[om] Concatenation of lists ??

Michael Kohlhase Michael_Kohlhase at asuka.mt.cs.cmu.edu
Thu May 3 18:32:53 CEST 2001


Dear Mike,

> I agree 100% that we need to get more people involved in developing the
> CDs, however I'm not convinced that this kind of approach is the answer.
> Having said that there is no need for a single development model for
> all CDs, individual groups can adopt whatever mechanism they wish to use
> and if one proves significantly better than the others then natural
> selection ought to take over :-)

I totally agree with you, and I would agree that the development model for
OpenMath (endorsed) CDs should be different (more careful) than "private
experimental CDs". Another thing is that even with CVS we do not have to
allow write access to anybody whom we do not trust.

> Developing CDs is quite a subtle process and our experience is that you
> need to have somebody who is clearly in charge: not so much a maintainer 
> as an editor.  An apparently minor and sensible change to the semantics of
> a symbol can potentially invalidate dozens of FMPs, examples etc., and it
> is hard to imagine tools which can help spot this automatically.

I think that the automated reasoning community has the technology to
address this problem, and that it would be interesting do work with them
(us actually: I am from this community) on this.

> I'm a big advocate of CVS and in fact introduced it into NAG (where it is
> now used throughout the company), but I don't believe its a "magic bullet".
> One only has to look at the Mozilla project to see how harmful allowing it
> to define a project structure can be.  There are other technologies such
> as egroup-style message boards which could also be useful, and maybe we need
> to support such things on the openmath.org website.

With my reference to CVS I did not want to restrict the range of
technologies, CVS is just what I know best. Also, as I said before, we
should not give everybody unrestricted write access. But having a central
(CVS-like) repository that allows me to keep my local copies of CDs up to
date by just typing cvs update would be very very useful. Best, if it had a
division into the official/supported/contributed branches with (decreasing)
access restrictions. This would give us the possibility of acquiring
increasing numbers of CDs and keep people from re-inventing the wheel in
incompatible ways. One way such a repository might work is by having
repository maintainers as in the CTAN network (I agree that the term
'editor' is better for the person responsible for the CD content).

> The important point is that the role of the OpenMath Society is to check and
> then adopt CDs, but anybody can develop CDs in any way they wish.  We should
> have an open discussion about what is the best infrastructure to set up
> (which should be as flexible and informal as possible) and then let people
> get on with it.  I'm open to suggestions and in theory would be happy to
> implement them on the openmath.org site (although as this is in Florida
> and I am in Oxford its difficult, but I don't see this as a long-term
> arrangement).

I also think that we should discuss this. But most importantly, I think
that we should act on this soon. After all, we the argument of natural
selection (and mutation: we can still change any system adopt now) will
work as well. 

I think that we should just start evolution.

     Michael


--
om at openmath.org  -  general discussion on OpenMath
Post public announcements to om-announce at openmath.org
Automatic list maintenance software at majordomo at openmath.org
Mail om-owner at openmath.org for assistance with any problems



More information about the Om mailing list