Eye2Eye: Larry Augustin Pt III

As a programmer, where do you stand on the discussion as to the comparative merits of Perl and Python?

That's the joke of the day. The reason why is that Eric Raymond has started rewriting the Linux Kernel configuration scripts. Everyone goes, "oh great, they're terrible," and then he said he's writing it in Python and everyone said, "oh, no." That ignited a big thread on Slashdot a couple of days ago about Perl versus Python.

Which do you prefer?

I've never had the time to learn Python at all although I use Perl regularly. Actually I did programming language and programming language design so that's exactly the kind of question I should know how to answer. But I don't know. I know that it's white space sensitive and that just scares the heck out of me. The Python people all argue you do it anyway so it's good. You might as well take advantage of the structure in the parser, and force the indentation. I can't quite get it through my head, beceause I've spent years learning to undo structures and to write in any free form style. And I'm a compiler writer and so I think, "Oh God, I have to count white space, I can't just throw it away? That makes the parser and the lexical analyser ugly. I don't even want to write a compiler for this language now."

In your keynote you also said that open source is inevitable. What do you see as the future for those companies that refuse to get involved with open source?

Imagine you're in the software business today. In addition to worrying about your competitors, you have to worry that some college student is going to write a piece of software that does exactly what you base your business on, release it for free. Fifteen other people are going to pick it up and before you know it, there's going to be a group of one hundred people on the Internet working together to build something that's better than what you're building inside your company in a proprietary, closed way. What do you do when that happens? How do you react to that? People have to start worrying about the open source community getting into their business. This is the reason you saw so many vendors here. They're realising it's a pretty significant market and you cant afford to ignore that.

What's your view of Mobile Linux and the work that Transmeta is doing? Do you think that there will be a lot of mobile devices using this technology?

That wouldn't surprise me. Because with Linux, you can strip it down to a pretty small footprint and get it to do what you want, it's a great programming environment. There are all sorts of tools and applications out there for it. And there are no license fees. If you are in any business where you need an operating system, even, and embedded one of these mobile phones, start developing on Linux. It's a great platform. I would guess the usual laptop manufacturers would be interested.

Why do think there has been so much interest in Transmeta?

I think that Linus brings a lot of enthusiasm, but Transmeta are in the business of competing with Intel. That's a business that's going to get a lot of press.

Do you think they're going to be able to succeed?

Taking on Intel is always tough, but it's never impossible. The thing is that what they've done a very good job of is going after a slice of the market where Intel hasn't shown a lot of focus, which is great. It's the kind of focus you need as a start-up. If anyone can succeed, that's the kind of model to succeed.

Is VA Linux thinking of doing anything with mobile devices?

We're not planning to go there, servers are our main focus for now. We always look at different stuff. Sometimes people beat us up because we ship Intel, but the fact is that we look at a lot of stuff. We've got everything running inside the company and it's a question of market demand, price, performance and availability. We played around with Alpha for a while and there are just a lot fewer people running Linux on Alpha chips.

Do you think that Itanium is a great chance for Linux in the server market?

We really see that Linux is scaling with Intel. Yes, it runs on many different processors, but if the most common Intel platform is four processors, Linux will run well on it. If the most common Intel platform is a 16 way 64-bit Itanium, then Linux will scale up and run on that. I think Itanium is where Linux will really begin to look at scaling above eight processors. I'd like to see a lot more chip vendors and hardware vendors working with developers to make sure that there's support there under Linux very early on in the cycle, so that when they announce it, there's Linux support. Think about it. We're doing development now on a port for which the hardware is still in beta test. We're going to have a great port of Linux by the time its final, which should be later this year.

Thanks for talking to ZDNet News

Go back to part one

Go back to part two

Has the Linux bubble burst? And if it has, is that necessarily a bad thing, or simply a sign of maturity? Regardless of the rhetoric, Mary Jo Foley believes there is enough promise in the basic concept that software is best developed via a cooperative, rather than a competitive model. Go to AnchorDesk UK for the news comment.