Finding Leopard

This time for sure! Maybe.

Predicting major improvements for sore spots in Mac OS X is a long-standing tradition. Like most Apple-related predictions, these things tend to be repeated until they come true, or until a decade passes, whichever comes first. When it comes to sore spots in Mac OS X, one thing immediately springs to mind: the Finder.

Dissatisfaction with the Mac OS X Finder is endemic in the Mac user community, ranging from mild frustration to deep-seated rage. Unsurprisingly, each new Mac OS X release has been the vehicle for a parade of Finder fantasies. Early on, when the Carbon/Cocoa split was still not widely understood and the less-known Cocoa API was assigned near-magical properties, all the ills of the Finder were blamed on its Carbon roots. A new Finder, rewritten using the Cocoa API, became the holy grail.

Mac OS X 10.1 arrived with no Cocoa Finder in sight, but hey, give them some time, right? The Cocoa Finder mania ramped up a bit for the first publicly branded big-cat release, 10.2 Jaguar, but it too lacked Cocoa Finder goodness. After Jaguar, the irrational attachment to Cocoa began to wane as the Mac community largely succeeded in educating itself (and as more high-quality Carbon applications proved that API choice does not dictate software quality).

When word of an “all-new UI” for the Finder in Mac OS X 10.3 Panther began to leak out, the Finder frenzy reached a peak. “Finally, after almost three years, they’re fixing the Finder!” The Cocoa Finder meme briefly flared again, but was quashed. A new Finder is a new Finder, right? Let’s not get greedy. But as it turns out, the “new” Finder was mostly just a slightly different look for the windows, plus a revised toolbar and very basic sidebar. This was no fundamental UI overhaul, nor was it the cleansing of long-standing bugs and performance pitfalls that even the most pessimistic Mac users expected.

During the development of Mac OS X 10.4 Tiger, Finder expectations were at an all-time low. And yet the rumors of an all-new (maybe even Cocoa!) Finder still dutifully appeared in the final run-up to Tiger’s release. And again, even those faint hopes were dashed.

So here we are in early 2006, nearly five years into the history of Mac OS X, awaiting the release of Mac OS X 10.5 Leopard some time in late 2006 or 2007. Does hope for the Finder spring eternal? You bet it does. In late 2005, a single, short post at MacosXrumors re-ignited the Finder fire.

[A]nonymous sources revealed to MacosXrumors the first major feature of Leopard and it looks like it has to do with the Finder. According to the sources, Apple will entirely re-design the Finder in its next major Mac OS X update.

From this one seed, planted in a low-traffic corner of the Mac web in October 2005, hundreds of tiny flowers bloomed. Reports of an “all-new Finder” appeared on other rumors sites, and then Mac news sites as well. Eventually, the echoes were so many and so distant that they appeared to be confirmations of the original rumor report.

And yet as far as I’ve been able to determine, every Leopard Finder rumor after October 2005 can be traced back not to private, independent “sources,” but to that single MacosXrumors post. That’s a reflection of both the collective exhaustion of the Mac community on the whole Finder issue, and the indomitability of the idea, the hope, the dream, that someday, somehow, we can all love the Finder again.

Now I know what some of you are thinking. “There goes Siracusa again, harping on his stupid Finder issues. Give it up! The Finder is fine now.” Yes, the Mac OS X Finder has gotten better over the years. But (to paraphrase a recent Mac Ach post) while I’d much rather be stabbed in the eye than shot in the head, they both still suck.

Earlier this month, we learned that Apple seems to agree with that assessment…or at least tacitly acknowledges that the sentiment is widespread outside of Apple. Take a look at this job posting for a Finder Software Engineer.

The Finder team is seeking an energetic, motivated software engineer to help develop next generation versions of the Finder, the notorious file browser for Mac OS X. You will be responsible for developing new features of an application that is often perceived by our users as the “face of the system”.

Emphasis added, but hoo boy! That sure isn’t the sound of a company that’s proud of its existing product.

I was actually slightly miffed when I saw that job listing because I’ve been planning to write something about the Finder for a while now. I guess it’s my own fault for slacking on the FatBits front. Call it a post-MWSF let-down. Anyway, I’ll now briefly post what I had planned to post earlier. Just imagine that you’re reading it before seeing the job listing and it’ll seem a lot more prescient. Maybe. Starting over…

So, the Leopard Finder…could this be the release? Yeah, we’ve been burned before (four times in five years, in fact), but it stands to reason that someday the longest-suffering Mac OS X application will get some software loving. I mean, right? Doesn’t it? We friggin' switched to Intel CPUs! It’s future time! What’s it gonna take, huh? Apple PDAs? iPods with cameras? Throw me a frickin' bone here!

But seriously folks, I actually think a new Finder is coming. Yes, maybe even a Cocoa Finder. But either way, I mean “all new” as in “looks different and works differently.” As always, it’s tempting to extrapolate from the UI of whatever the most recently released Apple software product happens to be. I will now give in to that temptation and admit that when I envision Apple’s Brand New Finder™, I see an application that takes a lot of stylistic and functional cues from Aperture. (That doesn’t mean anything other than that I’m easily influenced by shiny things, but bear with me…)

Obviously, in my vision, I also see an end to the stubborn bugs and sluggishness that have long plagued the Mac OS X Finder. In that area, perhaps the Aperture connection is a bit less favorable. For now, let’s just assume that things will actually work correctly and be reasonably fast and beachball-free.

So, this dark, sleek, Aperture-esque Finder with translucent black-tinted information palettes and metadata-powered auto-stacking features…will this finally fix the Finder? Will this be the Finder of my dreams? Sadly, these are two different questions. And here lies the ultimate tragedy inherent in even my Leopard Finder fantasies.

I believe that the Finder will (eventually) rise, phoenix-like, from the PowerPlant ashes1 of its shameful past. Furthermore, I applaud any such effort in service of a Brave New plan. At this point, the Finder desperately needs some sort of vision—any vision— to guide it. Unfortunately, my guess is that the vision behind the new Mac OS X Finder will have at its core the concept of “file browsing.”

Don’t get me wrong, I can appreciate a good file browser as much as the next guy—and the current Mac OS X Finder sure as hell isn’t it. So yeah, a better browser implementation will be great. It’s just that I think there’s more to file management than browsing.

A browser-mode-only Finder built in service of a coherent vision will surpass the current Finder by light years. It’ll do a little happy dance on the old Finder’s head. But it will likely make my daily Finder usage less satisfying due to the total removal of even a poor approximation of spatial features.

This is my fantasy, and this is my fear: that the Finder will be reborn, and it will be like The Omen. But I’m willing to take one for the team. If Apple wants to go full-bore browser-only, and does so with gusto, fielding a Finder with a UI that’s at least as daring as Aperture’s, I will thank the Gods that they finally did something. And I will try my best to rejoice in the glow of screenshots showing a beautiful interface that will make me less efficient, but will perhaps make others happy anyway.

Okay, back to the present day. So, in light of the Finder job posting, how’d I do? It looks like my prediction of possible Cocoa goodness, while an adorable nod to an ill-founded meme of the past, is soundly knocked down by the list of requirements for the new Finder team member:

Excellent knowledge of C++ …
Experienced in using STL, Boost …
Knowlege[sic] of Core Graphics, HIView and Carbon, Core Foundation

Okay, so, Carbon it is. How about my dire browser-only “spatial doomsday” scenario? Well, the job listing does describe the Finder as “the notorious file browser for Mac OS X.” It also says, “You will be working on user interfaces spanning various browser views…” That’s still not a confirmation of my worst fears, but it certainly doesn’t put me at ease.

And speaking of fears, there’s some more worrisome language in the job description. It says that the candidate will “participate in all of the various stages of feature development from design brainstorms, through feature development” and lists “experience with various phases of the UI design process” as a requirement.

Now far be it from me to impugn the very existence of the elusive software renaissance man (or woman), but I still find it troubling that Apple is trying to get the C++/STL/Carbon person to help design the user interface as well. If you find That Person, then great. But a programmer designing a really great user interface is the exception, not the rule (yes, even among Mac developers). A job listing like this really makes me wonder what the software design process is like inside Apple these days.

Finally, what does this all say about the timeline or scope of the Finder project? I predicted big things for the Finder in Leopard, but if Apple is still looking for people—people who will apparently be in on the initial design process—it does not bode well for a “big bang” Finder update in Leopard. Either the new Finder will be “mildly improved (again)” or it’ll appear in Mac OS X 10.6 Ocelot2 in 2008 or 2009. I’m not sure which possibility worse.

Ah, but the beauty of a Jobs-run Apple is that there’s always the possibility that any and all predictions are completely wrong. I think we’ll find out more at WWDC 2006. My Finders are crossed.

  1. Yes, I know the Finder shed PowerPlant in Tiger. No email, please.

  2. I made that up.

This article originally appeared at Ars Technica. It is reproduced here with permission.