Subject: Re: XFree86 4.0.2 released
From: David Dawes <dawes@XFree86.Org>
List: tech-x11
Date: 12/20/2000 12:37:19
On Wed, Dec 20, 2000 at 05:50:26PM +0900, T.SHIOZAKI wrote:
>From: David Dawes <dawes@XFree86.Org>
>Subject: XFree86 4.0.2 released
>Date: Tue, 19 Dec 2000 16:51:20 -0500
>Message-ID: <20001219165120.S874@XFree86.Org>
>> XFree86 4.0.2 is now available.
>You know that we are discussing about an important thing, that is to say,
>whether 4.0.2 should include Xutf8*, don't you?

I was aware of the discussion, yes.

>Please stop distributing it immediately.

Making demands like this does not help your arguments.

>Hiura-san, if XFree86 project neither stop distributing 4.0.2
>nor revoke Xutf8* from the CVS repository, we should plan a new
>X distribution project for UNIXen, which has the client set based
>on SMI's I18N stuffs, which has the server set derived from XFree86 DDX
>(and also incooperates non-IA32 DDX platforms), and which is more harmonious
>with some UNIX vendors, X.Org and Li18nux (and of course *BSD), shouldn't we?

Of course you're free to do that.  XFree86 is open source.

>Aside from Xutf8* being really wrong or not, I feel distrustful about
>the determining process whether 4.0.2 includes Xutf8*.
>I wouldn't put it past them to enforce even the locale-elimination
>advocated by Markus.  I'm really afraid of it.
>So, we would have no choice but to make a drastic cut in the near future,
>if we should accept Xutf8* now.

I think that many of you missed the point of Markus' proposal.  You
should read the whole message carefully rather than just the bits that
got you angry.

XFree86 is not planning to remove existing features or interfaces.
It is ridiculous to compare adding some new interfaces with removing
old ones.  One thing I can say is that we do not plan to intentionally
make changes in XFree86 that will either break the X11 protocol,
or break any of the published X-related standards.  From an API/ABI
point of view, we won't be making changes that would require bumping
the major version of shared libraries covered by published X-related
standards.  That means no semantic changes or removal of interfaces
covered by such standards.

We do from time to time add new protocol extensions and add new
interfaces.  Some of these are experimental.  As a vendor, I don't
believe that XFree86 is unique in doing this.  If you disagree with
that statement, I'd like to see it backed up with hard evidence.

We try to take backward compatibility seriously.  That issue was
probably my main reservation I personally had about including the
Xutf8 interfaces.  If it is eventually decided that there is a
better way of providing the funtionality they provide, then we'll
need to keep them around for a while for compatibility reasons.
My personal opinion is that it was worth doing it in this case.

What follows is all my personal opinion, and my not reflect that of
The XFree86 Project.

XFree86 has a choice of leading or following.  In the early days
we followed.  With the lessening relevance of an X Consortium/TOG
at around the time R6.4 was released, we had to make a choice
between doing nothing and taking the lead ourselves.  We decided
to do the latter.  The result has been a lot of new work, including
in areas outside of which XFree86 has traditionally focused itself.
With the rebirth of X.Org we've been waiting and watching to see
what would come of it.  I remember meeting with some people in
Atlanta in October 1999 about working on i18n through  To
be perfectly frank, I haven't seen anything much of that actually
get intergrated into a public release since then.  How long should
we wait (which I think was the point of Markus' message)?  X11 is
lagging behind in a number of areas.  If, and the other
vendors don't feel a sense of urgency here, and don't back it up
with results, then XFree86 is once again faced with the choice of
doing nothing or forging ahead ourselves.  Guess which choice we'll

Don't get me wrong.  I am a supporter of, and I'd like to
see succeed.  For that to happen there needs to be more vendor
cooperation in the standards process, and a more streamlined
standards process.  I think that XFree86 coming up with both real
proposals and implementations, and giving those implementations
some real world exposure can only help with this.  If other
members already have alternative solutions to problems like the
one that prompted this thread, then why aren't they being widely
used yet?  Is it for technical reasons?  Politictal reasons?
Inertia? Because they're proprietary?

XFree86 4.0.2 is released.  It's out there.  It couldn't be withdrawn
even if we wanted to (and I personally don't).

David Dawes                                      Email:
Founder/President, The XFree86 Project, Inc      Phone: +1 510 687 6857                          Fax:   +61 2 9897 3755