1 of 2
DAVID SIMS: I was really intrigued by the fact that the first place Marc Andreessen went the night [Netscape] released their source code was to a Linux users group. I read his comments about using a Mozilla front end on a Linux back end. I wondered what you thought of those sorts of collaborations to build end-user apps out of open source software?
LINUS TORVALDS: I think that it’s about time that we start getting end-user apps. For historical reasons, the open source projects have mostly been technical because it has mostly been technical people who have access to the Internet. It’s only been within the last two or three years that non-technical people are seriously using the Internet at all.

Especially for Netscape, it’s somewhat going back to having somewhat free source. I mean, they came from a fairly open environment initially. And, at the same time I don’t think people should assume that you only put open and open together, and then commercial and commercial together, or proprietary. Because what’s been going on for a much longer time is companies that take part open source and part their own proprietary information, and make a product that is a combination of the both.

*
"While I really enjoy having Netscape, and I think that it’s one of the major applications for Linux, it’s also not something that I’m going to bend over backwards for."
*
So I don’t want to give the message that this is a marriage made in heaven. I think it’s a good thing, but at the same time I’m sure that Netscape, for example, wants to continue -- I’m certain that they want to continue -- on other platforms than just Linux. And I’m also certain that while I really enjoy having Netscape and I think that it’s one of the major applications for Linux, it’s also not something that I’m going to bend over backwards for. I mean, I’ve always made my Linux decisions on purely technical merit and I think it’s a good match to have Linux and Netscape, but a wrong reason.

SIMS: I think a lot of people fear the command line prompt, and an advantage to a [graphical user interface] GUI, as restrictive as it is, is the learning curve. The basic learning curve is so easy. And that’s what caught my attention about this notion of a GUI front end on a Linux back end.
TORVALDS: Within a year, Linux already had the x-windows interfaces ported to it, so it already had a GUI. But yes, that GUI tended to have more technical programs, again. So what is happening is you’re getting more of the end-user, non-technical applications on top of that GUI, and then you really do have the best of both worlds, I mean, if you ever need it. But hopefully you’ll never need it.

C O N T I N U E D . . . 2 of 2
SIMS: They talked about different revenue models -- ways to approach open source. I wonder if there is something unique about an operating system that let's people say, "yes, we can work around this and support it and still be profitable."
TORVALDS: I think that what Linux really did was that Linux closed the circle, and that Linux was the last link in a multitude of open software projects. And to some degree Linux was the last link because it was -- I wouldn’t say it was the hardest part, but operating systems tend to have problems that other programs don’t have, so to some degree it was the hardest part, and that is why it was last. And because it was the closing link, it was also the thing that got all the glory.

SIMS: And along with the glory, part of what it deserves is the credit for showing people how to make a profit off of it.
*
"When you look back, a lot of software came from shareware or an open source-ish kind of background."
*
TORVALDS: Right, but other people had done profit. I mean, Cygnus has been around for a long while. There are non-obvious open source issues. For example, if you’re a hardware company, you’re actually not making money off selling your hardware, and the software is actually a loss. It’s not something you want to do, it’s just something that you have to do in order to be able to sell your hardware.

In certain areas, the open software development makes total sense. I mean, because it hasn’t been very popular in the same areas for other reasons because people are nervous about it. And they’re looking at all these other software areas where it’s not being used. But I think this is one area where it simply will open up fairly quickly because you have printer manufacturers that really want to sell their printers, and they have these front ends to the printers that really are losses to them. So let’s make them available in open source and increase your revenues and decrease your negative outflow.

A lot of what we consider to be highly successful and very critical software, maybe it’s never been open source --it probably never has. But it’s had a period of time when it tried to do kind of the same things. So, Netscape isn’t really related to Mosaic, but Mosaic was kind of the first incarnation of this graphical front end to the Web. And Mosaic doesn’t really count as open source because it had a very strict license. When you look back, a lot of software came from shareware or an open source-ish kind of background.