> Another issue in that vein... Everyone I know who has had extensive
> experience in both "commercial" software sources and publicly available
> sources agrees that the publicly available sources tend to be higher
> quality. In commercial software, the presures described above often result
> in rushed jobs just to meet marketing deadlines. Code reviews are few
> and scant. You write code and maybe a few other people will ever see it.
> You don't need to comment it and your boss would consider too much effort
> in documenting the code to be a waste of precious time (especially when
> his impossible schedule timeline has already slipped two weeks). Freeware
> code gets viewed by EVERYBODY! You want it to be something you are
> personally proud of. Your personal reputation is on that code and sometimes
> people remember you for your coding style (or lack there of). There is
> incentive there to write good code that just does not exist in commercial
> software, and it lacks the pressures to cheat and write quick bad code.
Caldera has, as you may have heard, bought Novell (formerly DR) DOS and is
in the process of publishing the sources.
It's quite amusing (as long as you don't need to rely on it) to hear the
conditions these sources were living under.
For examples, it looks like Novell needed about half a dozen versions of
at least three different C compilers (MS, Borland, Watcom) to compile, not
counting assemblers, linkers, and so on.
They've also got the rigts to the CP/M sources, and are publishing those.
Unfortunately, most of these sorces have been lost - that is, people who
still have sources (or even binaries) of the stuff are asked to contribute
those.
Compare that with Linux. Amazing how well-organized those commercial guys
are versus Linux anarchy, isn't it?
MfG Kai