Speaking just for myself...and, yes, I'm an academic, a philosopher (with mathematical interests)...
I have found LyX to deliver precisely what it purports to offer: I concentrate more on my writing and less on formatting. It seems crazy in retrospect, but when I was using a traditional word processor (WordPerfect, in my case), I'd spend a ridiculously long time worrying about hyphenation, line length, and the like, and that despite the fact that it didn't make a bit of difference, since I was probably going to re-write the paragraph I was so worried about the next day. (I've spoken to other people and have found this to be a common experience.) I'd worry about formatting section headings, where page breaks occurred, etc, etc. Now I don't. I just write and let LaTeX do all the work formatting my document. (And, of course, when I'm writing logic, well, you can't beat LaTeX when it comes to typesetting formulae.) So most of the time, there is no ERT in my LyX documents at all, and the preamble contains nothing more than what's needed to get fancyhdr to do its thing. Could the formatting be tweaked? Sure. If you /want/ to worry about that kind of thing with LyX or LaTeX, you're free to do so, and I've done some of that tweaking myself from time to time: That's when you have to delve into LaTeX. But precisely because LyX isn't WYSIWYG, formatting issues don't stare you in the face while you're trying to write, and so you are free to ignore them. (I find that I can write almost as freely in LyX as I can when I'm just writing longhand.) Indeed, I find that one of the first things people have to do, when they first come to LyX, is simply to /stop worrying/ /about formatting/ and learn to focus on content. It's a common complaint among college teachers that students seem to spend more time worrying about the appearance of their papers than they do their content, and, in that same vein, I often find myself wanting to ask people who post formatting questions to this list how much it really matters whether the gap about section headings is this big or only that big and, for that matter, whether they really think they know more about typesetting than the people who produced the styles they're using. That's not to say there aren't legitimate issues that arise, and if you're self-publishing books, like Steve Litt, for example, or trying to get your thesis into the appropriate format, then you're going to have more such issues to handle. But a lot of these have been encountered by other people, and a lot of them have been solved by people who posted the resulting LaTeX packages on CTAN. And for the most common issues, very comprehensive packages (like titlesec, say) have been written to expose the internals in a comprehensible way. I strongly suspect, however, that some very large percentage of the time, people just need to /chill on the formatting/. The core to LyX's advantage, one inherited from LaTeX, is thus that form is separated from content. As you may know, that's kind of a mantra these days. In this respect, LaTeX incorporates a primitive version of the kind of /semantic/ markup that's supposed to power the "semantic web". Because the markup is semantic, it's a relatively trivial matter to convert a LaTeX document to a spoken format for the blind or to reformat the document when it gets rejected by journal A and needs to be resubmitted to journal B. Yes, of course you can do a more semantic markup with Word or OO, but how many people actually do? Creating and debugging styles in traditional sorts of word processors is every bit as labor-intensive as doing it in LaTeX, and because traditional programs make spot-formatting so easy to do, people don't put in the effort to create or even to learn how to use styles. They just format line-by-line. LyX and LaTeX purposely discourage that kind of behavior, and that is a Good Thing. But to understand what a good thing it is, you have to unlearn some bad habits. It's true that bibliography formatting is a sore spot. If you can get away with using standard styles like apalike, then it's simple, but obviously not everyone can, and natbib, in particular, has some wonderful features that could be easier to use. (That said, LyX 1.4.x is a /big/ jump forward, and thanks to the developers for that.) Jurabib is even more flexible, and it would indeed be nice to see a GUI to configure it, as it can be pretty confusing, but I strongly suspect we will see that. But here again, I think a mental adjustment is in order. Journals do of course have their own styles, but I've never once had a journal send a paper back to me insisting that I put the references in form A or form B, and the journals that are really insistent all produce their own BibTeX styles, anyway. So for most people, natbib is going to be more than sufficient. What's true, however, is that most of the focus so far has been on scientific writing, and there are some issues connected with writing in the humanities that still, it seems to me, need sorting out at the level of BibTeX styles, and it's going to take some TeX coding to do that. But I expect again that this will happen, as more people in the humanities discover the advantages of LyX and LaTeX. I'm working on one such problem myself already. Richard Heck Jeremy Wells wrote: > For some time I have been evaluating Lyx as an academic word > processor, but find it wanting in a few critical areas. > > For instance, the stated goal of Lyx is to spend more time writing, > but less time on formatting. Based on my experience, however, and from > posts to this list, a great deal of time is spent inserting LaTeX tags > into documents. In fact, my assessment is that more time is spent > making Lyx work properly than is spent in dealing with a traditional > word-processing environment, be it MS Word or OpenOffice. Moreover, a > significant time investment is required to research the format of the > tag and where to insert it, and then to debug the results. How does > this save time? > > Is the eventual goal of Lyx to "GUIfy" more of the LaTeX backend to > avoid having to delve into adding tags? Or will this tool remain > relatively marginalized, only used by those willing to undertake the > significant time overhead needed to actually do productive work? > > Judging from the number of posts to this lists, citations and > bibliographies are a major issue. There is no easy to use method > (e.g., a GUI) that can define the options for natbib, jurabib, or any > number of bibliography styles. Most importantly, customizing these > styles again requires one to write more code, yet again, instead of > engaging in the writing process. > > I suppose what I'm hoping for is someone to say 1) "no, you're wrong, > because..."; 2) "wait x number of years and we'll be there"; or 3) "if > you don't like coding, use a different tool." > > This whole thing is extremely frustrating as I can see the huge > promise that the LaTeX/Lyx system can offer, but it's awfully rough > beneath the surface. > > -Jeremy