[forum] Re: A Gov Application Developer's Perspective - X Windows future?
- typo correction
Mon, 24 Mar 2003 08:03:07 -0500
Jim Schatzman wrote:
> Sorry about that - I meant to say that the Cygwin port of "XFree86" is incredibly slow.
> XWin32 is a commercial X server for Windows that in our tests is one of the fastest.
There are many reasons for this, I'll try to go into a few:
1) There are maybe 2-3 primary developers for the
Cygwin/XFree86 port who are doing this on their spare time.
These people are amoung the brightest out there, but they
have to work for a living, too. They have absolutely no
sponsership whatsoever. Thus, development has been slow.
DirectX/GDI functionality already exists, but it isn't
2) Cygwin still lacks a complete IPC implimentation which is
also optimized. Again, no one has stepped up to solve this
problem. Also, threading is ok, but could be further
optimized. Frankly, vfork is in dire need of a complete
overhaul, but that's just my opinion. There are only 3
part-time developers who work on this aspect, w/o sponsership.
3) Cygwin could probably do with a revamp of its mm system,
to increase the speed.
As a tax paying citizen of the U.S., I don't see why the
Feds can't divert a miniscule amount of that obscenely huge
defense budget they just passed to sponsoring (full time)
3-4 of the Cygwin & Cygwin/XFree86 developers to solve this
issue. Even if you'd only sponser 2, certainly that'd be a
fraction of the cost compared to how much it costs to
site-license and maintain Starnet X-Win. I don't know how
much they charge the Government, but $225/seat is highway
Redhat is still contracting for Cygwin development & Harold
Hunt would be the one you'd want to talk to about the
XFree86 port. I'm sure you are aware of the websites for
these projects (www.cygwin.com), so you know where to go for
Again, I stress the fact that unless someone steps up and
puts their money where their mouth is, these projects will
continue at the current pace. The nice thing about
sponsership is that YOU decide what features and
functionality YOU want to see.
At the very least, as a developer, you could contribute to
making the project better. Perhaps you could introduce it
as an official project for you to work on for your company?
Also, starting work on a completely windows native XFree86
from scratch would be highly inefficient as it is a partial
duplication of the current work being done for
Cygwin/XFree86. I really doubt that anyone out there would
want to tackle this as Cygwin/XFree86 is pretty close to
having a fully implimented rootless mode (which makes the
other port irrelevant).
I've made some assumptions and a few generalizations, but
this is my perspective as someone who is involved in the
> Thanks for reading my comments-
> Jim Schatzman
> Northrop Grumman Mission Systems
P.S. - All my comments were meant to be neutral and as such
do not reflect my political views.