> - boot code in comments should be just temporary help and gone in > final translation
Of course, but after the first few code lines I thought, It would be much easier for a reviewer to check. It's clear that there is a long way to go and boot code in comments will disappear. > - instead of 'pretend' to SExpression it seems better to define > accessors for OutputForm and use them I've already started to compile and found lots of typos. In particular, it sees that the nopile mode is not for me, because the error messages were not telling me much. So the code will be in the usual pile form. > - in case of outputTran it is clear that it performs two separate > functions. One is convertion of interpreter trees to OutputForm. The > second is general flattening + some substitutions. Ah, interesting. Maybe you know better, what an "interpreter tree" is. Yes, maybe we can split that, but I'd rather do a 1-to-1 translation first and we do the cleanup in a second step. I don't yet have a good feel for the code and how it behaves or how it should behave. > The general transforms are useful to all formaters, while the first > transform is only useful to interpreter. I would separate those > transformations. It requies some effort to find all places where > interpreter trees are passed to mathprint, but I think it is > reasonably well localized, so we can find them and apply convertion > there. That way we will decouple OutputForm from interpreter trees. Decoupling is a good idea, but until now I wondered only why there is so much stuff like "return" or "repeat" and the like in i-output. It just looked to me as if OutputForm grew historically from a common printer of AXIOM code expressions into something that is used for mathematical expressions. > In general, it makes sense to do some cleanups during translation. > One needs to be careful, but if translation is verified > incrementally, then we should be OK. May approach is to translate function by function and all in separate (git-)commits so that in each step one could check whether the output still works. Whether my git commits end up 1-to-1 in svn-commits is not clear, but I tend to be more in favour of that because that would also record the history of the translation. Ralf -- You received this message because you are subscribed to the Google Groups "FriCAS - computer algebra system" group. To unsubscribe from this group and stop receiving emails from it, send an email to [email protected]. To post to this group, send email to [email protected]. Visit this group at http://groups.google.com/group/fricas-devel. For more options, visit https://groups.google.com/d/optout.
