LyX, OpenOffice.org, structured authoring, and me
Po 02 května 2005
(This record is a slightly edited version of my message to lyx-users list. I would like to thank once more to the developers and all other participants on the list for their very civilized tone of the discussion.)
Sometimes around 1998+ I was working in law firm and was persuaded that Word is certainly not ideal, but probably the best what we can get and only occasionally I was dreaming about something better, which would be more helpful than just Word with its stupid styles (if you think, that Word style are best thing since sliced bread, or even better than that, then please define a style, which would have a first line intended except when it is a first paragraph after a heading style or list). Then I have learned more about HTML, structured authoring (separation of style from presentation etc.), XML, etc. and found out that (especially because I was moving out of legal field, so I did not need compatibility with the rest of the M$ world so much) what I really want is some tool which would enable me to focus on what I write and leave the presentation for sometimes later or to somebody else. After playing with some other stuff (namely lout and Amaya), I finally settled with LaTeX, mainly because of LyX (LyX supports partially Docbook/SGML, but many parts — including bibliography support — are missing), which allowed me to work on my stuff without being bothered about the underlying stuff and because truly good XML tools (Epic from ArborText) are terribly expensive and absolute overshoot for my needs (I do not want to maintain my thesis and personal letters in Oracle SQL database and paying $10k for that; thank you). Moreover, they usually do not run on Linux that well (or at all).
While working with LyX (currently I am working on my PhD thesis in sociology/criminology/law-and-society) I’ve got really addicted to BibTeX and nice support I get from LyX. Moreover, I love the simplicity of the environment. You know, writing scientific papers is actually all about sufficient number of characters put down on the paper — therefore, I am really not that interested in OLE, WordArt, included live spreadsheets, etc., but I need something where I could write a lot of text and my attention (not that strong in the first place) is not distracted by everything else (and the program is sufficiently fast in the first place). Moreover, it is possible to change a lot of the underlying programs to do more what could I do (so for example I am able to include export of Docbook/XML into LyX menus — actually it is just a small script which gets Docbook/SGML on input and then with sgml2xml utility creates Docbook/XML. Unfortunately, there is no way how to include support for some BibTeX-alike bibliographic system, so I am stuck.
While working with LyX I have unfortunately met some of its limitations and I am getting more and more frustrated, because I feel that there is actually so little what I really want (see below).
- Although there is a pretty good crowd of LyX users, there is too small pool of actual developers (BTW, I am not the one of them either, because I just cannot spend a time to learn C++ and code in it) — just some six of them in fact.
- So although LyX probably cannot be considered dead yet, its development is even slower than global warming and there are many issues from small to big which are just not addressed. Moreover, the program is in the state of perpetual rewrite and for couple of years users have been calmed down by saying, that when this cycle of complete rewrite will be finished programmers of the code will address our wishes and concerns. I guess most of users gave up on hoping that their concerns will be ever addressed, as I did. I am not saying that these rewrites are not necessary, I have really no clue about underlying programming, but I just do not feel that my concerns are addressed at all and LyX is IMHO becoming to be “programmers’ project” rather than “users’ project”, if you know what I mean.
- Moreover, one of the projects which was silently dropped and will be addressed probably sometimes around the time orange farms move from Florida to Boston (or from Spain to my native Czechia) is support for scripting of LyX. So users like myself cannot address themsevels some small issues in the program itself — for example, LyX doesn’t have Transpose Characters function (Ctrl-T in Emacs) and there is no way how to emulate that. Kind of WYHIWYH (What You Have Is What You Have — and that’s it).
- I am getting more and more tired with LaTeX itself — it is not real document format, but really just a heap of hacks how to make life at least bearable and normal LaTeX document is usually just a mess of code (environment and macro definitions), presentation stuff and some actual text. LyX document is usually much better because most of the non-document stuff can be hidden in LyX stylesheets (.layout files), but then you get to utter dependency on LyX itself, and as I said above I am getting more and more worried than LyX will soon die out completely.
- Even more so, because of OpenOffice, we cannot hope anymore for that moment of conversion, when people would finally figure out that structured authoring (and LyX) is better than WYSIWYG. If they do, they will probably end up with OOo.
- I have checked out what is actually available now in OOo and I was quite shocked, that the situation is actually much better than what I expected. Actually, I think, that in rudimentary (and not always useable) form most of the stuff I am interested is there. And what is more important, there is at least hope that most of my concerns will be addressed — either through huge crowd of programmers (I know that it is not ideal, but still there are more than six real programmers working on OOo, aren’t they?) or (and that is for me the value of my exercise with bibliography with Docbook) I can do actually a lot myself — either in StarBasic or with XSLT in export filters.
So what are my requirements?
- easy-to-use WYSIWYM (What You See Is What You Mean — LyX’s community term for something very similar to “structured authoring”) tool. No, I am not willing to go to PSGML/nXML or whatever mode of Emacs, that’s not enough.
- support for bibliography (actually, the thing I like about BibTeX is that I can happily maintain the databse just with my text editor — I have never really falled in love with the idea of dedicated bibliographic managers; which makes my excitment about MODS etc. slighlty smaller — I really will not want to maintain XML document with vim anymore). Current support of bibliography sucks royally, but I think it could be fixed to at least somehow workable state (limited to just numerical citations for example) quite easily — citations compression and inclusion of the page numbers into refernces would make OOo working for me.
- international support (I would love UTF-8, but I still survive with inputenc, fontenc, & babel and 8bit encodings). And yes, I hate that I still do not know how to persuade OOo that although my locale is cs_CZ I DO WANT to create mostly en_US documents and so I want its fields to generate things like “Bibliography” and not “Seznam použité literatury”.
- Stable document format — I have six years of work in LyX and LaTeX and I do not want to use for my long-term storage any format which will change in next five years six times. Which is the reason why I am interested in Docbook and slightly cautious about new OOo ISO document format — the format which is actually not shipped yet in any real application is not what I am hoping for. However, export to Docbook would have to be make covering all my document needs.
Actually the last point illustrates some of my other frustration — it is painful to see how many projects from many different sides are for so long time really close my ideals and none of them really hits the bull’s eye. I am mostly working in sociology, so I really do not need that much advanced equations, complicated graphs, etc. and other complicated stuff which is usually associated with “document format for scholarly authoring” Probably the only four things I would need above HTML (which I consider to be minimal format for this purpose) is:
- bibliographical citations
- table of contents
- generated cross references
Some of these could be addressed with some simple namespaces over XHTML, but there is just no-one who would provide complete set of tools and GUI WYSIWYM environment for it. And it is frustrating to see how there are many programs which are darn close — Amaya is the most obvious example for many many years (ThotEditor seems to be failed attempt to do something similar for Docbook). .
Category: computer Tagged: LyX BibTeX TeX structure authoring