Bob Tennent <[email protected]> wrote: > >|I have on two occasions asked on this list whether anybody wants to > >|help me convert M-Tx to Python. Christian Mondrup convinced me that > >|we shouldn't, as outside the Unix world people don't already have Python > >|anyway. > > That may have been the case when you asked but it isn't now: Python is > everywhere.
i know of no widnoze-based python users, but i know lots who use perl. while i can write perl moderately fluently, i wouldn't start any new project using it. > >|The objection does not apply to LuaTeX. All recent TeX distributions > >|have it, maybe at this stage only as an optional extra, but it is being > >|billed as the "next generation TeX engine". > >| > >|If we had LuaTex in 1992, musixflx could have been implemented in Lua > >| and there would be only one TeX pass. > >|If we had LuaTeX in 1996, PMX could have been implemented in Lua and > >| there would not have been pmxa and pmxb passes. > >|If we had LuaTeX in 1999, M-Tx could have been implemented in Lua and > >| there would not have been a prepmx pass. > >| > >|Now it is 2010 and we do have LuaTeX. > >| > >|We can go on as we used to: regard musixflx as cast in concrete, rely > >|on Don to keep maintaining PMX (nobody else except me, as far as I know, > >|has contributed even one line of Fortran code to it) and hope that someone > >|occasionally tweaks M-Tx to take account of some recent PMX feature (that > >|person is no longer me). > >| > >|Or we can gradually convert more and more of the functionality of these > >|packages into LuaTeX, thus taking advantage of the fact that the next > >|generation of TeX package writers will be fluent in it and will be able > >|to maintain the software. A single package luamusix.sty will do > >|everything. > >| > >|I think the choice is obvious. Don't you? > > No. Is Don willing to re-write PMX in Lua or anything else? I doubt it > and as PMX is still being developed, his cooperation is essential. indeed, co-operation is important here. > As for choosing Lua, I have my doubts. It sounds like a simple > scripting language with bindings to TeX. it's a compact scripting language which has been embedded in a new tex-like development. i don't write it fluently (yet) but am willing to accept people's assertions about its power. > Could not such bindings be written for > Python? The TeX community seems again to be doing something completely > idiosyncratic. there are perl- and python-based call-backs that some people use. jonathan fine wrote such a thing, using python, that's even slightly efficient. i wouldn't be surprised if i were told that there are others, such as ruby. luatex is different: the lua interpreter is part of the luatex executable, and there are luatex primitives that allow interaction with the interpreter. the interpreter can manage parts of the typesetting data structures, as well. this isn't a new technique (for example, the exim mailer that i use offers a build-time perl, as does the apache web server), but it's an obvious one. the choice of lua isn't as eccentric as you seem to imagine. there are research groups in my department who are developing a lua culture, in place of the perl culture that pervaded the group i worked in, in the 90s. (that's happening entirely independently of tex developments -- i just happened to stumble on something a research student was writing.) what's in the labs now may well be widespread on the street soon: it's a better thought-through language than perl, afaict (so far). robin ------------------------------- [email protected] mailing list If you want to unsubscribe or look at the archives, go to http://tug.org/mailman/listinfo/tex-music
