Re: tlug: Re: Many Faces on Linux

["Manuel M. T. Chakravarty" <chak@example.com>]

"Stephen J. Turnbull" <turnbull@example.com> wrote,

> >>>>> "Manu" == Manuel M T Chakravarty <chak@example.com> writes:
> 
>     Manu> The fact that some Linux softo doesn't run on commercial
>     Manu> Unixes is not so much a problem generated by the Linux
>     Manu> people, I think, than a problem generated by the people who
>     Manu> use and produce the commercial Unixes.
> 
> It's a bad interaction, but the Linux people are writing the software
> so they're generating the problem.

Well, but you can hardly blame them for doing the hard work...

>     Manu> In the end, people having the necessary hardware and OS have
>     Manu> to do the port.  So, if there is, say, no Solaris port for
>     Manu> some software, you should blame the Solaris users (or Sun)
>     Manu> who didn't try hard enough to port it, I think.  A second
>     Manu> problem is, of course, that if the software needs some
>     Manu> support from the kernel (eg, /proc file system), then, on
>     Manu> Linux, the adept hacker can add the required support to the
>     Manu> kernel and contribute the result.  On Solaris etc you are
>     Manu> stuck, as you don't have the source.
> 
> So far, no objection.
> 
>     Manu> But you can't expect the developer on Linux to make his or
>     Manu> her life more difficult by ignoring the features that are
>     Manu> specific to Linux.
> 
> That's not the problem.  It's lack of modularity.  Every software
> development textbook in the world tells you to isolate system
> dependencies in a few modules.  The "Linux disease" aka the "DOS
> disease" is that people don't.
> 
> True, it's almost always possible (hardware permitting) to port a
> given software.  But the author can make it easy or hard by following
> good or bad coding practices.  Many Linux authors do _not_ follow the
> example of the kernel developers, and spread (typically redundant)
> Linux dependencies throughout their source.

Ok, this is of course a Bad Thing.  But this is probably
less an attitude problem as lack of programming experience,
or often lack of ``envisioning that somebody apart from the
author ever uses the programm'' (I mean you start writing
and want a job g etting done fast and, when you begin giving
it to your friends, the code is already too big for easy
restructuring).

Manuel
---------------------------------------------------------------
Next Meeting: 10 October, 12:30 Tokyo Station Yaesu central gate
Next Nomikai: 20 November, 19:30  Tengu TokyoEkiMae 03-3275-3691
---------------------------------------------------------------
Sponsor: PHT, makers of TurboLinux http://www.pht.co.jp