Open is bad?

Once in a while, I enjoy columns about technology from people outside the technology industry. James Fallows, for example, doesn't suffer from all the unstated assumptions and personal agendas that many inside the industry have, so he regularly makes very useful observations from the "civilian" perspective.

On the other hand, Farhad Manjoo's column in Slate today really misses the mark. Here's his argument in a nutshell:

  • Apple has been arbitrary in deciding which iPhone applications are available through the iTunes store.
  • Google is being more open about applications written for their new Android platform, which will extend to the Google phone.
  • The more open you are, the harder it is to deal with variations in hardware and software. (A common term for these permutations is the "certification matrix.")
  • On balance, consumers benefit more from vendors like Apple, who are more protective of the limited options they provide, than Google, who opens the door to a lot of poor quality applications that run into problems because the certification matrix is just too big and complex to handle effectively.

I've had responsibility for the certification matrix. It's a pain. Worse, it's a source of a lot of unexpected problems that distract you from other projects while you rush to fix them, so that you don't lose any more good will from already frustrated customers. You can try to reduce the chaos through Jesuitical distinctions between what's supported and what's certified, but you'll never eliminate these headaches completely.

However, Manjoo is mixing two different parts of Apple's business, when it's important to treat them as one as separate policy decisions. Narrowing the range of software and hardware configurations you support saves you and your customers a lot of headaches. If you provide the hardware itself, you're in the best of all possible worlds, certification matrix-wise, since you don't have to worry about some other company building unexpected incompatibilities into new devices.

On the other hand, you can limit the certified hardware/software configurations without restricting the market for add-ons. For example, Firefox has won a loyal fan base through openness. Using the Firefox APIs, developers have built a wide range of plug-ins that do everything from blocking ads to providing instant translations of foreign words. As shiny and new as Google's browser, Chrome, might be, Firefox users aren't going to abandon all these useful additions just to get a new browser.

In this case, openness wins hands down, even though it has its cost. Some of the plug-ins are pretty useless, and more than a few have nasty bugs. Mozilla might have built into Firefox a way to prevent low-quality plug-ins from working without Mozilla's permission. Mozilla opted instead for the other side of openness, consumer information. The catalog on Mozilla's site includes ratings and comments from other users, so you're not exactly downloading in the dark.

The iTunes site already has ratings and comments. So why then does Apple need to exert tight control over what people sell, if it already has tight control over the technology on which it's based? Apple might have perfectly justifiable reasons, but the decision to block a podcasting widget because it's "redundant" has nothing to do with the certification matrix.


re: Open is bad?

Well, but as great as it is, Firefox is also completely bloated and unstable. Even its fans -- of which I'm one -- concede this. And if you ever have chronic stability problems with Firefox, what do you do? You start turning off plugins, because they're a chief source of problems.That's my point: Openness has advantages (plug-ins are awesome). It also has costs (you need to restart Firefox at least once a day, no?)

re: Open is bad?

Thanks for the response, Farhad. However, I have to respectfully disagree, since you've taken Firefox stability issues a little out of context.Yes, if you promiscuously add a lot of Firefox plug-ins, you run the risk of burdening the browser with something that wolfs down memory and CPU like popcorn. But I've also had to tweak Firefox, sans plug-ins, to put a ceiling on how much memory it consumes (especially if I keep a lot of tabs open). And plug-ins are hardly the only source of JavaScript memory leaks and other things that make your browser, Firefox or otherwise, go kersplat.On the other hand, Internet Explorer, which is less open than Firefox, suffers from more instability and other problems on every PC I've used. Some of these defects spring from deliberate design decisions, such as the unfortunate habit of IE to open an Office document in another browser window if you click on a link to it. This feature definitely eats up a lot of resources, since you're opening an Office application on top of IE. (Plus, many users can't quite figure out if you can edit the browser-delivered document or not.)IE has other sources of exactly the sort of problems you cite. Plain vanilla IE still freezes on my work laptop. Not surprising, since after Microsoft gained browser market dominance, it had less incentive to fix bugs quickly. Given that IE's implementation of JavaScript has its own peculiar twists, Microsoft has contributed to cross-platform bugs web application bugs, which is no small problem with the ubiquity of Ajax.It's possible to imagine the iPhone suffering a similar fate, if it becomes the dominant smartphone. We've already seen Apple make some highly questionable decisions about iTunes, such the flap over bundling Safari with iTunes upgrades. If iTunes faced any tough competition, I'd be surprised to see Apple make these missteps quite so often.Which is all a very long way of saying, yes, I do have to re-start Firefox once in a while. However, plug-ins are hardly the only source of instability, and it's doubtful they're even the primary source for most users--except, again, for the tiny minority who go nuts adding them. Since people can read about problems with a particular Firefox extension before downloading it, caveat emptor certainly applies.