sort out symbols for associative binary operators, e.g. union

[jbs1]

migrated from Trac, where originally posted by kohlhase on 23-Jan-2008 1:17pm

MathML allowed big operators for associative operators like union. We need a general template how to deal with these symbols.

In the phone call today, we distinguished three symbols that we will probably need. They are the same from the deep semantics but different on a representational level.

For set union, we have

union = unionset1 (the nary-associative union operator)

union-coll = big_unionset3 (from the Eindhoven CDs).

union-expr = unionMathML (the binding operator over expression with a bvar).

all of these names are just for the local discussion. The first two have the natural role "application" the latter role "binding".

For union-expr, the MathML WG tentatively suggested the name big_unionset1.

We need to decide what good names for these symbols should be, and into which content dictionaries they should go.

[jbs1]

migrated from Trac, where originally posted by kohlhase on 23-Jan-2008 1:33pm

[jbs1]

migrated from Trac, where originally posted by kohlhase on 23-Jan-2008 1:34pm

Replying to [ticket:23 kohlhase]:

We need to decide what good names for these symbols should be, and into which content dictionaries they should go.

I think they should all go into set1, we should not have more than one CD per topic in the MathML set.

[jbs1]

migrated from Trac, where originally posted by kohlhase on 23-Jan-2008 1:38pm

James Davenport suggested that we have a general lifting operator (see issue #24) for the associative operators and that we have three symbols (for all aspects), whereever the concepts are widely used in Math (e.g. evidenced by the fact that LaTeX has a \big operator), and add FMPs (DefMPs if that flies, see issue #5) that define the -coll and -exp symbols in terms of the simple ones.

[jbs1]

migrated from Trac, where originally posted by kohlhase on 24-Jan-2008 7:14am

I have been looking at product and sum some more, and there the situation gets even thornier. We have the mathml token element, which in the classification above is plus-exp, and we have the sumarith1 symbol in OpenMath, which is actually plus-coll.

So we have to decide where the simple backwards compatibility wins, MathML or OpenMath. In my mind plus-exp should be named 'sum' somehow. But it should not be sumarith1, which is already taken.

This raises the general question, what the names of the content dictionaries will be for the MathML CDs should be. Maybe we have to invent new names to make sure that the old OpenMath CDs from the MathML Group are deprecated. Maybe we should use m:arith.cd3 for the joint MathML3/OpenMath3 CD for arithmetic smbols.

I am raising the issue on the OM3 mailing list and splitting this issue out to #25.

[jbs1]

migrated from Trac, where originally posted by kohlhase on 24-Jan-2008 7:55am

So, if we can come to the conclusion to have new CD names (see #25; taking m:arith for the discussion), then I would propose to use

summ:arith for plus-exp and sumoverm:arith for plus-coll.

[jbs1]

migrated from Trac, where originally posted by kohlhase on 25-Feb-2011 12:39pm

fixed with MathML3