Summary of some discussions and conclusions at the 6th OpenMath Workshop,

recorded by

Gaston H. Gonnet

Informatik, ETH Zürich, July 27-28, 1996.

If you find any inaccuracies, please e-mail me any suggestions/corrections/additions/deletions.

Content Dictionaries

Since the meeting in Bath (Jan 1996), the concepts of context and semantic dictionaries evolved in a way of increasing overlap. At this time it was decided that these two concepts should be unified, and renamed. Five names were nominated for a vote (contexts, semantic dictionaries, context dictionaries, content dictionaries and contents ). Voting was done in two rounds, the first round selecting the names context dictionaries and content dictionaries and the second electing content dictionaries by a majority of one. So the official name is a content dictionary or CD for short.

It was agreed that a Content Dictionary is an OpenMath object. This means that a CD can be described using the syntax for OpenMath objects, and could be transmitted electronically between any OM-communicating processes.

It was also agreed, and it is a logical conclusion, that the format of the CD's can be described in a CD. This is standard in the field of databases, e.g. relational databases. The CD which describes the format and semantics of the CDs will be called Meta.

Implementation. The following people are responsible for heading the committees which will define a few of the basic CDs. These CDs will be part of the version 1.0 of OpenMath to be released before December 31th, 1996.
Meta - Gaston Gonnet
Basic - Angel Diaz
Polynomials - Mark Gaetano
LinearAlgebra - Arne Storjohann (subject to confirmation)

By August 15th all the committees should be formed, and there will be some mechanism to broadcast e-mail messages to the participants of these groups (John Abbott will handle the logistics of the e-mail groups). To be on a particular group, send a message to with a single line "subscribe openmath-xxxx" in its body, where xxxx is either "dev" or the group you want to participate.

By October 15th a draft of all the CDs should be available to the OpenMath community at large. Comments/corrections/changes will follow.

By the Dublin Workshop or very shortly thereafter but before the end of the year, the definition of the most basic CDs will be frozen to satisfy the deadlines from the EU.

It was agreed that the CD writers will make their results available as soon as possible to allow the groups who plan implementations to start working as soon as possible. It was viewed as extremely desirable to have some implementations to show to the funders by mid 1997.


Mika Seppala distributed copies of the INFO 2000 program. It is possible that we could apply for support for the OpenMath Consortium under their MIDAS prorgram.

We would like to apply to obtain support for two full-time positions and for support for the two workshops per year. Of these two positions, one would be mostly in charge of the technical development of the OpenMath definition, and the other would be mostly in charge of the dissemination and application of the idea.

Mika Seppala proposed the following tentative budget

(all in ECUs)
1 Post Doc's 2 @ 50,000 ECU/year, 3 years 300,000
2 Travel for Post Doc's 10,000 ECU/year 30,000
3 Workshops and similar meetings (promotional) 60,000
4 Overheads 78,000
Total 468,000

Tasks. Implement the present OpenMath definition, promote its use, develop it further and create an environment which makes scientific publishing in the WWW possible.

Main site. INRIA Sophia-Antipolis.

Other sites. Bath, CAN, ZIB, RISC-Linz (?), ETH, Helsinki (?).

Industrial collaboration. Springer-Verlag, Elsevier.

The deadline for submitting the proposal is September 13th, 1996. James Davenport agreed to distribute the tasks among groups by August 1st. People should notice that there is relatively little time to get this proposal written and submitted.

Computation commands in OpenMath

A somewhat lengthy and lively discussion centred around whether OpenMath should transmit objects alone, or objects and commands. It is my view that nobody disagreed with the desirability of sending commands (like evaluate, plot, edit, display, print, etc.). The disagreement comes with the fact that the semantics of some commands, most notably "evaluate" will preclude any type of easy agreement between the very different systems which aspire to be OM-compliant.

Mika Seppala pointed out that for December we are supposed to produce a specification for the transmission of mathematical objects alone, and that we should concentrate on this aspect for now. There was an informal agreement that we may offer a list of such commands without a very precise semantic attached to any of them.

Next Workshop

The next workshop, the 7th Workshop, will take place in Dublin on December 7-8, 1996 (Saturday and Sunday). Professor Richard Timoney ( will be hosting the meeting. Committees should schedule their meetings on Friday December 6th.
Last update: Mon Jul 29 16:18:22 MET DST 1996