This article originally appeared in TidBITS on 2011-07-08 at 1:29 p.m.
The permanent URL for this article is:
Include images: Off

Apple Needs Public Betas for Mac OS X

by Adam C. Engst

It’s unlikely that we’ll see a Mac OS X 10.6.9, given the nearness of Mac OS X 10.7 Lion’s release, but if Apple were to release 10.6.9, I would be forced to recommend that TidBITS readers not install it immediately. After all, both 10.6.7 and 10.6.8 introduced entirely new problems that affected large numbers of people while resolving extremely specific bugs that likely affected only a small set of the user base.

That’s unlike Apple. Although there have been exceptions, Mac OS X releases have, for the most part, been either positive steps forward or at least neutral. As a result, I’ve had little reason to recommend extreme caution when upgrading. Early upgrades — the 10.x.1 and 10.x.2 releases — generally address common bugs and problems and are thus worth getting quickly. And later upgrades — those that come after 10.x.5 — are largely aimed at mopping up highly particular problems and lurking infelicities, and seldom introduce new problems.

So what explains the font problems in 10.6.7 that forced the Snow Leopard Font Update (see “Apple Releases Snow Leopard Font Update [1],” 26 April 2011), or the printing problems in 10.6.8 that are currently best addressed by replacing four Unix applications with their previous versions (see “Mac OS X 10.6.8 Suffers Printing and Audio Problems [2],” 1 July 2011)? Are Apple’s testers just getting sloppy? Are they too focused on testing Lion to give sufficient attention to Snow Leopard? What about developers? Perhaps they too have been more focused on developer previews of 10.7 than on new versions of 10.6?

Or maybe this is part of the price of success? What if these problems are directly related to the ever-increasing popularity of the Mac and the concomitant growth of the surrounding industry of software and hardware manufacturers? And, of course, to Apple maintaining this growth over the last 10+ years, which has given us not just multiple generations of software and hardware, but an expanding number of professional users who rely on their Macs for day-to-day work.

In short, perhaps the entire Mac ecosystem has grown so large that it has exceeded Apple’s internal ability to test sufficiently against the myriad real-world configurations that have sprung from each and every Macintosh purchase. And, I would argue, even if Apple doesn’t target the Mac at the enterprise market, the consequences of introduced bugs have more of an impact on professionals than ever before, simply due to the increased number of people relying on the Mac for their livelihoods.

What then is the solution? Certainly, Apple could just employ more testers and move more cautiously with Mac OS X releases. Or we users could attempt to hold developers responsible for better testing — it’s somewhat troubling that both Parallels and PGP failed to discover and address (either in code or with communication) their conflicts with 10.6.8 before it shipped.

But neither of these solutions is satisfying. Apple can’t possibly test every conceivable user configuration, nor would we want the company to delay necessary fixes for obsessive testing. And yes, it would be nice if developers did a better job of testing too, but all developers have limited resources to devote to the task.

I propose a more modern, democratic solution: public betas of interim Mac OS X releases, starting with 10.7.1. As much as I know the very concept goes against Apple’s secretive, controlling grain, I think the company should consider it.

[Update: In fact, as I’ve learned from some of the comments on this article, Apple actually has a program somewhat along these lines. Called AppleSeed [3], it’s an invitation-only program that enables everyday users to test pre-release versions of Apple software and provide Apple with feedback. From what I can tell (and what I read in this recent Ars Technica article [4]) it’s almost entirely secret, though, so it’s unclear how it compares with what I suggest below, and the simple fact of the matter is that, just like Apple’s internal testing and developer testing, it failed to bring the problems that plagued 10.6.7 and 10.6.8 to the forefront.]

Of course, I’d hope Apple would do a better job than many other companies that do public betas. For instance:

Again, the very concept of allowing users to see betas of Mac OS X goes against Apple’s desire to control the message, but as far as I can tell, Apple isn’t enforcing its developer non-disclosure agreements at all — many Web sites are now publishing NDA-covered information about Lion with impunity. So perhaps Apple doesn’t care about technical details leaking out, and with small updates like 10.7.1, it’s not as though there would be any damage to Apple’s competitive position if someone were to discover that a particular bug has been fixed or some other problem has been introduced.

And, on the plus side, releasing public betas could significantly improve customer confidence in Apple’s ability to provide a solid operating system. It takes a long time to win trust that can be lost nearly instantly, and the last two releases of Snow Leopard have distinctly hurt Apple’s reputation.