cygwin and ms-sfu

cygwin and ms-sfu

[James R. Phillips]

OK, here's something unrelated to biology in general and hippos in particular.

News reports have it that ms is integrating features of sfu into their next
windows server release, leading me to suspect it may become an integral part of
longhorn-professional too.  _And_ they hired the founder of Gentoo (reportedly)
to set up  a portage-like directory of ported unix software for use with sfu.

This could really rock, esp if they make x.org one of the first ported apps. 
Is cygwin in danger of being surpassed/obviated?

The cygwin architecture is based on win32, which is necessary to support win9x,
but ms isn't worried about that.  ms-sfu is programmed without using the win32
subsystem, which just represents another layer to go through in the winnt
architecture.  Should cygwin consider an architectural change, dropping win32
support, and programming directly to the native winnt api?

And, why do I have to mention hippos?

Re: cygwin and ms-sfu

[Christopher Faylor]

On Sun, Aug 14, 2005 at 08:02:50PM -0700, James R. Phillips wrote:
>OK, here's something unrelated to biology in general and hippos in
>particular.  News reports have it that ms is integrating features of
>sfu into their next windows server release, leading me to suspect it
>may become an integral part of longhorn-professional too.  _And_ they
>hired the founder of Gentoo (reportedly) to set up a portage-like
>directory of ported unix software for use with sfu.
>
>This could really rock, esp if they make x.org one of the first ported
>apps.  Is cygwin in danger of being surpassed/obviated?

If so, c'est la vie.

>The cygwin architecture is based on win32, which is necessary to
>support win9x, but ms isn't worried about that.  ms-sfu is programmed
>without using the win32 subsystem, which just represents another layer
>to go through in the winnt architecture.  Should cygwin consider an
>architectural change, dropping win32 support, and programming directly
>to the native winnt api?

There is no architectural change needed.  Cygwin already uses the WINNT
API for a few crucial things and has for years.  But, it hardly seems
necessary to go to additional effort if Windows is going to allow easy
porting of UNIX programs.

I don't see any reason to struggle to keep cygwin alive if there is
something superior out there which is readily available.

cgf