Linux Weekly News, Alan Cox interview (part II)

Linux Weekly News

Contact us

ChangeLog interviews Alan Cox
Maya Tamiya, driving force behind ChangeLog.net (the Japanese translation of LWN) was working on a translation of our January, 1999 Alan Cox interview. In the process she came up with more questions, which Alan was nice enough to answer. Maya has forwarded the result back to us. So, here it is: the Alan Cox interview, part II:
MT: The LWN interview was held early in January. Now, only about two months later, the business world's shift to Linux has been accelerating, and the circumstance seems to have changed since.
Would you tell us what you think about this circumstance? Specifically, you mentioned in the interview that 2000/2001 "is the time scale for big Unix vendors to begin openly switching to Linux." Do you think that is going to happen any earlier now?
AC: I think it has been interesting how loudly some of the vendors got behind Linux. I don't think it has changed anyones timescales though. It takes a finite time for their customer base to forget they were saying "We don't support Linux", "Linux is a hacker OS", and similar things so they can embrace it as serious OS.
MT: ...and, may we understand that "switching to Linux" means IBM quitting developing AIX and concentrating only on Linux, for example?
AC: I couldn't see IBM switched entirely to Linux. Linux is a very good general purpose OS. Right now there are plenty of specific things that each vendor has specialised as part of their business that Linux doesn't do. It's the question of which product they offer first when you say "I want to buy a web server", rather than which products they offer in total.
MT: The second question is about this part of your diary (http://www.linux.org.uk/diary/)
March 1st:
?> Tomorrow LinuxWorld starts and everyone can go and celebrate all the
?> binary vendor software being released for Linux, and then perhaps
?> start open source projects to make it look slow and obsolete. One
?> thing Linux and especially the current desktop projects like Gnome
?> and KDE are going to do is to weed the 'good' commercial software
?> out from the junk.
Since I am not a native English speaker, I do not have much confidence in my understanding. Would you please confirm if I am correct/wrong?
My understanding is that you think junk binary software will vanish because of the open source projects, however, if the software is 'good' enough, they will survive. Put another way, business that sell software (not support) will not completely vanish in the future. If it is 'good', such software that is used widely by ordinary people will sell in the future, even if it is not specially customized nor applied advanced technology like machine translation etc.
Am I correct? If I am wrong, would you please let me know what you meant with this part of diary?
AC: I think poor quality binary software will vanish. Anyone who is selling something that can easily be written freely to the same standard will find that it gets rewritten.
Good quality binary only software I'm sure will survive. I can see the fact its binary becoming a negative point, but not the only thing that matters. Perhaps the conversation in a board meeting in 2001 will go
"We could use Oracle for our database"
"But thats binary only, which makes it a risk"
"True but it is the only thing that scales to our data size"

MT: And, would you tell us what 'good' means?
- if it is 'good' how should it be? (features, functionally?)
- which existing binary product would you call 'good'? (if any)
AC: "Good" means 'does the job people want it to do, and does it well' - its not a very precise definition.
I'd call Quake a good binary product for example. And the Linux community loves quake, and bought lots of copies. Writing a free quake clone would have been a big task at the time it was released.