TUAW Interview series with Paul Kafasis: The Leopard delay - does it change anything?

The interview series with developers on Apple's delay of Mac OS X 10.5 Leopard has gone really well, with great responses from Brent Simmons and Wil Shipley already offering insight into the mind and perspective of a developer. Next up to the plate is Paul Kafasis of Rogue Amoeba, makers of such apps as Fission, AirFoil and Audio Hijack Pro. I never realized that Paul apparently does most of his coding from his beach-side manor (I kid, of course), but I nevertheless was thankful that he took the time out of his day to offer his take on Leopard's delay, backwards compatibility and whether Apple should open up the Apple TV and iPhone to 3rd party goodness. Paul offers some great thoughts on all this and more, along with some choice advise for budding developers. Read on for my interview with Paul Kafasis on whether Leopard's delay changes anything for the state of Apple.

TUAW: Many critics and general users are freaking out about the Leopard delay. They've been criticizing Apple for becoming less of a computer company and more of a gadget maker, and events like the dropping of 'Computer' from their name and now the delay of their desktop OS are fueling these complaints. Since it's a big topic, however, I wanted to open by getting your thoughts as a Mac OS X developer.

Paul: Becoming "less of a computer company" makes perfect sense as the separation between computers and electronics falls away in the real world. In 1976, a computer was a rare and very specialized thing. In 2007, in addition to your Mac, your new car is a computer (actually several computers), your TiVo is a computer, and your iPod is a computer.

As far as the delay on Leopard, my biggest issue is that they promised a release date and then missed it. That is, I'm more concerned with the failure to meet a promise than with the fact that Leopard is slipping. As Brent asked in a blog post, is 10.4 really such an awful OS to use? It works pretty well for me. Heck, for testing, I keep a machine running the current OS, and one running the current OS - 1, and I never feel compelled to update the older machine.

The pace of OS X updates was much faster than I'd have liked really, once we got past the terrible 10.0 and the mediocre 10.1, so I don't mind Leopard taking a bit longer. I only wish Apple wouldn't have made a promise they couldn't keep.

If you've already gone Leopard-only with a product or two (or were considering it), does the October delay change that decision at all? Would it be worth it - or even possible - to switch gears/backtrack to build in 10.4 support and ship sooner?

We were considering making our newest (as yet unreleased) product 10.5- only, but we ultimately did not go this route.

There are some good reasons to do it - as a new application, we'd be able to take advantage of a lot of technology only available in 10.5, and we wouldn't be bothering anyone who'd used a previous version on an older OS (as there is no "previous version"). As well, at the Leopard Tech Talks Apple put on, they were really lobbying hard for developers to go Leopard-only, promising promotional opportunities.

I found this interesting, as it seems Apple really wants to use the third party developers to help push users to new versions of the OS. This makes a lot of sense - again 10.4 is a pretty mature OS. The problem is that as a developer, you have to hitch your wagon to Apple. You need to trust Apple that they'll ship on time and that adoption rates will be as high as expected. If everything works out, this could be nice for you too, but there are no guarantees. We decided to go a safer, if less bold route, which in this instance has worked out for us.

As far as "switching back" to supporting 10.4 as well, I know that wouldn't have been viable for us, nor a good investment of time. Think back to Lotus and their work to get 1-2-3 to work with 640 KB computers. By the time they finished, 640 KB computers were gone and the market had started to move on. Shifting focus here would be a similar waste of time. By the time you finished, OS X 10.5 would be available anyway, and you'd have expended a lot of effort and lost a lot of great new stuff. It would make more sense to shift just your focus for a bit, perhaps to another application.

Also as a Mac OS X developer, what are your thoughts on the pseudo-closed nature of Apple's new gadgets? The Apple TV isn't *officially* a 3rd party device, but it appears as though Apple has left the door wide open. The iPhone, as far as we know, is still a closed device. Do you want to develop for these devices? Any product ideas you care to share?

I definitely have to disagree with your statement that the door to the AppleTV is "wide open". Is the AppleTV hackable? Definitely, far more than, say, the iPod. But for your average user? That door is a bank vault door, and they don't know the combination. Until the average user can easily (and I'd say wirelessly) install software, it's pretty well closed. The only way to do that, as far as I can see, will be security holes which are much more likely to be patched.

We'd love to develop for the AppleTV, extending it the same way we've allowed the AirPort Express to work with any audio and making it a more robust device. But right now, we can't get inside, so we really have to attack it from the other direction, through iTunes. This is less than ideal.

As far as the iPhone, I think everyone wants to be able to develop for it in some capacity. It's poised to be the next huge Apple platform (after the Mac and the iPod), and who wouldn't want to get on board?

Do you think Apple should have designed at least one or the other to be wide open to 3rd parties with a publicized plugin architecture? Should the company modify this practice with its future peripheral/gadget offerings?

I think it'd be great if they did open these devices - we'd see some amazing new functionality. So, do I think they should open them up? Absolutely. Do I expect them to, however? No, that's really not Apple's way. I hope they'll change this, and realize that going it alone (or with a couple high-profile names like Google) may work, but it's less than ideal.

From your experience with this delay, do you have any advice for budding developers in terms of building their software for one or more versions of Mac OS X? Should they put all their bytes in one basket, or support at least one previous version?

We've long done what Safari has done with supporting the current OS minus one (so currently, 10.3 and 10.4), and this has worked quite well for us. I will say that deciding to support only an OS that isn't shipping is at best a gamble, as evidenced by what's happened with Leopard. You can get access to a lot of nice new features, but your tied to Apple as to when you can release. We prefer to be more in control of our own destiny. So think about it long and hard, and ask yourself if you really need what the newest OS has to offer and if you can afford a delay if Apple doesn't meet their promises.

I also have to say I strongly disagree with Wil here. The people who upgrade their OS are also the most likely to buy new software, I certainly don't disagree there. To say that no one on an older OS will buy software, however, is just plain wrong. We've got plenty of customers still on OS X 10.3.9 and we're seeing plenty of sales from them. Our one 10.4-only application is Fission, and we've seen plenty of requests for a 10.3.9 version.

While we're not going to backport Fission to an OS that's losing users every day, plenty of users aren't on the bleeding edge. Many, many users don't upgrade instantly, or even after several months. OS X 10.3 was pretty good with OS X 10.4 being even better. Leopard does provide a lot of great new technology, but we're not going to see 100% adoption instantly.

Any closing thoughts on the Leopard delay, Apple's direction as a company or the impact of the iPhone? On the latter, be honest: is your credit card already eager and waiting? Or are you sticking with your [insert mobile phone here, if you have one]?

As for the iPhone, I'm interested largely thanks to its convergence of an iPod, a cell phone, and a half-decent camera. My biggest fear with merging devices is that the whole will be far less than the sum of its parts, but I think Apple can avoid this pitfall. I'd certainly be a whole lot more interested in it if I knew I could add on to it with third party software, like a Palm Pilot. At the same time, I've never added anything to my existing iPod nor my existing phone. I'll certainly be checking it out once it's available.

The Leopard delay stinks for a handful of developers, there's no question. But as far as the users go, everyone should take a deep breath and realize that Tiger is a pretty damned good OS. Getting Leopard a few months late isn't going to kill anyone. You'll be fine, your Mac will be fine, and Apple will be fine.

- - - - -

And there we have it; Paul hasn't quite rushed out to stock up on duct tape just yet, so I feel comfortable with letting the bomb shelter collect dust for another time. I'd like to thank Paul Kafasis for his time, as well as for sending an updated response once he saw Wil's interview, as conversation between developers just makes things more interesting.

Stay tuned for two more in this series, as interviews from Gus Mueller (of Flying Meat software) and Allan Odgaard (of Macromates) will follow very soon.