Improving communication between GNU Emacs and XEmacs

Improving communication between GNU Emacs and XEmacs

[Michael Sperber [Mr. Preprocessor]]

Allow me to introduce myself: I am a member of the XEmacs review
board.

In a recent discussion on comp.emacs.xemacs and in private mail, Eli
Zaretskii and I were wondering how to ease communication between the
development teams of GNU Emacs and XEmacs.

Partly as a result of this discussion, I initiated the creation of a
new mailing list "xemacs-design", to which we're moving most traffic
related to the future development of XEmacs.  Here's the complete
charter:

> xemacs-design is an open list for discussing the design of
> XEmacs. This includes discussion about planned and ongoing changes to
> functionality and API changes and additions as well as requests for
> them. This is the primary channel for this kind of discussion; related
> code changes will usually not be applied until they have been
> discussed here. This does not include bug reports, which go to
> xemacs-beta.

GNU Emacs developers are very welcome on this list.  I hope this is at
least a small step forward, and hope that we can get a little bit more
cross-pollination between the Emacsen that way.

More information about the XEmacs mailing lists and the subscription
process are at

http://www.xemacs.org/Lists/index.html

-- 
Cheers =8-} Mike
Friede, Völkerverständigung und überhaupt blabla

Re: Improving communication between GNU Emacs and XEmacs

[Richard Stallman]

Since we cannot use the code in XEmacs for lack of legal papers from
the authors (and even records about who the authors were), there is
not much possibility for cooperation with you to be useful for us.  I
don't see sense in helping you design changes to code we can't use
anyway.

In order for code to be available to you and to us, it has to start
from the code in Emacs (or else from zero).  On that sort of project,
it might make sense for us to work together.

Re: Improving communication between GNU Emacs and XEmacs

[Michael Sperber [Mr. Preprocessor]]

>>>>> "RMS" == Richard Stallman <rms@gnu.org> writes:

RMS> Since we cannot use the code in XEmacs for lack of legal papers from
RMS> the authors (and even records about who the authors were), there is
RMS> not much possibility for cooperation with you to be useful for
RMS> us.

This is an unsubstantiated claim, and it doesn't get better with
repetition---many XEmacs maintainers have signed papers (including
myself), and if you have questions about authorship, you need only
ask.  I'll admit there are probably murky areas, but these are not
everywhere.  We'll try to oblige---why don't you try it out?

RMS> I don't see sense in helping you design changes to code we can't
RMS> use anyway.

Well, as a sign of goodwill, many XEmacs developers have written and
maintained code which are available for GNU Emacs as well---as vice
versa.  If we adopted your viewpoint, then you're obviously saying we
should stop doing this since this is really code we can't use anyway
either.

RMS> In order for code to be available to you and to us, it has to start
RMS> from the code in Emacs (or else from zero).  

In fact, you just asked me to assist in bringing over code I wrote
from XEmacs to GNU Emacs.  Where's the availability problem?

If you personally would like to abstain from any cooperation, that's
your right.  It's just an offer of goodwill.  I hope other GNU Emacs
developers will continue to cooperate in good spirit, as they have
done in the past.

-- 
Cheers =8-} Mike
Friede, Völkerverständigung und überhaupt blabla

Re: Improving communication between GNU Emacs and XEmacs

[Per Abrahamsen]

Richard Stallman <rms@gnu.org> writes:

> Since we cannot use the code in XEmacs for lack of legal papers from
> the authors (and even records about who the authors were), there is
> not much possibility for cooperation with you to be useful for us. 

Even without code sharing, we can share programming interfaces.  I
can't speak for "us", but that would be useful me.

> In order for code to be available to you and to us, it has to start
> from the code in Emacs (or else from zero).

Lots of code developed for XEmacs start from "zero", or from code
developed by people who have signed papers.  Some of it would probably
be more likely to be useful in Emacs, if Emacs developers gave input
to it early in the design process.

Re: Improving communication between GNU Emacs and XEmacs

[Eli Zaretskii]

> From: sperber@informatik.uni-tuebingen.de (Michael Sperber [Mr.  Preprocessor])
> Date: Thu, 11 Apr 2002 17:22:08 +0200
> 
> RMS> In order for code to be available to you and to us, it has to start
> RMS> from the code in Emacs (or else from zero).  
> 
> If you personally would like to abstain from any cooperation, that's
> your right.  It's just an offer of goodwill.  I hope other GNU Emacs
> developers will continue to cooperate in good spirit, as they have
> done in the past.

I suggest that we try the cooperative approach in some practical
matter, instead of arguing about principles.  My experience is that
it's much easier to agree about practical issues than about
principles.

Re: Improving communication between GNU Emacs and XEmacs

[Richard Stallman]

    > In order for code to be available to you and to us, it has to start
    > from the code in Emacs (or else from zero).

    Lots of code developed for XEmacs start from "zero", or from code
    developed by people who have signed papers.

In those cases, cooperation could make sense.

Re: Improving communication between GNU Emacs and XEmacs

[Richard Stallman]

    This is an unsubstantiated claim, and it doesn't get better with
    repetition---many XEmacs maintainers have signed papers (including
    myself), and if you have questions about authorship, you need only
    ask.

You proposed cooperation, so I thought we were having a civil
conversation, not an argument where people say "I dare you to prove
that!"  In the interests of cooperation I will explain the situation
in a little more detail--perhaps then you will see the difficulty we
are in.

Some XEmacs developers have signed papers, and some have not.  So we
can consider using a piece of code from XEmacs.  Sometimes the authors
are people who have signed papers (or will do so), and sometimes they
are not.

But the hardest problem is that in many cases we cannot reliably
identify all the authors of a substantial piece of code.  I wish it
were true that we could simply ask someone, but XEmacs has a history
of ten years, during which many different people maintained it, and
for much of that time without keeping records.  Even if all those
people wished to cooperate, with all the good will in the world their
memories are not up to it.  The task facing them would be to name
everyone that wrote more than 15 lines over the past N years.  The
unaided human memory can't do that task reliably.  That is why it is
important to keep records, and why we usually can't use code in XEmacs
unless its history was particularly short and simple.

There are cases, such as your package code, which are easier--where it
does seem that we can identify the authors and they have signed
papers.  For such code, we can cooperate.  I wish this were true for
all of XEmacs, but I know from various experiences over the last 6
years there are large parts for which it is not.

    RMS> I don't see sense in helping you design changes to code we can't
    RMS> use anyway.

    Well, as a sign of goodwill, many XEmacs developers have written and
    maintained code which are available for GNU Emacs as well---as vice
    versa.  If we adopted your viewpoint, then you're obviously saying we
    should stop doing this since this is really code we can't use anyway
    either.

If the XEmacs developers saw a legal difficulty in using code from
Emacs, then they would rationally reach such a conclusion.  However,
as far as I have heard, they regard all the code in Emacs as available
for their use.

Re: Improving communication between GNU Emacs and XEmacs

[Stefan Monnier]

>     This is an unsubstantiated claim, and it doesn't get better with
>     repetition---many XEmacs maintainers have signed papers (including
>     myself), and if you have questions about authorship, you need only
>     ask.
> 
> You proposed cooperation, so I thought we were having a civil
> conversation,

His answer was quite civil, I believe.  He just pointed out that
he knows about those problems tracking authorship and that his
proposal is hence just about agreeing on principle to try and
cooperate within the bounds of what is possible and to try to improve
those bounds.

I think the reason he came this way is because there is a feeling within
the XEmacs community (or so I believe) that the Emacs development has
generally been opposed to cooperation (even in the absence of any legal
obstacle).  Whether this feeling is justified or not is of course irrelevant.
All that matters is that we make it clear on both sides that we consider
cooperation as something desirable.

I think this is very desirable indeed.  I have been reading the
xemacs-beta list for a while and I hope some XEmacs developers can find
the time to follow the emacs-devel list as well.

If the willingness to cooperate is clear, then we will find ways to
cooperate in various occasions.  Those can be minor and small for a start.


	Stefan

Re: Improving communication between GNU Emacs and XEmacs

[Francesco Potorti`]

   If the willingness to cooperate is clear, then we will find ways to
   cooperate in various occasions.  Those can be minor and small for a start.

For a start, thanks to efforts by Martin Buchholz which I heartily
aided, the etags versions of Emacs and Xemacs are one and the same, and
so have been for about one year.

Yes, it was a very minor effort, and certainly a very easy and limited
one, but we did it :-)

Re: Improving communication between GNU Emacs and XEmacs

[Eli Zaretskii]

> From: "Stefan Monnier" <monnier+gnu/emacs@RUM.cs.yale.edu>
> Date: Fri, 12 Apr 2002 16:09:28 -0400
> 
> All that matters is that we make it clear on both sides that we consider
> cooperation as something desirable.
> 
> I think this is very desirable indeed.

Full agreement.

FWIW, I don't think anyone will disagree with that.  The challenge is
to be able to translate this agreement in principle into practical
cooperation.

> If the willingness to cooperate is clear, then we will find ways to
> cooperate in various occasions.  Those can be minor and small for a start.

Yes, small ventures is the way to start this, IMHO.

Re: Improving communication between GNU Emacs and XEmacs

[Michael Sperber [Mr. Preprocessor]]

>>>>> "RMS" == Richard Stallman <rms@gnu.org> writes:

RMS>     This is an unsubstantiated claim, and it doesn't get better with
RMS>     repetition---many XEmacs maintainers have signed papers (including
RMS>     myself), and if you have questions about authorship, you need only
RMS>     ask.

RMS> You proposed cooperation, so I thought we were having a civil
RMS> conversation, not an argument where people say "I dare you to prove
RMS> that!" 

I'm sorry if it sounded that way, it is not what I intended.  I was
suggesting that you identify pieces of code (like you did with the
startup-paths code) that would be good starting points for shared
development and ask the authorship question.  It is very possible we
won't be able to track it all down to the point of making you
comfortable, or that not all the authors have signed papers.  On the
other hand, it is also possible that authorship can be identified, and
that papers exist or can be obtained.  It may also be possible in some
instances to replace code with unclear authorship or lack of papers by
code with clear authorship and papers.  I'll try to help with this.

RMS> I wish it were true that we could simply ask someone, but XEmacs
RMS> has a history of ten years, during which many different people
RMS> maintained it, and for much of that time without keeping records.

Yes, but for much of that time (when Steve Baur took over, and I
suspect even before that), we have kept records.  I can say for myself
that I'll try my best to help.

-- 
Cheers =8-} Mike
Friede, Völkerverständigung und überhaupt blabla

Re: Improving communication between GNU Emacs and XEmacs

[Richard Stallman]

    His answer was quite civil, I believe.

The words practically accused me of deliberately repeating a lie.
But it seems that he didn't mean it that way, so I won't
hold it against him.  I don't think he is a native English speaker;
perhaps he did not realize what those words normally imply.

    I think the reason he came this way is because there is a feeling within
    the XEmacs community (or so I believe) that the Emacs development has
    generally been opposed to cooperation (even in the absence of any legal
    obstacle).  Whether this feeling is justified or not is of course irrelevant.

That is a very grave accusation against me, and whether it is
justified ought to be important to anyone who works with me,
just as it is to me.

Re: Improving communication between GNU Emacs and XEmacs

[Richard Stallman]

    I'm sorry if it sounded that way, it is not what I intended.

If you didn't mean it that way, then I am not upset about it.
Don't worry about it.