Hi Andreas, >From time to time I was also thinking that the equation rendering should be >doable somehow... In the end, this usually led me to the conclusion that it is >probably not so easy. For instance, Searching for MathML rendering doesn't >give very many hits (GtkMathView seems to use Pango to render MathML). Going >the other route (via latex/mathjax) might actually be easier, but requires >quite some stuff to be installed. Also if one has a nice svg renderer and >requires latex anyways one might as well use pgfplot et al for the whole plot?
Maybe one could have a look at matplotlib to get an idea how they do this? Best, Alex. On Thursday, 18 September 2014 11:11:52 UTC+2, Andreas Lobinger wrote: > Hello colleague, > > on one of these rainy sunday afternoons i sat there and thought: Hey, this > can't be that complicated... > > Math type setting (still) seems to be some black art and an awful of > heuristics are put into code and only there. So there is no material in > algorithmic form available that could be re-implemented somewhere. I mean, > even TeX itself is available only a literate programming. > > So options i considered: > * wait until it's integrated in Pango (you will wait a long time) > * try to integrate a latex/pdflatex run to produce dvi or pdf and use that. > And find a dvi or pdf to cairo renderer. > * run mathjax.js via a java script engine to svg and use Rsvg or similat to > render on cairo. > * rewrite mathjax.js in julia (might have licensing issues). > > >