mailing-list for TeXmacs Users

[Nicolas Girard <address@hidden>]

On Thu, Aug 01, 2002 at 10:41:47AM +0200, David Allouche wrote:
> > Anyway this kind of interaction would be of the highest interest for
> > many other uses, such as collaborative documents
> 
> What is on your mind?
> 
> We are thinking a lot about collaborative documents. Well not quite
> right now, because we are thinking a lot about GUI issues, but
> collaborative work is an important planned feature.
>  

Well, this is a complex problematics, and there's a clear lack of
related software solutions. I see at least 5 points where TeXmacs could
be the killer app:

1) Revision history support

By "support", i mean to provide:
- interaction with an external tool to allow the user to create
  different revisions
- a way of displaying semantic differences between two revisions

The first need should be quite easy to implement. The second one is more
tricky. It requires to be able to determine the semantic differences
between 2 documents A and B. If you consider A and B as simple trees,
well I had a look to a french phd thesis which roughly said that it was
*very* difficult. I could try to find the document again if you're
interested. This is more easy -- though far from trivial -- with xml
documents. A robust and definitive tool is still lacking, but the guys
from Logilab have done a good work with XMLdiff
(http://www.logilab.org/xmldiff).

This makes me think that:
- it would be very hard to reinvent the wheel, and much more
  well-advised to get inspired from what has ever been done with
  xml documents
- it is a good thing that TeXmacs document format was so close to xml. I
  don't know if it is a good idea to switch to xml either. I fear too
  much about readability, but i guess it would make it easier to compute
  differences between two TeXmacs documents


2) Collaborative edition of documents

This is related to 1) but requires additional features. I think this
requires interaction with a server (perhaps TeXmacs-server ?) and this
is where i thought the kind of interaction would be useful.

First, in all company documents you can see a nice, flashy table titled
"Revision History" which is finally never filled... because it shouldn't
be ! This contains no new informations that people would need to
precise, but on the contrary these are informations summarizing the
evolution of the document, therefore should be automatically generated.

So in TeXmacs this should just take the following form:

  <revision history>

And to this tag should be associated a macro which precises how the
table looks like. Then the data should come from a server, or whatever
you can imagine as a concurrent versioning system.

Second need, the ability to handle and display concurrent editions. For
this i have to particular idea, i just feel it would still require
interaction wich a server.

But keep in mind that this is a huge need in many companies or public
corporations, for which only incomplete solutions exist. Microsoft Word,
to speak about the more frequently used software, offers a collaborative
system but last time i tried it, it was of very little interest. Most of
time (i saw that) people just type their modifications in another color,
and merging has to be done *by hand*. That is to say, welcome back to
the Middle Ages... but this is a reality.

3) Collaborative annotations

I'm not sure about it but i think even if the technical guts to allow
collaborative annotations are the same as 2), this is a different need
in the way as in 2) the structure of the document itself evolves,
whereas here i speak about annotations which are meta-informations
associated to parts of the document.

A possible form would be a <note> tag, though this should not be
regarded as part of the document but as a meta-information either
created by the local TeXmacs user, or given by the "server" as requested
(maybe to allow such feature in the document you could put a
<synchronize notes> tag on the top of it, someting like that). The
"server" should centralize all annotations for this document (well,
rather a complete document containing all meta-informations) and perform
synchronization as requested. So the <note> tag could imply that, once
created, it informs the "server" about its contents.

Another very interesting use: if all the above stuff was possible, it
would allow articles to be refereed, maybe this could require little
workflow management but i'm convinced people mostly want things to
reliable, simple and stupid so to my mind it would not be necessary.

4) Group scheduling

Okay: this may look ridiculous at first sight. But wait a minute. How do
you think groups keep informed about their disponibilities ? Roughly
speaking, there are two cases. First case, they bought Lotus Notes, and
it rules Second case, they try to find the secretary but she's not here,
or she doesn't have the informations needed, because people didn't give
them to her.

Why this gap ? Because if it requires the slightest effort to keep their
personal informations updated, people don't do it. So why not TeXmacs ?
As it is evolving, it could become the software many people have in
front of them. As it handles structured informations, on can imagine
that it could easily send structured informations to a server as people
write them.

So to my mind one could imagine that people write their disponibilities
in a dedicated TeXmacs document that:
- displays them in a sexy way
- sends them to a server which keeps updated a document common to the
  workgroup, or which can generate such a document on-demand.

5) Dynamic reports creation 

Know Crystal Reports or Business Objects ? These are very expensive
software which basically (i mean, this is what people need them to do in
99% of the time) perform queries on a database and create nice tables
containing the needed data. You would be terrified if you knew how these
products cost, especially regarding their final use.

This would be a very useful thing if TeXmacs could interact with a dbms
(directly or not, this is not the key point) to allow reports to be
dynamically created. I mean, this is a thing to be able to generate a
report, and it is another one to let people create it dynamically, and i
think this is what peole expect.


This is roughly what i had in mind when i was speaking about
collaborative documents. I'm also very interested with the potential use
of TeXmacs as a "personal informations-aware" companion, but i think i
could write my ideas in another thread.

Cheers,

Nicolas