On Thu, Feb 27, 2014 at 12:02 PM, Prannoy Pilligundla < prannoy.b...@gmail.com> wrote:
> ᐧ > > stefano franchi wrote: >> >> > Well, pandoc handles LaTex already (although not very successfully in >> > my tests. Management of citation, for instance leaves a lot to be >> > desired), so all it's missing is a docx "reader" to make the >> > back-conversion possible. >> > >> > That is, I understand pandoc to work this way: >> > >> > Reader module from format X produces ---> internal representation in >> > pandoc format which is used by --> writer module to produce format Y >> > >> > So we would need to add a doc writer module and possibly a lyx reader >> module. >> > I assume both would have to be written in Haskell (haven't really >> > looked carefully into this). >> > >> >> Correction after morning coffee: >> >> The needed modules would be: >> >> 1. a docx|odt reader module >> 2. (possibly) a LyX reader module >> > > I feel we need to decide on the minimum set of features to support > first.If we want the features like cross references,bibliography,Indexes > etc which are there in Latex but not Lyx,we can use pandoc but then we > need to think on how to get back from docx\odt to Lyx. Otherwise as we have > access to the parsed data of Lyx then we can go ahead in these lines and > decide on how to convert to docx|odt. > > I'd like to hear other people's opinions on this. Here is what I think: - Bibliography support with suitable styles is a must. This feature is as crucial to someone working in the Humanities, as math support is for someone working in the sciences. With the difference that scientists can often avoid conversion to Word, while Humanists just can't. - I am not so sure about indexing. Indexing only makes sense for books, really, and books have a different workflow. The index is typically prepared last, *after* formatting issues have been settled (for obvious reasons: page numbers change when you change format). That means, to me, that indexing will be done from within the tool used for final production (be it word/inDesign or latex). In other words, once we get to the indexing stage, we are out of the round trip and/or conversion loop for good. - Cross references are important too, but not so relevant, I believe, for the choice of the initial format. That is because we have the needed info to produce cross-references in Lyx already (witness the xhtml translation). - As for more complex latex-produced output (tikz, pstricks, etcetera), I really don't know, but it seems to me that it would be really really difficult to get it right, if at all possible. The only possibility would be to save these portion of the text as metadata, as Rainer originally suggested. > Before deciding to write reader modules for Lyx and docx|odt in pandoc i > feel we should think on what are the formats that we want to support in the > future and seeing whether pandoc has support for them or not.One advantage > i see is that if we write a reader module for pandoc then we can convert it > into both into docx and odt(reverse is still a problem).But since the > results are also poor in pandoc,we should think twice before choosing this > way. > I'd like to look into pandoc furher. My early results were poor indeed, but I think I haven't really managed to drive the tool properly yet, especially for bibliographies. I'll post updated conversion results soon, both for biblios and for math. > If we decide not to go with pandoc then question comes back to which > format needs to be supported first,docx or odt. If we are targeting docx,i > still feel we can modify elyxer and use some python docx library for > converting into docx and then modifying and using Rob's program for getting > back .lyx and completing the round trip > > That was my original idea too. But the bibliographic support is the outstanding issue. S. -- __________________________________________________ Stefano Franchi Associate Research Professor Department of Hispanic Studies Ph: +1 (979) 845-2125 Texas A&M University Fax: +1 (979) 845-6421 College Station, Texas, USA stef...@tamu.edu http://stefano.cleinias.org
