This is the mail archive of the cygwin@sourceware.cygnus.com mailing list for the Cygwin project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]

Re: Linux "bazaar" model (Re: Beta-19 and configurations....)


   Are you under the impression that linux somehow sets itself up automatically
   out of the box?  It doesn't.  In the rawest form, you download sources
   from ftp.kernel.org and extract them onto your hard disk using (gasp)
   tar and gzip.  Some package package maintainers have made things easier
   with various tools (e.g., RPM).  These are third parties however and are
   not part of the "official" linux distribution system.

Linux installation is *far* more automated these days.  You can
essentially start from scratch with the CD and install a complete system.

   Frankly, if a potential developer can't handle the chore of setting up
   his/her system to build the cygwin sources I can't imagine that they would
   be able to contribute to the project in any meaningful way.

It is not the developers.  It is the set of skilled users (like myself)
who want UNIX functionality, but are far too strapped for time to wade
through loads of information to get a usable system.  I have been a unix
developer at many times in the past.  I understand that there is a lot
of things that *can* be configured, but that is not a necessity.  I say
again, try the UWIN distribution from AT&T.  This is a model that I
think would be a good one to emulate.  It installs out of the box and is
immediately useful.  It is not quite as functional as gnuwin32 has the
potential to be, but it is much more convenient.

   The release maintainers *could* take some time to set things up so that
   even the most unskilled newbie could set up his system to build the sources
   but what is the benefit of that?  I'd rather that time was spent improving
   cygwin, not diddling with scripts to make life easier for people who don't
   want to or can't understand how to set up their systems themselves.

Again, we are talking about taking the knowledge that goes into all of
the FAQs floating out there and coding it into installation scripts.
Yes, this is a lot of work.  However, if it is done properly (and Linux
does this), then you will get an order of magnitude more users.  When
looking for bugs, this can be a good thing.  Note that not all of the
people who are calling for complete installation are newbies.  Note that
outside of the working gnuwin community, gnuwin has a reputation as too
raw to be useful.  This is *NOT* a criticism of the package, merely a
criticism of the set of things provided and the difficulty (and skilled
time) involved in setting it up.

   Anyway, you're *still* not getting it.  If you think this is a good idea,
   why not take it upon yourself to set up some kind of auto-install system?
   Submit if for approval to the mailing list.

See past couple of messages about the problems involved with integrated
user-community contributions.

   As far as including editors, etc., with the package; I'm still in favor of
   the linux model.  You can download whatever you want from the net.  Great
   effort should be taken to ensure that UNIX applications compile "out of the
   box" under CYGWIN.  Then, if you need an editor, download the sources and
   compile it.

You are not getting it.  Why not have a complete, usable release?  You
are still caught in the "GURU" unix mode.  You are no-doubt a top-notch
hacker (probably make 3 times my professorly salary).  However, it is
this kind of mentality (among other market forces) that have doomed the
unix community to being a serious minority, *DESPITE* superior
technology.  Sigh. 

   It would be nice for Cygnus to provide a web site with ported binaries
   but it should be no great hardship to use other sites that are
   frequently advertised on the mailing list.

Yes.  It is a serious time committment.  And if there are 12 little
configuration things that need to be done to get something working,
chances are that most people will forget 1 of them and spend hours
debugging as a result.  This is not something that I want to spend my
time on, especially since someone already figured out what is involved
in getting things working.

   All that I'm saying is that this "Why don't they just..." attitude is
   rampant on this list.  There are only a handful of contributors here.
   Mumit Khan is a contributor.  Vischne (sp?) is a contributor.  Sergey is
   a (BIG) contributor.  There are five or six other people who are contributing
   code.  Everyone else is saying "It seems to me..." and "I don't see why they
   don't..."

Take all the people who are complaining, give them an organization and
framework in which to contribute to a *COMMON* release, and you will see
an explosion of usable code.  GNUWIN's potential has grown far beyond
the couple of people who are currently contributing.  The type of
organizational work required to really make this work is not all that
fun, but if done properly, could really make GNUWIN be the flagship
unix-for-NT package.  And *THAT* would be worth the time and effort.  

--KUBI--
John Kubiatowicz,
Computer Science Department,
UC Berkeley
kubitron@cs.berkeley.edu

p.s. Now just admit it.  Wouldn't you love to be able to just *get* a
working package from someone without having to spend a day compiling and
configurating it?  Why should the compilation/configuration process be
repeated by everyone who does it?  Can we not pool our collective
resources??? 

-
For help on using this list (especially unsubscribing), send a message to
"gnu-win32-request@cygnus.com" with one line of text: "help".


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]