Review While the third update to Mac OS X, Panther, was an essential upgrade for Mac users, the fourth has presented Apple's marketeers with something of a challenge. The ritual that we call the annual OS upgrade is Apple's best publicity showcase after January MacWorld - a chance to remind the world that it doesn't just make iPods. And it's a sensible occasion to introduce major system wide updates. It's also an opportunity to charge rent - and a predictable revenue stream is something software vendors have longed for years. Microsoft, Oracle and Sun amongst them. But Microsoft's Licensing 6.0 scheme has flopped even amongst business customers, not least because enterprises are sceptical that the company can deliver the goods within the lifetime of the subscription.
Apple has a different problem. As OS X improves, it becomes harder to convince OS X users to make the jump. If Microsoft had announced that the next version of Windows XP would sleep and wake up within three seconds with near 99.99 per cent reliability, would pick up a WiFi network within 10 seconds with similar consistency, and was now free of viruses, then users would flock to upgrade. But even the first, barely usable version of Mac OS X boasted this when it first appeared March 2001. The issue here isn't tempting Windows switchers, but whether an annual $129 can be justified.
2003's Panther release was everything OS X should have been in the first place, fixing many long standing performance issues. Tiger continues that all round improvement, and feels crisp compared to Panther. (It's hard to overestimate the cumulative effect of these tiny performance gains. Computer enthusiasts will often spend hours or days writing on a macro that saves a few seconds, and isn't used very often. Yet the tiny improvement in Preview, and its ability to allow you to view a slideshow, will probably gain the typical user as much as a complicated workflow.)
So OS X Tiger sees the Mac in excellent health. However for many existing OS X users there isn't a single compelling reason to upgrade. As a consequence, reckon some users, it is a focus on the cosmetic and the superficial. Which is exactly where the focus shouldn't be.
"Today's reality is that Apple has to convince consumers that making these [kernel] changes is worthwhile. The only way they can do this is to hype visible, cosmetic changes with lots of attention getting bling," writes one poster on an Apple enthusiast board, in a thread entitled "Were the Most Anticipated Features of Mac OS X 10.4 Just A Gimmick?"
Is this justified?
After a month of hands on use, it's hard not to sympathize with the accusation.
Apple claims to have made over 200 improvements in Tiger. But on closer examinations these include such essentials as "Buy Printing Supplies", a graphics equalizer for the DVD Player, and "Export Bookmarks" from Safari, giving the impression that Apple was stumped as to how to sell the Tiger upgrade. Each bundled Dashboard widget is listed as a separate improvement.
We counted over 20 new desktop backgrounds in Tiger - why aren't these listed as new features, too?
More troubling however is that in areas where Mac OS X Tiger does offer impressive potential advantages over its predecessors, these are hampered by poor and often inexplicable interface design decisions. Tiger also loses points by removing features computer users have long taken for granted. Let's deal with the latter first.
Saving an MP3 that you've loaded from a web page and played in Safari now requires an additional $29.99 payment for Quick Time Pro. That will be reason enough not to upgrade for many. Roxio's Toast no longer burns songs purchased from Apple's Music Store. Right-clicking to download a file from Safari still works, but for how long is anyone's guess. The trend isn't in the right direction - Apple has gradually been removing multimedia features from its software products (see Apple de-socializes iTunes).
It's hard to escape the conclusion that Apple now views the Mac as a platform for a closed home entertainment system - based on iTunes and QuickTime - rather than an open computing platform. AirPort Express is a great example of how a little vision, and terrific engineering, can be spoiled by this new approach. Using Airport Express, it ought to be possible to pipe audio wirelessly from any Mac application to the remote speakers - which should appear as another sound output device in the control panel. But Apple crippled the software, forcing the user to pipe music through iTunes.
The Mac is becoming the incredible vanishing media platform!
The Life Aquatic
Apple makes much of three Tiger features: Dashboard, Automator and Spotlight. Of these two are laudable attempts to solve long standing issues with personal computing. The other is a silly gimmick that typifies the demoware approach to software development.
Dashboard widgets ripple onto the desktop with what we hope will be the last Aquatic metaphor from the UI team. But as a metaphor, a watery desktop never made much sense to begin with. Documents "liquify" to and from the Dock - but when documents get wet, shouldn't they curl up, and the ink smudge, too? Configuring Dashboard slides the entire watery pool to the top of the screen, rather like a Bond villain's swimming pool sliding back to reveal an ICBM launch pad. When metaphors start silly, they can only get worse.
On Apple's own support boards, concern about the performance impact of Dashboard is widely expressed. But this is overstated, and isn't the real issue with Dashboard. (With 30MB free memory - on a machine with 1.5GB of RAM - Dashboard claims half of what's available when invoked. It doesn't matter how many widgets are active. The memory is slowly reclaimed by the garbage collector, and this compares very favorably with Konfabulator, a clunky rip-off of Stardock's DesktopX product for Windows.) The real issue with Dashboard is that it's a solution looking for a problem. We've had equivalents such as Desktop X and Konfabulator for several years now, and they've yielded thousands of clocks, media controllers and dancing Hula girls. Shed the gimmicks and redundant applications and what's left isn't too different from the set of desk accessories that shipped with the original Mac.
As a consequence, Sherlock has been neglected, and the third party application that inspired it, Watson, has been acquired by Sun Microsystem and forgotten. Someone needs to go dumpster diving at Sun and rescue the latter. We always thought the hype behind web services was overstated - but Sherlock is still an excellent way of navigating eBay and the easiest way of finding CD album cover artwork (try the FirstRiver Amazon plug-in if you don't believe us).
The marketing focus on the slick, but useless Dashboard does rather detract from Spotlight and Automator, which are extremely promising.
We'll leave an in-depth review of Automator for when it's had time to mature (our version has two top-level "Help" menus); but it was surprising to see it didn't pick up AppleEvents exposed by third party applications. It does however support shell scripts. Which leaves Tiger's crown jewel, Spotlight.
In the Spotlight
Spotlight is a system level content indexing engine that's also available as a service for developers. Mail and the Address Book make good use of it. You can group all your contacts at Sunshine Desserts in a moment, or email on particular topics, or from certain users.
We were delighted to discover that the 101kb limit on content that Spotlight indexes has been removed since the first Tiger beta. This handicap renders Google useless as a serious research tool, although the search engines have been loosening their corsets over the past year, indexing slightly larger files. We can't say what the upper bound is, but test documents over 1.5MB were fully indexed in Tiger.
Because the kernel now supports notifications, Spotlight queries are kept right up to date. Add a word to a file and it's reflected in the queries almost immediately. So from a technical point of view, the file system team has done a terrific job.
The problem with Spotlight is in everyday use.
For example, some simple searches are now much harder. If you merely want to search for a particular file by name, you'll need to use an undocumented feature, and wrap the search term in quotes. Unlike the search results window in Panther, or in Mac OS versions prior to Sherlock, the Spotlight window is now an orphan. If it loses the focus, you need to use another undocumented feature: apple-space-space to return the focus. If you want a simple list of files over a certain size, you need to resort the results displayed.
More seriously, the user interface severely hampers what queries can be made. The dearth of boolean operators (AND is permitted) means that it isn't possible to query for documents containing "Microsoft" and "Antitrust" but not "EU". There is a query builder that allows content to be specified as a field, but without even the limited boolean qualification that an iTunes Smart Playlist allows: permit all or any, it's useless. As a "smart folder" this is pretty dumb.
A "smart" folder: the only boolean it understands is 'AND'. Dumb.
Nor is it possible to tell if it the document contains one or many instances of the phrase.
Most serious of all however is the lack of context provided by the search results. Google returns a couple of fragments from sentences surrounding the text, and mature standalone search tools such as dtSearch and Copernic highlight the text in a two pane view. But in its first implementation the Spotlight API doesn't provide this information.
Well organized users may be able to infer the context. If you're spectacularly well-organized, you may have a folder called "EU Court of Appeals Decisions", but Spotlight results don't display that alongside the result (you need to click each item returned), and of course most of us aren't that methodical.
If Google results offered this much context, Larry and Sergey would still be at school
Hopefully this will be implemented in the next update. It's so very nearly there. For example, a query for "Tevanian deposition QuickTime" found a PDF document and loaded it with the search term. But Preview doesn't support multiple word queries, and so mistakenly returns "0 occurances".
Mail is another example where searching is slower than before.
Quite idiotically, it isn't possible to search by sender or subject without first embarking on a mailbox-wide query. The previous version of mail permitted this. Now with the Spotlight-enabled mail search (which takes up twice as much disk space as before, 2GB in this case) you need to start a search, stop it right away, and then use one of the buttons that appear only after the first results from the aborted search have been displayed. There is no dialog that allows you to limit the query to specific fields or build a specific query.
Does no one at Apple use Mail?
Persistent queries, or as Apple calls them Smart Folders, show the promise of system level searching, but also show how poor UI designs can scare users away. "All documents in the last three days" displays images, even though Apple lists Images (or "Presentations") as a Document. Want to look only for files by extension? Good luck - search by extension has disappeared. A simple Boolean "NOT" would solve this kind of nonsense, but it isn't an option.
A UI summary
Some of the cosmetic changes smack of change for change's sake - such as the pale blue look for the Mail application. We won't dwell on these particular UI aspects - John Siracusa has done a fine job over the years and goes into depth here.
What's of more concern is that Tiger's best feature, Spotlight, violates two of the founder's favorite maxims.
One was expressed in a Wired interview in 1996. " Some people think design means how it looks. But of course, if you dig deeper, it's really how it works. The design of the Mac wasn't what it looked like, although that was part of it. Primarily, it was how it worked," he said, correctly.
The other is the often repeated, but unofficial corporate goal of "making great technology easy to use."
Tiger fails on both of these counts. Spotlight is great technology, but it fails because the poor UI lets it down: its potential isn't tapped. And Dashboard was only ever about bling.
How did this happen? Plenty of people will blame the demo sensibility: if it looks great in a demo then Apple considers the job done. But I suspect it owes more to corporate paranoia. By guarding Tiger secrets so closely Apple has created an echo chamber. Surely someone, somewhere could have pointed out the lack of boolean operators
Is Tiger in good shape? Resoundingly so. Is it worth the upgrade? Well, at Patip Plaza in Bangkok I was offered Tiger on DVD for 600 Baht. At $15, or £8.20, you'd be silly not to.®