Interview with Bill Hayden.
11 July 2002, 16:23gmt, by , Senior Journalist
From the he-did-what??? department...

The BeOS Journal finally decided to release round one of the interview with Bill Hayden, in order to not delay it any further. You will remember we asked for your questions back in June, 2002, and here they are, along with a few of our own. We'll post round two when Bill has more free time on his hands to complete it, and finish up this nice, lengthy interview sometime before the middle of August. (hopefully. ;-)

T.B.J. : Thanks Bill, for taking some time out to answer a few questions about your work...

T.B.J.What are your thoughts on third party development with Cosmoe, specifically other developers who will add their own contributions to the project?

Bill : Third party development is most certainly welcome in Cosmoe. That's a good question to ask since Kurt Skauen, the author or Atheos -- the OS that Cosmoe is largely based on -- was quite adamant about NOT accepting third-party changes. In fact, his stance was one of the main reasons for Cosmoe.

kmusick asks, : "Reluctantly, I believe the decision to use a Linux kernel is a sound one. There is so much new hardware support and development from companies like IBM and Sun. However, the BeOS kernel has some interesting benefits like file system add-ons, that I am concerned that we would lose in the process. Do you see any ability to provide this level of compatibility with the existing BeOS kernel?"

Bill : I love the BeOS as much as most of the audience probably does. If I had an exact BeOS clone to offer, that would be great. Now the screen goes wavy and we move back to the real world. Realistically, the Linux kernel offers both the quickest and IMHO the best path to the future of Cosmoe. It doesn't offer all that the BeOS kernel did in some areas, but in other areas it already surpasses it.

The main issue in operating system development these days is not the GUI or the kernel, it's the drivers. Without drivers it's almost impossible to get off the ground. For anyone but die-hard hackers to give it a try, the OS needs to support most or all of the equipment they use. Linux shines here because it has driver support on par with the mainstream Oses such as Windows and MacOS. That's what gives it a fighting shot. The advantage here far outweighs the cool but rarely used BeOS kernel features.

This is probably a good time to mention that Cosmoe is not intended to be a BeOS clone. I have a high level of respect for what Be did with BeOS, and I've been greatly inspired by it, but my goal is not a feature-for-feature exact clone. But, given that my goal *is* to provide BeOS source compatibility, you will find much of what you are looking for in Cosmoe.

T.B.J. : The first developer release is intended for Linux only, what's the ETA (if possible to give one) of a release people can test running on BeOS and/or Windows?

Bill : To get Cosmoe running on BeOS or especially Windows would be a lot of effort, and it's not the direction I'm headed anyway. I suppose a motivated individual could do it, but that individual is not me. FreeBSD or Darwin would be a much more straightforward port. Cosmoe may evolve to the point that it can be compiled and used on top of several Unix-y kernels such as those.

alex asks, : "When you release Cosmoe, what support for alternate languages will it have natively?"

Bill : Support for alternate languages will largely depend on the underlying OS. Linux has quite extensive support for alternate languages. As long as developers of Cosmoe and Cosmoe applications are mindful to support this standard, Cosmoe will not hinder multiple-language support.

T.B.J. : What can the BeOS user/developer community do to support your efforts?

Bill : Well, the main thing that's needed at this point is developers to examine my code, make contributions, and generally help the project get off the ground. To be honest, 99.9% of my work since the project was announced has been support people trying to compile and use it, and not to develop Cosmoe further. Hopefully a critical mass can be reached soon so I can get back to doing the under-the-hood work that makes Cosmoe better.

Unfortunately, the BeOS user will probably be disappointed at this point. The video is slow, there are few apps, and there is a learning curve. It will take developers to fix those things before users will be interested in more than a casual way.

Zaranthos asks, : "Why not put the effort into the OpenBeOS or Glass Elevator projects? Do we really need another OS project?"

Bill : That's the beauty of Open Source. That's the failure of Open Source. Pick your viewpoint. :-) I develop Cosmoe because I enjoy doing so. I don't always agree with the direction of other projects. For example, OpenBeOS seeks to be an exact clone of BeOS. Cosmoe does not. I'm seeking usability and compatibility (albeit in a very BeOS-like way).

T.B.J. : How far do intend on taking Cosmoe? What are your plans for it?

Bill : I intent for Cosmoe to be a self-contained, installable OS one day. For now, it makes more sense to continue in the role as an alternative Linux GUI, since it lowers the bar for people to join in and help out. I can foresee it being available as both an add-on component to existing Linux distros, and as it's own distro.

lichtgestalt asks, : "To my layman's ears this seems to be an extremely ambitious project. Being more or less compatible with Windows, Linux, Mac and BeOS... Frankly, it sounds a bit monstrous to me. Are you still on your own on this project or have others joined in?"

Bill : I have already received patches for outside developers, but more people are still just trying to get things compiling and running at all. Once the bootstrap process gets smoothed out, the help should be more plentiful.

lichtgestalt asks, : "You mention that the components are under different licenses, according from where you grabbed them; GPL, LGPL, BSD... What license will you use for code that you'll do from scratch? Could other projects like OBOS use those, should they be interested?"

Bill : There probably won't be many (if any) separate projects coming directly from me for which this decision will need to be made. Here's the current state of affairs:

• Appserver - GPL,/li>
• libcosmoe - LGPL
• Apps - whatever the author decides to use (some are GPL, some BSD)

If the situation were to arise, I would use the LGPL for from-scratch projects. If you were really just interested in my political stance, then I would have to say that I prefer the LGPL and BSD-type licenses best.

The GPL is the tactical nuclear weapon of licenses -- IMHO it should only be used in the rarest of special circumstances. I don't like it's viral nature. The LGPL keeps the code open and free without the GPL's tar-baby effect.

lichtgestalt asks, : "Bill Hayden sounds a bit familiar. Did you do some development under BeOS? If yes, which apps and when and why did you leave?"

Bill : I was a BeOS developer. Any BeOS developer worth his salt know his developer number, but I've lost/forgotten mine. I think it was 387. (People not in the BeOS developer scene are probably yawning now) I was a big advocate for maintaining the PowerPC port. When that effort looked hopeless, I slowly lost interest. I always secretly (and sometimes not so secretly) longed to have control over the process to make the PowerPC port happen. In a way, now I have that. That was a good bit of the motivation for Cosmoe as well.

T.B.J. : What is your relationship with Yellowtab? Current and/or future.

Bill : It's quite funny to see the rumors that fly when people get excited about a new product. I'd never even heard of YellowTab until after I was supposedly in talks with them. :-) MacOSrumors.com also had "an exclusive peek" at Cosmoe and was quite impressed with it. If you know anything about MacOSRumors.com then you know they made that up as well.