In the following document the second a is much nearer placed to the V:
\starttext $V \mathbin{a} W \qquad V \mathbin{\mathbin{a}} W $ \stoptext
(in the second case an italic kern seems to be missing)
Can one avoid this side effect of nested \math-command? When building symbols it is not always easy to avoid such nestings.
On 4/30/2018 6:50 PM, Ulrike Fischer wrote:
In the following document the second a is much nearer placed to the V:
\starttext $V \mathbin{a} W \qquad V \mathbin{\mathbin{a}} W $ \stoptext
(in the second case an italic kern seems to be missing)
Can one avoid this side effect of nested \math-command? When building symbols it is not always easy to avoid such nestings.
it becomes a chicken-egg issue because one can argue that a wrapped mathbin is no longer a pure bin and so ... it's again an indication that a wrong boundingbox (width) relying on always some kind of italic to fix it is a bad approach and recursively going into a bin (or whatever) node to see if there is a nucleus chactacer would be yet another 'configureble option' ... all is possible but the more we get of that the more one needs to be aware of it and provide control over the opposite (keep in mind that such control is then a formula wide one as math lists are only processed when fully collected)
Hans
----------------------------------------------------------------- Hans Hagen | PRAGMA ADE Ridderstraat 27 | 8061 GH Hasselt | The Netherlands tel: 038 477 53 69 | www.pragma-ade.nl | www.pragma-pod.nl -----------------------------------------------------------------
Am Tue, 1 May 2018 10:12:42 +0200 schrieb Hans Hagen:
On 4/30/2018 6:50 PM, Ulrike Fischer wrote:
In the following document the second a is much nearer placed to the V:
\starttext $V \mathbin{a} W \qquad V \mathbin{\mathbin{a}} W $ \stoptext
(in the second case an italic kern seems to be missing)
Can one avoid this side effect of nested \math-command? When building symbols it is not always easy to avoid such nestings.
it becomes a chicken-egg issue because one can argue that a wrapped mathbin is no longer a pure bin and so ... it's again an indication that a wrong boundingbox (width) relying on always some kind of italic to fix it is a bad approach
Well yes, I already suspected this when I saw "italic" in the log.
What would be a better approach?
The question who started the investigation wanted to get a binary symbol "\hat{\otimes}". How does one define it so the spacing is correct? It doesn't need be the fontloader that catches all the case. It would be quite okay to insert some code that says "hey handle me like \otimes". And it would also okay to assume a "good font".
But even with cambria it doesn't work correctly in context currently. The displacement is slight (only 0.3pt) but at large zoom one can see it:
%\setupbodyfont[cambria] %without it the spacing is more pronounced \starttext $ V \otimes W $
$ V \mathbin{\hat{\otimes}} W $
$ V \Umathchar 2 0 "2297 W $
$ V \mathbin{\Umathchar 0 0 "2297} W $
$ V \mathbin{\Umathchar 2 0 "2297} W $
$ V \mathbin{\Umathaccent 0 0 "302 \Umathchar 0 0 "2297} W $ \stoptext
On 4/30/2018 6:50 PM, Ulrike Fischer wrote:
In the following document the second a is much nearer placed to the V:
\starttext $V \mathbin{a} W \qquad V \mathbin{\mathbin{a}} W $ \stoptext
(in the second case an italic kern seems to be missing)
Can one avoid this side effect of nested \math-command? When building symbols it is not always easy to avoid such nestings.
this is again a side effect of bad glyph boundingboxes
anyway, the tex engine collapses nested ord nodes so i generalized that in the upcoming engine
\mathflattenmode \numexpr 0 % off + 1 % ord + 2 % bin + 4 % rel + 8 % punct + 16 % inner \relax
so you can get rid of nested same-class (actually this is only done when no super or subscript) .. as a side effect you get a sequence of math char nodes and that will then listen to the other italics enforcing flag (for fonts with weird character boundingboxes)
i didn't test side effects but at least we can influence it (the default is of course 1 which is what tex naturally does)
Hans
----------------------------------------------------------------- Hans Hagen | PRAGMA ADE Ridderstraat 27 | 8061 GH Hasselt | The Netherlands tel: 038 477 53 69 | www.pragma-ade.nl | www.pragma-pod.nl -----------------------------------------------------------------
-
Hans Hagen -
Ulrike Fischer