The "Paper Feed" pane
  picks the paper tray for the first page and all subsequent pages,
  or for all pages in a job, something that the LaserWriter driver
  has always had in the main (or first pane) of the "Print" dialog
  box but not something you change for every job. "Error Handling"
  provides options for reporting PostScript errors and how to
  proceed if a tray runs out of paper but another tray has paper of
  the same size.

  Some printers also include a "Printer Features" pane, sometimes
  with a tabbed interface. You see that when the printer's
  PostScript Printer Description (PPD) file says it has features not
  handled by any of the other panes. If the PPD file organizes these
  features into groups, Mac OS X displays the group names as tab
  names to show their relationship. If not, the tabs are called
  "Features 1," "Features 2," and so on.

  Mac OS X currently supports only two kind of PPD feature options:
  those described as "PickOne" (a pop-up menu), or "Boolean" (a
  checkbox - but if a boolean option isn't defined with exactly two
  options, "True" and "False", it's a pop-up menu too). There's a
  "PickMany" option in the PPD specification, and while you might
  think of it as a group of checkboxes that lets you pick as many as
  you like, Mac OS X doesn't support it. That's a layout issue -
  each features pane is limited to five one-line options. A
  "PickMany" option could include a few dozen choices, and laying
  out that many checkboxes on the fly is a fairly significant
  interface problem on smaller screens. Mac OS X responds by not
  trying. If an option group has more than five features, they're
  split into multiple tabs - "Stapling 1" or "Stapling 2," for
  example.

  Those are all the standard panels. No "ColorSync" panel? Nope -
  color matching is performed automatically by Mac OS X, though some
  printers may offer additional settings. No "Font Options"? Uh-uh.
  Mac OS X always downloads versions of the fonts it is using unless
  it knows you're using the same PostScript version of a font that's
  in the printer, because otherwise you may be subject to the same
  kind of layout errors prevalent in Mac OS 8 and 9. (Non-PostScript
  printers don't have to worry about it anyway.) There's no
  "Background Printing," either, because _all_ printing in Mac OS X
  is background printing. The only reason it's stayed in Mac OS 9 so
  long is because some systems didn't have enough disk space to
  spool huge professional jobs, but in days of 80GB hard drives for
  US$150, Apple has decided that's not a problem.

**Print Center** -- Replacing the Chooser and PrintMonitor (or
  desktop printing in later days) is a new application you've
  undoubtedly experienced if you've printed anything in Mac OS X.
  Print Center is a universal queue manager and printer discovery
  program. Rather than have icons on the desktop for each printer -
  or at the "Computer" level of the Finder, where they'd logically
  fit - Mac OS X lists all of your printers in the Print Center
  application. Technically, they're not "printers", they're
  _print_queues_, structures created to store jobs for each printer
  the system knows about. That's important for a few tips later on.

  You create print queues in Print Center via the "Add Printer..."
  button. It displays a dialog sheet asking you to pick a kind of
  printer and specify the printer within the kind. Default kinds
  include LPR printers using IP, USB printers, and AppleTalk
  printers, just like the old Chooser would have done. Mac OS X
  separates the _transport_ (like AppleTalk) from the _printer_
  (such as a PostScript printer). You can choose any printer
  available over the transport you choose. Some printer developers
  pair transport and printer together, so you may have choices like
  "EPSON AppleTalk" and "EPSON USB" in addition to "AppleTalk" or
  "USB", but it's still a pretty clear system.

  Print Center doesn't know how to find printers connected to your
  computer - it relies on plug-ins to show you what printers are
  connected via AppleTalk, USB, LPR over IP, or even specified in
  NetInfo ("Directory Services"). Those plug-ins are called
  _printer_browser_modules_, or PBMs. They provide everything you
  see for choosing a printer once you've picked a communications
  method - a PBM - in Print Center's "Add Printer" sheet.

  Apple's printer browsers are generic. If you pick "USB", you'll
  see any printer on the bus identifying itself as a printer, with a
  pop-up menu so you can say what kind of printer it is so the OS
  can pick the right driver. That's why EPSON provides its own
  "EPSON USB" printer browser. It detects only EPSON printers on the
  bus and knows which model is connected, eliminating extra
  configuration steps. Pick the printer from the printer browser
  module and Print Center creates a print queue for it.

  Once you have a print queue, Print Center acts as the status
  center for all printing jobs. It's where you put jobs on hold,
  restart them, or delete them. There's no support for printing at a
  specified future time, like PrintMonitor allowed. Print Center
  presents normal (boring) status, or updates you on its progress
  through its Dock icon. It seems like a pretty staid and
  uninspiring application, but Print Center is indeed the central
  way you interact with Mac OS X printing. The lack of exotic
  options may seem dull, but it also means that printing behaves in
  predictable and consistent ways - something of a novelty for the
  Macintosh.

**The printing flow**

  Behind the scenes, the same philosophy of doing everything
  possible guides print jobs from applications to paper. It starts
  with the application. Just as before, programs "draw" into a
  printing graphics environment, whether with QuickDraw or Quartz2D.
  Unlike Mac OS 8 and 9, though, Mac OS X itself implements the
  printing APIs, so they all work the same way no matter what your
  default printer is - the printer driver isn't even involved at
  this point, unlike the old world where the driver did
  _everything_. Mac OS X can therefore enforce certain conventions
  that were impossible to standardize in Mac OS 9. Each application
  starts a job, then starts a page, draws the page, and closes the
  page. It repeats those last three steps for each page to be
  printed, then closes the entire print job. In Mac OS 9, strange
  things happened if you got these steps out of order. Mac OS X
  returns errors and makes programs do it the right way.

  Many applications want to put their own items in the "Page Setup"
  or "Print" dialog boxes. Mac OS X allows that through
  _printing_dialog_extensions_. Unlike QuickDraw GX printing
  extensions that could affect every part of printing, these PDEs
  _only_ add items to the human interface. PDEs can be part of
  applications or other system code, or even part of a printer
  driver - that's how drivers can implement their own custom
  interfaces for features not provided by Mac OS X. (PostScript PDEs
  must tell the system if they provide an interface for some feature
  in the PPD file so Mac OS X doesn't provide its own generic
  interface for that feature.) PDEs must also supply a text-only
  summary of what they do for the "Summary" pane if the system asks
  for it.

  Mac OS X enforces a standard spool file, and it separates the
  drawing of pages from other data, like copies and page ranges and
  anything else in a printing dialog box (whether from a built-in
  feature or from a printing dialog extension). It's no surprise the
  limited 120-byte print record has been replaced by an extensible,
  opaque _job_ticket_ for all non-drawing information associated
  with a job. That leaves all drawing commands separated, and in
  fact, Mac OS X spools all page drawing into a PDF file. The PDF
  file plus the job ticket comprise the print job. That's also how
  "Preview" works for free with every application. Since PDF can
  perfectly represent every page printed, viewing that PDF is a
  perfect preview of the job. The printer driver isn't even involved
  at that point.

  The PDF file and the job ticket travel to the _print_server_ - the
  root-level process named "PrintServer" always running in the
  background. It's not where most of the work happens. PrintServer
  is the brains behind Print Center, since Print Center isn't
  running every time you print. It manages the queues and responds
  to requests from Print Center to stop or resume jobs. You tell
  Print Center and it tells PrintServer, reporting results back to
  Print Center to report to you. PrintServer also gets status from
  other parts of the printing system so Print Center can display it.

  While managing status and queues, PrintServer hands off each job
  to the _Print_Job_Manager_, the workhorse of the Mac OS X printing
  system. Three other kinds of modules complete the printing system,
  and Print Job Manager coordinates them all. Two of these modules
  are external to the OS. _Printer_modules_ are what most people
  imagine as "printer drivers," providing device-specific printer
  commands to put ink on paper. _I/O_modules_ implement a
  communication path, like AppleTalk or USB, and usually work with
  printer browser modules to help you set up printers in Print
  Center. _Converters_ transform graphics between one format and
  another. Mac OS X includes converters that change PDF graphics
  back into Quartz2D commands, or into PostScript, or into
  rasterized bitmap data ready for inkjet printers to process.

  Now we can complete the picture. The Print Job Manager examines
  each incoming job ticket to see which printer gets the job. It
  finds the right printer module and the transport, like USB or
  AppleTalk. The PJM then asks the printer module what kind of data
  it needs. An inkjet printer module probably wants rasterized
  pixels, but a PostScript printer driver would want PostScript
  code. A native version of PDFWriter might want the data already in
  PDF format so it could just optimize it and add new features. The
  Print Job Manager calls the right converter, gets the graphics
  back in the right format, and sends that data to the printer
  module.	

  The printer module does what a printer driver should - it turns
  the graphics data into whatever commands the printer hardware
  requires to put ink on paper. Those commands, as a big block of
  data, go back to the Print Job Manager, where the appropriate
  transport has already been identified. The PJM sends the commands
  to the printer through the I/O module. Both the I/O module and
  printer module may return status messages to the Print Job
  Manager, where they're passed back through PrintServer to Print
  Center to you.

**Implications of power** -- It's certainly more complicated than
  "the printer driver does everything," but it's got all the power
  anyone needs. First, it's _modular_. In every version of Mac OS
  before 10.0, _all_ of these modules are built into _every_ printer
  driver. Each driver includes a Chooser component, its own printing
  dialog boxes (all of it, not just the parts different from a
  standard), communications code, conversion from QuickDraw to
  whatever format the printer wants, status reports back to the
  user, background printing, spooling, and more. In Mac OS X, a
  printer module needs only convert graphics to printer commands -
  the system handles all the standard parts. Apple supplies most of
  the converters and I/O modules, too.

  Because the system includes so much of the standard code, updates
  to the system can improve everyone's printing. Using the same
  graphics engine to rasterize pixels for the screen and for printer
  drivers makes sure you get what you see - no more driver tricks to
  force QuickDraw into unnatural resolution tricks to match advanced
  printer features. Quartz2D is up to the task. If the PostScript
  converter improves, all PostScript printer drivers improve with it
  (though since the built-in driver is PPD-based, there probably
  won't be many more PostScript drivers).

  Mac OS X printing modules are _much_ easier to write than classic
  printer drivers because they're smaller. They also promote reusing
  code. If you have a printer with both a USB and FireWire
  interface, the manufacturer can release a printer browser and
  printer module now piggybacking off Apple's USB I/O module. Later,
  when there's more time, it can add a FireWire I/O module (if Apple
  doesn't write one soon), giving you new printing options without
  revising the other code, or at least without revising it much.

  Since it's year-ago release, Mac OS X has seen lots of printer
  drivers - mostly from the big players, but they already had
  rasterization and communication code completed that they could
  port fairly quickly to the new system. Writing new printer modules
  should take a few months, not a few years. The graphics flow
  through the system in PDF format, useful not only to the system
  but to you. Above all, absolutely everything that _can_ be
  standardized in the operating system _is_. No more depending on
  printer driver writers to implement the entire printing process
  for you.

**What about Carbon?** -- The Mac OS X printing system applies to
  all native applications. It does not, however, work for _Carbon_
  applications. This entire new architecture can't work in Classic
  or under Mac OS 8 or 9 because, to do so, all of those modules and
  the background-only applications that control them (like