@DavidCarlisle -- the book class is badly designed in a number of ways, but people keep using it, and basing other classes on it, anyway. amsbook tried to take all of the usual ams requirements into consideration, sticking as close to the user interface of book as possible, but there are always exceptions that nobody thought of. so \addtocontents is a necessary evil. besides, it's in lamport, so who's to say it's "wrong"? (lamport certainly doesn't.)
@egreg I think I am misunderstanding noclearpage from imakeidx. I thought this was used to prevent indexes from starting on a new page, but apparently that's not the case or I am misusing it. Could you take a look (if you have time) to a little example document?
@GonzaloMedina: Hi Gonzalo! Hypothetical question: how hard would be to make a graphic progressively go from bottom left to top right in a beamer document based on the page number? :)
@barbarabeeton It's wrong in the same way that \vspace is wrong, Almost always if they are used mid document it means the document author is fighting the document class design decisions. It's OK to have that fight but it should be in the preamble, not mid document. \addtocontents isn't wrong in general: it's the way headings add their content lines, but if the headings aren't adding the right space the thing to do should be to fix that not just bolt on extra space in the document.
@PauloCereda And you want the image in the background and uncovering as your presentation advances? Cool! I might "borrow" the idea (with some other image) :)
@egreg -- well that's certainly bad, but i'll counter with the glomming of all author info into \author{...}rather than segmenting it nicely, identifying each bit of info separately for each separate author, and then putting it together properly according to what a publisher needs and wants. the top matter is really a mess.
@DavidCarlisle -- okay, i won't disagree that style decisions are better made in the preamble. especially spacing of the sort being discussed. but can you provide a reasonable automatic method for avoiding a page break between a chapter listing and its associated sections? or between the fifth and sixth (the last) sections of a chapter? those are places i've found \addtocontents most useful.
my tex4ht build is starting to get too big. 5 days now to compile what is about 8,000 pages in pdf. all the time is spend on math generation to svg images. I need to find a better way to speed this up. Or buy a supercomputer.
@barbarabeeton If you never want a page break after a chapter toc line \l@chapter should add the \nobreak in all cases you shouldn't have to adjust the individual ones. Well for final copyediting as you know you may need adjustments or \enlargethispage or re-wording the title or whatever, but I see these things in first drafts not in fine tuned final publication copy:-)
@DavidCarlisle when I add a small section, I have to recompile everything again, which means I have to wait 5 days each time. At the end of this, I'll end up waiting one month for each build, since I have 20,000 more pages to add. I need to change my build makefiles to do something smarter. But tex4ht needs the whole latex file to make links.
@Nasser why are you doing it as a single latex file anyway? but even if you do you only need to convert a math equation to svg if you have not converted it before.
@DavidCarlisle my build is like this: I run Mathematica and Maple, they generate latex, then I compile them using pdflatex and tex4ht. the problem is everything goes to one latex file. so even if I change one line, have to recompile everything.
@DavidCarlisle I need to do it in one file, so that tex4ht can generate the links ! If I break the file to small separate files, then I have to make the links myself. May be I have to do this.
@Nasser it's only that way because you made it that way, but as I say even if you make it all one file then you still don't have to convert math you have previously converted
@PauloCereda You're welcome! I should have waited for the question in the main site; this is the kind of simple stuff that fools people around here and gets you a lot of rep. points.
@MartinSchröder I do not know how to parallelize things. too advanced for me. I do not think tex4ht is designed for this sort of thing. I just run tex4ht foo.tex and wait and wait....:)
@DavidCarlisle yes. I add one line, have to compile 8,000 pages again. again, only made one file to get the links does right by tex4ht. (chapter, section, subsection, toc etc...) But will looks at xr or add the links manually if xr does not work with tex4ht.
@michal.h21 it is one pdf in one link on my page. Here is the link fyi: 12000.org/my_notes/CAS_integration_tests/index.htm it is the last one in that table which is LARGE, the one which says "algberaic functions" the pdf build fast, but the tex4ht is very slow.
I am now breaking the build to small files, then have to make the toc by hand or use xr package.
@Nasser ok, fair enough. But why do you need HTML? If it's only you looking at them, and you don't want them broken up into separate files, then why isn't PDF/some other direct LaTeX output format good enough?
@michal.h21 I did not know about -p, but how will this speed things? Since it is image generation which is slow? Does the order matter? But will look into -p option and see if it will help, thanks!
@PaulGessler I do not like pdf files for browsing. Imaging if wikipedia articles were all each made of a pdf file. Would you enjoy browsing wikipedia then? Having to download each pdf file, open it using local reader, scroll down, etc... vs. just looking at html page directly. There is a role for pdf files ofcourse, but for browsing, I like to look at something in html, then download pdf if I want as an option.
@Nasser it will not speed image compilation, of course :) it is good when you want to just to make quick fix, like fixing spelling etc. you need to regenerate images in the end, of course, to be sure that everything is ok
@PaulGessler but my HTML page is not one 8,000 pages. That is the whole point of using tex4ht. It splits one large document into many separate HTML pages (by section, subsection, subsubsection, etc... as configured) and sets up the link directly to them from main toc.
So instead of making links manually to many separate pages, tex4ht does all that automatically. That is the main advantage of using it vs. manual HTML coding where one have to do all these by hand.
@egreg -- if you are going to undertake updates to imakeidx, i have a request: provide an option that doesn't clobber the page headers and footers, just formats the content. although imakeidx is way above any other multi-index processor, we have to ask authors to specify [original], otherwise the page style is irreparably damaged. (and don't ask us to revise how the page footer is formatted; it took michael downes close to a week to get it to work in the first place.)
Hello all. Has anyone used inkscape's textext extension to write math formulas in an .svg graphics file? I am having trouble inserting a \begin{cases} \end{cases} environment.... ???
@DavidCarlisle -- okay, thanks. that gives me a place to look. (how about adding the \enlargethispage from the body of the document? more attractive than manually altering the .toc file.) you sure are lucky, never having to cope with unruly authors.
@becko ok, but you can do all that Latex. You can use tikz for the picture, add Latex code there and compile all using Latex. ofcourse, you'd have to use tikz to make the pic, which might not be as easy as inkscape.
