SCO CEO Speaks Out

"The point about open source that I believe is really cool is this notion that you have thousands of eyes around the world looking at a similar problem, and obviously when you have more people focused on something, you can solve things better."
[CNET News]

This is one of the greatest misassumptions out there. The traditional open source model is horrendously inefficient. As we all know from Brook's great work, adding developers to a project only creates more scheduling problems. Perhaps this is why some open source projects never seem to get out the door, or at least get out the door very late, with less features than their commercial competition. Additionally, without a formal architecture review process, you wind up with all these 13 year old "software architects" running around and making wonderful decisions about how this or that class should work, and how this or that method should run (ever tried to figure out the format of some random linux config file?... heck, we've had a standard config format since INI's Win 3.1). The result is that the old code has to be constantly scrapped and thrown into the garbage pile so that some commercial software company can bring in their own money and their own team to attempt to do things the right way. These kinds of commercial partnerships are becoming more and more the norm for large open source projects, and it makes me really wonder some times. If the open source movement is so great and powerful, why does it need all the extra help?

4 Comments

  • I am amazed by this comment. Adding developers to a product tends to slip the schedule of a product because of difficulty communicating. The open source model doesnt depend on strong communication. It depends on passionate people making contributions that are vetted through various project leaders. There arent any Linux weekly software meetings. No monthly new linux programmer seminar. People just come out of the woodwork with contributions when they are ready. This clearly works. And it cant be measured in the same sense that traditional software efficiency models are measured.





    Open source development doesnt have to be efficient. It is done by people who want to spend time doing it. Market share doesnt matter. Nothing matters other than solving the problem domain.





    Mozilla, KDE, GNOME, Xine, Linux, BSD, Gimp. All of these are world-class software products. The entire city of Largo down the road from me has switched to free software. Do you think we'll see less of this in the future?





    As far as configuration settings. Please. This is a problem on all software platforms. INI files, the registrry, xml configs, configs in databases, binary config files, etc, etc, etc. This is a systemic problem in software.





    Companies contribute to open source when they want to see a project address a specific issue. This is the same reason individuals contribute. Not understanding why a project would need help from a company is the same thing as not wondering why it would need help from individuals.





    Anyway... The computer landscape has never looked so interesting! Competition is great!





    Cheers!





    Robert

  • Yes, many apps create their own kinds of files, however, the vast majority of them do not--especially the ones that expect you to be able to hand edit the configuration settings. Linux has no consistancy whatsoever. It would not be so bad if there was a decent UI to work with the settings (a la windows control panel), but it just aint there. Yes, KDE and Gnome and some of the distros are making good progress there (like the Mandrake Hardware Wizard), but there is still a LONG way to go.


    Yah, you aren't gaurenteed a feature just because you ask MS for it, but you aren't gaurenteed a feature just because you ask X open source group to put it in either. Both sides will put it in (hopefully) only if it is a good idea. You don't want everyone and his mom deciding what "cool" feature they want in your products.





    BitTorrent is a great little utility, but its hardly on the scale of office or windows.

  • PS: I should point out that most of the configuration files in question are the system ones. Again, the problem isn't as much the format as much as it is that you have to know the format to work with the file. If (a) the format was consistant or (b) their were some easy tools to work with all these settings, it would not be a big deal.

  • Who decides what's cool? If there's a public bugdb I can vote for it, I can provide use cases etc. With IE, I can't even give feedback. I'm deaf, dumb and blind. Not pleasant. As far as I'm concerned, the IE team might as well not exist for all the good it does me.

Comments have been disabled for this content.