On Wed, Feb 12, 2014 at 4:43 AM, Csikos Bela <bcsikos...@freemail.hu> wrote:
> Csikos Bela <bcsikos...@freemail.hu> írta: > >stefano franchi <stefano.fran...@gmail.com> írta: > >>My next struggle with word conversions came much sooner than I > thought.Suggestions >>are welcome on how to tackle the conversion a > document with the following >>characteristics: > >>~ 16,000 wordsclass: articleengine: LuaTexBib: biblatex + biberno mathno > imagesno X->>references, branches, etc.lots of footnotes > >>In short, your standard Humanities article...Here is what I tried, with > related >>problems:1. >Lyx's own Xhtmla - does not know what to do with > biblatex, hence >>all references >are just bib keys and there is no > bibliography > > > >Did you try latex2rtf? > > > >I don't know if it works with biblatex and footnotes but for me it worked > well with bibtex. > > > >Might worth a try. > > Sorry, I did not notice you are using LuaTeX. I guess latex2rtf does not > work with it. > > By the way what can it offer that latex can not? > > Success! I was finally able to do the job with tex4ht and the ooxelatex script. ooxelatex is a script that configures tex4ht to produce output in odt format from a xelatex source. It took some hunting, because this script (and many other similar scripts) have apparently been removed from TexLive 2013 installation of tex4ht. However, the version available on the svn repository for tex4ht did the trick. (see point 4 below for tex4ht's peculiar status in TexLive) I now have an odt file---which I could easily convert to word's doc---with proper footnotes and biblatex/biber-processed bibliography. The only difference from my original setup was to switch from luatex to xetex, but that was painless enough. I only really need (Lua|Xe)Tex in order to work with unicode input sources, and either does the job. I do prefer LuaTex because of its better compatibility with microformatting when producing pdf output, but of course that is irrelevant when converting to odt or html. Lessons learned from this experience in view of a more general lyx-doc conversion project: 1. The production of bibliography (and associated in text references) require latex processing, hence the conversion must go from latex to odt/doc and not from lyx to odt/doc. This may be true for other *semantic* components of a text that require (multiple) latex processing (X-references, indices, and so on). 1.1 This means that there are really two different use-cases for a word conversion-tool, depending on whether the final product is doc or pdf. In the former case, latex processing (or a simulation of latex processing carried out from within lyx) is necessary. In the latter case is not. Several people may collaborate on a paper sending versions back and forth and roundtripping between lyx and doc (Rainer's use-case, I guess) with plenty of references, cross-references, etc, ***as long as the lyx person will produce the final pdf** (and as long as a correct system to preserve those information through the roundtrip has been devised). 2. tex4ht can preserve all the relevant information from a latex file because it lets latex itself do the processing instead of trying to parse the latex file. To be more precise, it first runs latex with a special package (tex4ht.sty) in order to produce a (modified) DVI file. Then it runs a (java) program on the DVI file to produce (x)html, odt, docbook, etcetera I wonder if a lyx-doc conversion shouldn't use the same approach, either by relying on tex4ht itself or by trying to replicate, on a much smaller scale, its approach. tex4ht is a very ambitious and therefore very complex program. Perhaps a more focused (odt only) version could avoid much of the complexity? 3. I haven't looked into the math issue. tex4ht is capable of producing MathML from latex sources, and, according to tex4ht's own website, "The OpenDocument code employs MathML for formulas, and XSL-FO for formatting." I really have no idea about the meaning of that last clause or whether an adt-MathML formula would be correctly exported to word's doc/docx format. 4. As some of you many know, tex4ht is an almost orphaned project after the sudden and unexpected death of its creator, Eitan Gurari, in 2009. Karl Berry and Radhakrishnan CV are maintaining the project, but there has been very little activity since 2009. There have been frequent updates to maintain compatibility with biblatex (which was moving very fast in those years), but little else. Indeed the official release is still Eitan's last of 2009. This peculiar situation may be worrisome for a conversion tool relying on tex4ht Cheers, Stefano -- __________________________________________________ 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
