Precedences the wrong way around

[uholzer]

According to the MKM08 paper about notations, a lower precedence means "binds stronger". (That is, if I understand that correctly, when the output precedence is smaller than the input precedence, no brackets are added.) Christoph Lange told me that the precedences are interpreted as in Prolog, where a smaller precedence means "binds stronger" as well.

JOMDoc however currently (revision 2057) does it the other way around (at least for MathML). I attached a test case which contains its own notation definitions. What JOMDoc produces is (below shown as ASCII instead of the real MathML output)

(ax) + b
a + xb

but I expect

ax + b
(a + x)b

[uholzer]

test case

[vzholudev]

Dimitar, please take care of this ticket with a high priority. Thnks

[dmisev]

Let's first check if such a change would not break something somewhere (sTeX maybe?). Michael, Christoph?

[clange]

Michael will have to answer this ultimately. AFAIK the handling in JOMDoc and sTeX is currently the same (otherwise he would have noticed a mismatch earlier); so if we change it, both sTeX and JOMDoc would have to be fixed.

[kohlhase]

I am aware of the fact that the implemented preferences are different from those of the MKM paper, but I do not consider this a problem. They are also different in that they range over all integers (I think). I am very opposed to changing the implementations of JOMDoc and sTeX, what we should rather do is (finally) make the correct specification in OMDoc1.3 that explains the whole thing.

Please do not change a running system over a simple transformation.

[clange]

Replying to kohlhase:

what we should rather do is (finally) make the correct specification in OMDoc1.3 that explains the whole thing.

Then I'd suggest that you create a ticket in the OMDoc Trac for that.

[kohlhase]

I hav e done that, see  https://trac.omdoc.org/OMDoc/ticket/1616.