« first day (3181 days earlier)      last day (1795 days later) » 

4:11 AM
@DavidCarlisle Wow
2 hours later…
6:26 AM
@PauloCereda 15 Brazilian Reals? Nothing else?
4 hours later…
10:51 AM
If I have more than 1 set of grouping (catcode 1,2) pairs, let's say {} and <>, is there a canonical way to absorb an argument in such a way as to know the charcodes of the grouping tokens; that is, whether it was grouped as {}, <>, {> or <}?
My question pertains in particular to the case where I am trying to parse an input stream, so I am not free to take the group in question and arbitrarily cast it as an argument to a 2nd "decoder" macro.
11:14 AM
@StevenB.Segletes no
@DavidCarlisle Thanks for making it easy!
@StevenB.Segletes you can always (in theory) parse character by character and work things out by hand, but tex is not going to help here at all, if you actually grab the argument you can't tell if it was \fbox{x} or \fbox x the delimters are stripped at a very early stage.
@DavidCarlisle So if I grab it "token by token", I presume I would use \futurelet. But since \futurelet produces an implicit token, how to I find the charcode of such a implicit token?
@StevenB.Segletes take apart \meaning (look at bm package, it does a lot of that:-)
@DavidCarlisle Thanks for the suggestion
11:24 AM
@StevenB.Segletes but you can't just use futurelet as otherwise you can not tell \foo{zzz\egroup zz} from \foo{zzz} zz\egroup
@DavidCarlisle Not that I know how to address it, but I understand that differentiating explicit from implicit is another problem area.
@StevenB.Segletes safer(if you can grab the whole thing in advance) would be to \detokenize to get a safe list of tokens then go through character by character annotating with the catcodes
@DavidCarlisle That would sound nicer, if I could pass an arbitrary group to this theoretical detokenizing decoder, but how to grab that group from an input stream (with grouping intact) seems tough.
@StevenB.Segletes you should set the standard catcodes then do \let\catcode\undefined to stop people messing things up.
@DavidCarlisle LOL!!
11:32 AM
@StevenB.Segletes one notable aspect of the latex3 format (as it was in 1992) was exactly that all tex primitives were undefined, for exactly this reason.
@DavidCarlisle great package by the way
@PauloCereda naturally
12:03 PM
@DavidCarlisle See email
@JosephWright looking...
@DavidCarlisle @JosephWright sounds like one more problems with too much quotes.
@UlrikeFischer Yup, think so
@JosephWright --outfile="cat"-eps-converted-to.pdf we could add another unquote here or better really would be as I suggested to frank normalize away all quotes in the core code and then just add them everywhere they are needed in the back end file references, but no chance I can look at that during the day
12:19 PM
@DavidCarlisle Sure, was just trying to track down the remaining issues
@DavidCarlisle would epstopdf be actually able to handle file names with spaces?
@DavidCarlisle Possibly I will need to sort out how to use a custom format for typesetting, but on the main release branch I could just install, rebuild my system formats, then use it, so I'm not 100% sure it's worth the effort. It's a very special case.
@UlrikeFischer I don't see why not, but it needs to pass "foo bar.eps" with quotes on the outside to ghostscript/ps2pdf
@JosephWright what is so difficult in using pdflatex-dev instead of pdflatex here? It sounds more logical and less errorprone to do everything with dev-files instead of mixing dev and master.
@UlrikeFischer The typeset target doesn't build a format, it just uses the system one at present. So switching to pdflatex-dev would still need a local install and to build formats outside of l3build. I suspect this change is going to be a one-off issue, hence thinking that I can (1) force the dev stream to build using latexrelease and (b) when it comes to a final release, simply do a local install first, build my 'normal' system formats and go with that
@UlrikeFischer Steps would have to be git checkout dev, l3build install, fmtutil-sys --all, l3build ctan. Doesn't really help a lot ...
12:42 PM
@JosephWright shouldn't it be enough to run l3build install in base and then do fmtutil-sys --byfmt pdflatex-dev?
@UlrikeFischer Yes, but the problem remains: having to use a local installation to get things to work. That's not really any different from what I said I'll do for a final release. The longer-term solution would be to build the format every time typesetting is carried out. But I suspect that is overkill
@JosephWright actually I'm confused if you were just using the system files for typesetting rotex would presumably work, so you are using the system format and search path (so system epstopdf) but the graphics.sty from the dev branch?
12:58 PM
@JosephWright I'm confused. Aren't you building the ctan zip locally? Where it is done then?
@DavidCarlisle Yes
@JosephWright so it would also work to drop a locally fixed epstopdf into the support directory in the dev branch? (I think I did something similar with the filehooks package to get things working)
@DavidCarlisle That would work, yes
@UlrikeFischer Yes, but with an unmodified TL installation. The moment I need to do 'local fiddling', just adding latexrelease to a few source files seems safer than buildingmy system-wide formats. Like I say, a proper fix would be to use the format that gets used for testing
@JosephWright thinking about it sounds a bit like the typical latex work flow: always using the auxiliary files from the previous compilation ;-). So when building a new master you are using the previous latex format. It wouldn't work with luaotfload if the docu should demonstrate a new feature.
@UlrikeFischer Yes, exactly, hence saying I really have to fix it properly. Sigh
@UlrikeFischer I guess I do need to think about better support for a format target, at least internally. Probably I need to refactor parts of the build file for the kernel fist: it's got stuff that is now done better within the l3build core
1:16 PM
@JosephWright hm
1:33 PM
@barbarabeeton do you think this is worth reporting as a TexBook error? (I realise you are no longer the gatekeeper but...)
A: Question about exercise 17.16 in TeXbook

David CarlisleA \vcenter is handled as a \vbox except that a vcent atom is wrapped around the construct which finally means that the resulting box is adjusted so its height+depth are centred on the math axis. So... As far as I can see the construction of the inner vertical list acts as if it was in a vbox in ...

@PauloCereda I'm just not sure if it's worth the work
@JosephWright It is, given that I am working on it. :)
@PauloCereda First step in my mind is to alter how the build targets are done in build.lua for the kernel: they are manual, rather than the table-based approach we now have
@JosephWright I think now that we have auxtrees and the possibility to set texmfcnf, it could work to "install" everything in a sandbox-texmf instead of putting everything in the same folder.
@UlrikeFischer Yes, but I as thinking of just taking the existing code for custom formats used by check, and making it more general. Just needs to be able to know which formats to build, and where to run
2:08 PM
@JosephWright hm
2:21 PM
@PauloCereda Brand new user. No MWE, but plenty of ducks. Already knows the drill. tex.stackexchange.com/q/500324/2693
@AlanMunn LOL
2:35 PM
@PauloCereda ping
@AlanMunn pong (opening)
3:28 PM
Hi folks. I found tug.org/TUGboat/Articles/tb09-1/tb20bechtolsheim.pdf, which appears to be as detailed a guide to \expandafter as is available, at least freely.
Is this a good and sufficient coverage of the topic? There's also a reference to some TeX book the author wrote, which is apparently available in 4 volumes, but I can find out almost nothing else about it on the net.
@FaheemMitha makes it look a lot more complicated than it is, and still misses a few cases just looking at the first page
@DavidCarlisle Hmm. Is anything the author writes wrong?
Actually, I would say that one thing that TeX reliably is, is complicated.
It's like one of those math books with starts with a simple list of axioms, and 50 pages on, you no longer have any idea what is being talked about.
@DavidCarlisle Alternative recommendations welcome. I guess I could look at the TeXbook, but it's not exactly chatty. And not all books cover such topics.
@FaheemMitha yes, in the list at the bottom of column 1 top of column 2 item 1 and 3 are fine but item 2 should only have two cases the token is expandable r not expandable, he splits it up into several sub cases but for example his list of expandable primitives he just lists \expandafter and \the which misses several (\if, \jobname, \csname,.... so if he is listing the expandable primitives it should be a longer list but there is no need to list them there.
@DavidCarlisle "the token is expandable r not expandable"
Should that be "or"?
I somehow got the impression that all primitives were not expandable, and that is why they were called primitives. But I guess that's wrong.
@FaheemMitha sorry yes
@FaheemMitha lots of primitives are expandable, \expandafter for example is an expandable primitive
There are lots of examples on this site of course, eg
3:41 PM
The details of TeX are also quite easy to get wrong. Again, like maths.
Q: When to use \edef, \noexpand, and \expandafter?

Loop SpaceI'm quite happy hacking TeX macros and cobbling together bits and pieces from different style files to suit my own ends, but I have a suspicion that my resulting hacks are not quite as elegant as they could be. In particular, with regard to when to expand and when not to expand macros. A common...

I find most programming languages are a bit more predictable.
@DavidCarlisle I've seen that one, yes. But none of the answers are very detailed.
@FaheemMitha no, not really. all languages have quirks that catch the unwary even simple small languages like C, trying to work out how many * and const to add to a prototype makes \expandafter seem trivially obvious.
@FaheemMitha there really isn't a lot you can say about \expandafter it is a very simple operation. Admittedly it can be put to some surprising uses but its behaviour (and implementation) are trivial.
@DavidCarlisle C is a bit of an outlier. But for example Python is more predicable. And I found Common Lisp very predicable. A very nicely designed language, that. OTOH, R is full of nasty surprises.
@DavidCarlisle Well, I'll read through the article anyway. Maybe I'll learn something.
@FaheemMitha yes the author is generally reliable (but not noted for short descriptions)
3:46 PM
@DavidCarlisle And that's the sort of thing experts say. Others would not necessarily concur.
@FaheemMitha if you have two tokens a and b in a list ab then normally tex will expand a but given \expandafter ab it first expands b. That is all there is to it. You can be shown any number of examples as to why that might be useful, but there is nothing else you can say about the working of the command.
@DavidCarlisle Yes, I understood that much. But according to the article, it only expands b once before returning to a.
Does "expand" always mean "expand once"? It's an ambiguous word in this context.
BTW, I was wondering why there wasn't a short form for multiple \expandafters.
@FaheemMitha yes
Like \multiexpandafter3.
@FaheemMitha multiple expandafters are actually not needed so often (ceratinly more than three very rarely except for trick answers on sites like this) and with new primitives like \expanded and \unexpanded multiple \expandafter are needed even less.
3:52 PM
@DavidCarlisle I see.
It looks like \expanded is quite new.
@FaheemMitha but what would that do? \expandafter has a trivial definition and the behaviour of multiple ones just follows from normal processing, but \multiexpandafter3 would be quite hard to define most likely
@DavidCarlisle Equivalent to 3 \expandafters.
@FaheemMitha it was in a non-released pdftex around the 2000;s and has been in luatex since the start, Joseph added it to pdftex/xetex/ptex last year.
@DavidCarlisle Oh. I didn't know you were still adding primitives.
I mean, I thought all these things were frozen.
@FaheemMitha define equivalent: does \multexpandafter3 expand literally to \expandafter\expandafter\expandafter in one expansnion step?
3:55 PM
Hmm. See previous comments about TeX not being so simple.
@DavidCarlisle No idea, sorry. Perhaps.
@FaheemMitha that is is \expandafter\@gobble\multiexpandafter3 \expandafter\expandafter after two expansion steps?
Would that work, or not?
@DavidCarlisle Sorry, I don't know what \@gobble is.
@FaheemMitha you could define it that way but it isn't very tex-like to generate a list of tokens in a single expansion step like that.
@FaheemMitha \def\@gobble#1{}
@DavidCarlisle Oh
@DavidCarlisle Oh, it just eats the following token?
@FaheemMitha yes so \@gobble\expandafter\expandafter\expandafter expands to \expandafter\expandafter
3:58 PM
@DavidCarlisle \@turkey
@DavidCarlisle Would it be preferable to expand it in n-1 steps, then?
@PauloCereda \@duck
@DavidCarlisle ooh
@FaheemMitha there are a gazillion ways that you could define it so it would then take careful documentation to say how it was actually defined, whereas \expandafter\expandafter\expandafter it is easy to see what it does if you know what \expandafter does.
@DavidCarlisle Ok.
4:00 PM
To be clear, when the term "expanded" is used without qualification, does it mean one expansion? As opposed to "fully expanded"? Or is there no generally agreed upon interpretation?
BTW, what's involved in writing a new primitive? Doesn't sound like a terribly easy thing. I read Joseph's blog entry, but he doesn't go into detail.
@FaheemMitha it usually means once, except when it is clear from the context that it doesn't. That is, if you can trust the author. If you mean in general tex discussion and/or questions expansion could mean once, multiple times, non-expandable operations like assignment or anything else tex does.
@DavidCarlisle So in formal speak, it means once?
For the sake of sanity, let's confine the question to TeX books, and Tugboat articles. If that is a reasonable subset to consider.
@FaheemMitha in the case of \expanded it was a matter of adding a dozen or so lines here and there, and re-ordering some bit but the hard part in tex source is working out exactly where to make the changes, it is all one big monolithic file with lots of implicit dependencies.
@DavidCarlisle Sounds hairy. Lots of debugging.
4:18 PM
@FaheemMitha That one wasn't too bad: Heiko wrote the code, it was clear where it went. The XeTeX additions were hard
@FaheemMitha You have to put in a few 'hooks' to let TeX know about it, then you have to write the implementation. The former bit is easy enough, particularly now I've cleaned up some of the code order. The implementation difficulty depends on the desired effect.
@JosephWright already thinking about the pdf 2.0 stuff ? ;-)
@UlrikeFischer Yes
@UlrikeFischer I'd say 'medium' in terms of difficulty
@UlrikeFischer, @DavidCarlisle (@FaheemMitha) The worrying thing here is I am likely one of about half a dozen people who've actually done this at all recently, and I don't know Pascal or WEB
@JosephWright What about unexpected effects? Especially side effects?
AFAIK, it's impossible to do anything in a modular fashion in TeX. Sort of by design.
@FaheemMitha They usually come down to 'binaries don't build'
@FaheemMitha TeX's sources are ... interesting. I understand Knuth's idea, but I also see why it did not catch on.
@JosephWright I was thinking of runtime changes. Do you run everything to a TRIP test or similar?
Unit testing, basically.
I'm only aware of TRIP, which was written by Knuth himself, if I'm not mistaken.
@JosephWright Which idea?
4:24 PM
@FaheemMitha We added a few tests to make sure the new primitives still work. Because the one I was adding are strictly separate from other things, either (a) everything else still works or (b) you kill the binary (by messing up those hooks)
@JosephWright I'm not really worried.
@JosephWright I see. It builds using GCC, right?
@FaheemMitha Having the code written to produce version for typesetting, ordered by concepts not by the needs of the compiler. There are lots of 'bits' stacked together. Problem is, it really only has a benefit when you actually want to typeset the code and read in that way.
@UlrikeFischer Er, :)
@FaheemMitha Erm, probably: I just used the automated set up from the TL GitHub mirror
@JosephWright I'm not sure I entirely follow this, but ok.
@FaheemMitha Think about a typical TeX source. The lines are ordered based on the needs of the computer. Say for example you had five concepts, and it was easier to write about (1) and (3) before (2), but in the output you wanted (1), (2), (3).
@FaheemMitha In a single source file, you can't (easily) do that. WEB makes re-ordering bits easy, but the source is then 'out of order'.
@FaheemMitha Modern software sources are written in an order to 'help the computer': you wouldn't normally expect to be having code which is essentially patched in-place by the compiler.
@FaheemMitha Read tex.web :)
4:31 PM
Ordered by concepts means that Knuth covers things in whatever order he feels like it?
@FaheemMitha more or less, essentially it's the source code for the book tex-the-program, with annotations to allow the actual compilable source to be extracted. "literate programming"
@FaheemMitha In the order he wanted to deal with concepts. So for example if you need some init code for routine B, but for reasons it has to come before routine A when actually making the binary, you write the source with A, init-B, B, then it gets re-arranged to init-B, A, B when actually building the binary.
@DavidCarlisle @JosephWright I see.
@JosephWright oh what a tangled web we weave
@DavidCarlisle DocStrip when over-done goes the same way ...
4:35 PM
@DavidCarlisle I see what you did there...
@JosephWright by documented inspiration
@AlanMunn actually it was some bloke called Knuth that did it, not me:-)
@DavidCarlisle Well, I was thinking backend code, but yeah
@AlanMunn Also Pink Floyd's Dogs of war reference, for the sake of Great Britain
4:50 PM
@DavidCarlisle, @UlrikeFischer Think I've spotted an easier way to deal with the custom-format business: I just have to arrange that format() builds pdflatex.fmt, which currently is missed as we don't test that way. Have a few adjustments to make, but largely sorted in my head :)
@JosephWright uh-oh
1 hour later…
6:00 PM
This is from the same folk that wanted me to take a course on how to write paragraphs. Note the wonderful English shown in the "This is for:" and the list...
6:30 PM
I was about to ask if there was a good way to do a simple flowchart in LaTeX, then realized making it in Powerpoint and exporting it as a PNG is probably way less work then learning a new package
6:43 PM
@Canageek Another option is to export as a PDF then crop it. :) I also like this approach. :)
@PauloCereda Cropping PDFs is a pain though, isn't it?
@Canageek A flowchart is a pretty simple thing to do with TikZ, though.
@Canageek Ah it depends on how you do. :) There's a tool in Linux named Krop that is quite straightforward. :)
@AlanMunn It is going to be about 4 boxes with arrows between them. It is totally unnecessary, but it will make a pee reviewer happy. I'm sure they are an engineer, engineers LOVE unnecessary flow charts.
@Canageek Yeah, that's a breeze to do even if you don't know much TikZ
@Canageek Or you could go old school state of the art with @DavidCarlisle and use picture mode.
6:56 PM
@AlanMunn If I was going old school, I've got a copy of Visio 1.0, and a Dosbox environment set up to run Windows 3.1 ;)
@DavidCarlisle -- I'm not sure enough of the details to be able to say your answer is the only possibility. So, send it in, and let the "vetting committee" decide. (I'm on a boat until next Monday, so not in a good position to forward it.)
@Canageek Look at Section 3 here: tug.org/TUGboat/tb39-1/tb121duck-tikz.pdf and please let me know if you find it useful. I'm very interested in an expert user's opinion, thank you!
@CarLaTeX I'm far from an expert!
@Canageek You're certainly not a beginner :)
@CarLaTeX I mess around with it every few months or years when I need to write a paper, but I'd put myself at the "understand basic paper writing, but get lost trying to follow complex package descriptions sometimes"
7:07 PM
@Canageek You're my ideal reader! :)
@CarLaTeX Once I have the rest of the edits made, I'll look at it
@Canageek Thanks!
1 hour later…
8:27 PM
@barbarabeeton turns out my answer was flat wrong:-)
Hi, if someone writes a thesis and links 1000 pictures in the tex document, what is the best way to share the tex files with someone who wants to fine tune the preamble and the used packages?
@JonasStein send them a zip file, or share a github repo or ....
I thought about the draft mode first. But is there something better?
@JonasStein What do you mean by 'links'? But just zipping everything should work fine.
something which does not require to collect all files from the harddisk, if the helper does not need the pictures
with link I just mean the includegraphic URL to the file
8:34 PM
@JonasStein hard to say in general, depends what help the helper is offering. If you don't include the images then )in general) the sizes will change so all page breaking, float positions etc will be different, that may or may not matter depending on the intention
sure. The helper does not the final polish, but just basic fixes and a review of the codestyle
@JonasStein draft mode it is then. Perhaps they could mention that having 1000 pics in a thesis is excessive:-)
@JonasStein I guess the simple thing would be try draft mode, and see whether you run into trouble.
perhaps I have to write a wrapper, which greps all URLs in \includegraphics
@JonasStein if the page break is not important one can simple drop the pictures, use demo mode and sent the tex files as a zip. If the pictures are needed some cloud space like dropbox or github or something like this is fine.
8:40 PM
@DavidCarlisle Unless you're doing this: tex.stackexchange.com/q/13594/2693
@AlanMunn oh a 1000 character thesis, all images, quite short then..
@DavidCarlisle Exactly!
@DavidCarlisle A bit repetitive though, I suspect.
@DavidCarlisle In fact, it could contain a single sentence. tex.stackexchange.com/a/423089/2693
@UlrikeFischer pagebreak is not needed at the moment.
@DavidCarlisle, @UlrikeFischer Hopefully I have a simple plan: building for release now
@PhelypeOleinik Guess I should include you too ^^^
Does anyone know what command with MathJaX to write a large integral? It was important for me to be able to edit some answers or questions about Physics SE. Thank you for your valuable help.
8:49 PM
@PhelypeOleinik You'll soon find what you've let yourself in for ...
@Sebastiano \int it will be larger in displaystyle just as in real tex
@JosephWright The super secret plan?
@DavidCarlisle I've tried, but it always stays the same size. I'll probably make a mistake.
@JosephWright wouldn't it be nice to have group pinging? something like @latexteam which pings everyone? ;-)
@UlrikeFischer :)
@PhelypeOleinik not just yet .. dev release
8:53 PM
@AlanMunn well on the document level it is the best we have, but I thought I should better look for a real solution - @DavidCarlisle said I still get the blame for all hyperref stuff.
we have group pings on IRC for Gentoo. It is very useful.
@Sebastiano a large and small int jbergknoff.github.io/mathjax-sandbox/…
If someone wants to play with luametatex but doesn't like ConTeXt, I just uploaded an experimental, pre-alpha LaTeX format for luametatex to a git repo: git.math.hamburg/marcel/luametalatex
@MarcelKrüger oh scary, @PhelypeOleinik will have to take the blame for all latex support issues on that platform:-) (@MarcelKrüger thanks for that, did you find some documentation somewhere or did you reverse engineer the context setup?)
@DavidCarlisle If it's on that platform only... ;-)
9:00 PM
@MarcelKrüger wow I'm impressed. That's why you found that lualibs broke luametatex. Does this need a changed lualibs?
@DavidCarlisle No, reading documentation would be boring :P I tried to stay as far away from ConTeXt as possible because I don't like GPL related issues, so I mostly tried to write a PDF backend in LuaTeX and then chnaged every undefined error...
@MarcelKrüger @UlrikeFischer ^^^^ (the first bit)
@UlrikeFischer Changed lualibs, the latest PR for luaotfload and a shared library build of kpathsea... It was hard to write, so it should be hard to install😂😂 :)
@DavidCarlisle reading documentation of notable authors can't be boring - and some contain even ducks.
@DavidCarlisle Is the large size as large as possible? When there are fractions the integral is very small. It is probably better to use the (num)/(den) notation. Thank you very much for the link.
9:04 PM
@Sebastiano some fonts have straight integrals with extending middle parts but not many (and I don't think mathjax could access that anyway)
@DavidCarlisle I think I've understood: that is, two symbols in half large to have a large integral.
@MarcelKrüger I already have the newest version - I normally always run the newest branch. But the kpatchsea stuff sounds as if I would get problems on windows ;-)
@Sebastiano that is how \left(...\right) grow, the font has the parts to make brackets of arbitrary size, using a repeating straight segment. But most fonts have a sloping integral at just two sizes
@DavidCarlisle You were very precise and the explanation impeccable. Thank you very much.
9:20 PM
@UlrikeFischer I have no idea how this works on Windows, but I think at least MikTeX has a kpathsea.dll. So it might work there(?) Of course the "kpathsea" of MikTeX probably doesn't deserve it's name anyway, but maybe TeX Live uses a dynamic version too?
@MarcelKrüger so is ffi generally enabled with lmtx or do you need --shell-escape or equivalent?
@DavidCarlisle AFAICT there is no built-in shell-escape option, so this would have to be built manually. Especially there are no such restrictions by default.
I want to implement shell-escape at some later point, but then kpse can be made an exception.
@MarcelKrüger suspected as much, thanks
@MarcelKrüger texlive has a kpathsea631.dll, miktex has also a dll, but I have some doubts that it is really equivalent, imho it doesn't really use kpathsea but a similar implementation from Christian.
9:56 PM
@UlrikeFischer I see. I might try to make it work with the TeXLive version, but I'm not sure I will hae time for this tomorrow.
@MarcelKrüger don't hurry, I have lots to do and shouldn't play anyway ;-(
10:41 PM
hey, can we typeset latex/mathjax equations on the main site?
like we can on the math sites, etc
like we can on math.stackexchange.com i mean
@tjt263 no
because i have a latex question that involves a mathematical function/formula that needs to be incorporated into the actual code and..
there goes that idea
@tjt263 if you want to show tex output upload an image
it would be weird in context i think. it just would have been so nice and convenient. it's not the first time either. i thought i was just doing it wrong
@tjt263 In the vast majority of cases having it enabled here would make things much more difficult, since there's already enough confusion about MathJax being LaTeX (which it's not.)
10:48 PM
but of course the LaTeX site is one of the sites that doesn't actually support LaTeX typesetting.
@tjt263 No, it does support LaTeX typesetting... :) That's why it doesn't support MathJax.
it does?
@tjt263 at other sites it is not an issue that mathjax is not tex
well thats fine then, so what's the syntax
@tjt263 Sorry that was intended as humour.
10:51 PM
@tjt263 mathjax is pure javascript, it interprets a subset of latex syntax but the actual typesetting engine is unrelated, so at a site discussing details of tex typesetting it would be very confusing.
@tjt263 Try $\LaTeX$ in a site where MathJax is supported and you'll see a good reason for TeX.SX not using it. ;-)
it's just a utility that interprets latex math syntax and renders it not unlike other fuller LaTeX engines i.e. pdfLaTeX, etc. it just happens to be written/implemented in javascript, because.. well it has to be written in some language, and javascript is just the most suitable frontend tool for working in a browser with CSS and so on. Nobody's confused about their engine being coded in C or C++ or whatever it might be
just sayin
@egreg seems to work on math.stack
you mean the a looks off
11:18 PM
@tjt263 but it's not a re-implementation of the same algorithms and does not implement most of the lower level commands so virtually no answers on site explaining details of tex typesetting would apply. That's not to say it's a bad thing, and on a site where you just want to display an equation and not discuss its typesetting it is reasonable to enable it.
11:29 PM
i don't think i disagree with you. i mean, that's exactly what I want it for
i know how to typeset the equation

« first day (3181 days earlier)      last day (1795 days later) »