« first day (1324 days earlier)      last day (3608 days later) » 

1:21 AM
I think tex4ht is ignored by Latex community. This makes me wonder how folks translate all their Latex document to the web then? Since there is no other option that I can see. Or may be no body cares about HTML here and only pdf is what is important? I think this is a big mistake.
 
 
6 hours later…
7:16 AM
@Nasser Not really - tex4ht did have to be taken over after the original author passed away, which obviously caused some issues
@Nasser Remember that for publishers, a likely workflow is to store data not as LaTeX but as XML, and to create LaTeX, HTML, etc. as derived formats
@Nasser Also, recall that what TeX is good at is typesetting and carefully-controlled output: HTML, even with modern CSS, simply isn't like that (reflowing = no real control over exact placement of material)
 
@JosephWright I do not understand when you say for publishers. Are you saying someone who wants to convert a latex document to html is a publisher? I am talking about every person who uses latex all day. Don't they need to convert these to HTML? I guess many do not. And that is why there is little interest in tex4ht
 
@Nasser I'm talking about people with money who use LaTeX a lot for their business: there is a lot of LaTeX use that 'end users' never see
@Nasser The usual answer here is that if you know in advance that your end target is not just PDF but also HTML then using LaTeX (or even ConTeXt) as an input format is probably the wrong choice
@Nasser Note that ConTeXt scores better here, as they've got more of a consistent system and have been able as part of MkIV development to deprecate some 'TeX-like' input idioms that are hard to convert to HTML (\item and \section-like concepts are obvious examples)
 
But this is the problem: What if you want output to be both PDF and HTML? And also use math? There is no other option than Latex/tex4ht/pdflatex for me. I do not know Context. Is it better than latex/tex4ht? Any examples of its HTML conversion?
 
@Nasser I don't know what your full context is: it sounds like professional work, in which case surely you have a budget for something like 3B2
 
I am student at school. I use latex for HW's and projects. that is all.
 
7:29 AM
@Nasser In that case, why do you need HTML output?
@Nasser Or at least why can't you use e.g. Pandoc
 
Because I like to browse my HW's on the web. Much easier than having to download a PDF for each small document. Imaging if wikipedia was all pdf files and no HTML? There is advantage for each format. Looked at pandoc, hard to do math. Nothing better than latex for math.
 
@Nasser Like I say, some of the issues arise from the twin facts that people who need to spend money on HTML conversion don't do it using TeX but using (very expensive) commercial publishing platforms, and the fact that the author of tex4ht died and finding a volunteer team to take over is not so easy
@Nasser Also, as I've tried to indicate, the design LaTeX isn't ideal for conversion to HTML: you need more structure and less user flexibility (or at least it has to be carefully thought through)
 
But what other software is there, that can do as good math as latex and generate to HTML? I know of nothing that can do what you are saying. If I have one million dollar to buy this software, where is it? What is it called? Does 3B2 do better math than latex? I tried many software, and they are all bad. never tried 3B2.
 
@Nasser 'Better' is a tricky term: 3B2 is derived from an older product which used to use TeX 'behind the scenes'. I've not seen it myself, but I know it's used for a lot of journal publishing. What those people are after is automation: hundreds or thousands of manuscripts, PDF/HTML/database output without any human intervention, etc.
 
7:44 AM
@JosephWright yes, I know this too about it. That is the problem, I know many good software that can target to HTML very well, as long as there is little to no math involved. indesign for example. When complex math comes in, then they all fall apart. I would use indesign, but it does not do math well at all (terrible equation editor software).
 
@Nasser InDesign is different: it's a GUI. Like I say, I've not seen 3B2 but as I understand it it's not a simple GUI for layout, it's an automation system. As such, input is auto-scraped from LaTeX/PDF/Word files into XML. It's really very different from an 'end user' package in that sense.
 
Also tried frameMaker, good for technical, as long as little math involved.
 
@Nasser Again, you are thinking about single documents: the business end of HTML production is automated 'black box' systems for large scale conversion. That's where the money goes in.
@Nasser For serious maths, I'd be tempted to ask Barbara Beeton about what the AMS do
 
@JosephWright You think AMS has some hidden software better than htlatex? :)
 
@Nasser No
@Nasser I suspect they hand-tidy into a known subset of LaTeX that works well with conversion to HTML
 
7:51 AM
I did not know AMS publish to HTML much. Very few journals do.
 
@Nasser Depends on your area: all of the journals in chemistry have HTML versions, done using 3B2/ArborText mainly
@Nasser Taking a look at the AMS pages, I see they stick to PDF
@Nasser I suspect one issue is that for very maths-heavy stuff the current HTML situation (MathML, etc.) is not developed enough that professionals in the area want to risk such things. PDF works for them.
 
You see, I told you :) That is the problem. Wikipedia went for HTML only (no pdf) and journals go for PDF using pdflatex. Hard to do both well.
 
@Nasser Like I say, different requirements (the AMS have to have the 'correct' output: Wikipedia isn't quite the same!)
@Nasser One problem is the nature of the output. If you look at Wikipedia, they use a lot of images for maths, but people like @DavidCarlisle are working on 'proper' maths support for browsers, etc. Until that is all standard to the point where you don't have to worry about it failing to load, it's very risky (in general) to produce a 'pure HTML' solution.
 
@JosephWright I tried output mathml and render using mathjax. So no image for math. Very nice looking output. The problem is tex4ht has few bugs in its mathml outout, and some math was failing. I posted all these on this forum. The worst was that it could not do siunitx package. When I used \SI[], then the HTML will not even get generated ! So I swithed back to png for math, at least it works.
someone said the \SI[] problem was with mathjax and not tex4ht. But I have no idea. over my head. I just stopped using mathax for this reason.
 
8:06 AM
... and after managing everything on time yester, no matter the strikes, I slept late today and missed the only two morning buses to the port :(
 
@Nasser Problem there is that I've set up siuntx to do it's best to give useful output from tex4ht but you've surprised it by trying to use math mode!
@Nasser One thing people use tex4ht for is conversion to Word, and for that to work I have to force text mode (spacing weirdness in OpenOffice)
@Nasser I'm actually working on a new version of siunitx and this sort of thing is one of the reasons. The problem is partly that trying to support all possible outcomes is very tricky, at least with the way I initially defined \SI as working. (Back with the ConTeXt business: if I was starting from a blank slate life would be easier.)
 
@JosephWright but I have to use \SI in math mode many times. like inside \begin{align} and such. Most of my use of it is inside math mode. I am engineering student, and so most of my equations have units :)
 
@Nasser No, you misunderstand me :-)
@Nasser The most common desire when converting to HTML seems to be to want a unit to come out 'just as text' but with superscripts/subscripts correct: \SI{10}{\square\metre} => 10 m<sup>2</sup>
@Nasser Like I say, I have a plan that I think addresses the issues, but it's non-trivial and I have other things on my plate (day job, for a start)
 
@JosephWright if you have a wikipedia account you can select mathjax output, they're working on making that the default
 
8:19 AM
@DavidCarlisle Still a hack, really
 
@JosephWright no need to tell me:-) But point is documents last for thousands of years so a hack to work with current browsers but allowing standard documents isn't so bad and the chrome developers think it isn't a hack but the future: they're on record that they don't really believe in "documents" but rather see thier job to make a snazzy javascript+css platform on which people can develop web applications, and see documents as a rather small side issue......
@Nasser the best way to support siunitx in any tex-to-xx convertor) would be to have a specific siunitx configuration (for mathjax or tex4ht ot latexml or ...)that way you can map the high level markup straight to the converted document. The alternative is to hope the convertors actually understand @JosephWright's tex definitions in the package but (a) that is unlikely and (b) if they do the result of looking at it via the expanded tex definitions is inevitably that you generate a mess:-)
 
@DavidCarlisle Yes, true
@DavidCarlisle I guess I should work out how to arrange such a thing
 
@JosephWright easiest is to put an enhancement request on the mathjax github issues list and hope Davide does it:-)
 
@DavidCarlisle I meant the tex4ht part
 
@JosephWright My understanding of tex4ht always relied heavily on mailing Eitan and asking him to configure it for me:(
 
8:31 AM
@DavidCarlisle Yes, back with my earlier point about the issues since he died
 
@JosephWright is it to be expected there are no updates between last prestest and first full 2014, I think ive switched repository back to warwick and it doesn't complain about the year but says no updates available
 
@DavidCarlisle I got some updates, but hadn't been doing a pretest one every day
@DavidCarlisle Certainly the mirror is up-to-date (I use the same one)
@Nasser Will work on tex4ht issue and siunitx: proper config is as @DavidCarlisle says the way to go
@DavidCarlisle Did you see that l3build is now on CTAN?
 
@JosephWright I saw you said it was, hadn't actually looked yet:-)
@JosephWright is macros/contrib/latex the natural place for it?
 
@JosephWright Yay, congrats! :)
 
@JosephWright I once tried to move tex4ht configuration in siunitx.sty to siunitx.4ht (correct way to configure packages with tex4ht), but I have found catcode clash between LaTeX3 commands and tex4ht (both redefine _ and :). in fact it is maybe good question, how to support LaTeX 3 macros in tex4ht configurations
 
8:53 AM
@michal.h21 Is "Blame David." a valid answer? :)
 
@DavidCarlisle As you're here at the moment: Is there an easy way to convince XMLTeX to output <br/> instead of <br></br>; otherwise it seems I'm in trouble trying to generate HTML.
 
Apr 23 at 12:11, by David Carlisle
@PauloCereda no
 
@DavidCarlisle LOL
 
@StephanLehmke xmltex doesn't output anything, it reads xml ?
 
@DavidCarlisle Well I can output what I read with \protected@write, no?
 
@PauloCereda it seems it is universally accepted answer. also as he is connected with LaTeX 3, it is perfectly valid :)
 
@michal.h21 LOL
 
@StephanLehmke yes but you can write whatever you like, xmltex doesn't help you does it? (I suspect I'd be confused even if I had looked at the xmltex sources:-)
 
@DavidCarlisle AHA!
% #1 should be empty, between the / and the >
% probably should put some internal form here rather than literally adding
% end tag to be reparsed, but this simplifies grab code.
\def\XML@endempty#1>{
  \expandafter\XML@startelement
  \expandafter<\expandafter/\begintag>}
 
@michal.h21 You've been spending too long here, you're getting tobe as rude as @PauloCereda and :egreg
 
9:24 AM
Nice to know it was easier that way ;-)
 
@David: Stephan is doing a great job with xmltex, you should include him as maintainer. :)
 
@StephanLehmke yes but once you've triggered your macro template for br, you know (or can re-check) it's empty so write out the /> form (or as it's html the > form with no end tag) can't you?
 
@DavidCarlisle I'm talking about inline HTML stuff which was read with \xmlgrab and then written out again.
 
@StephanLehmke yes but the grabbing bit (which I did) can treat /> and </br> the same, it's the writing bit (which you did) which needs to write it as />)
 
@michal: don't mind David, he's sad because his only hopes on getting the money jar are on Chile's hands. :)
 
9:28 AM
The following work really well: (a) read XML and interpret it (b) generate HTML from TeX code. But not (c) read HTML intermixed with text and write it again as HTML.
 
@PauloCereda I may sell the xmltex IP to Stephan and retire rich.
 
@DavidCarlisle ooh XML money!
 
@StephanLehmke I'm sure it's not my fault but apart from that I don't understand, perhaps you should make a MWE:-)
 
@DavidCarlisle Well if I grab some arbitrary text interspersed with HTML markup like foo <b>bar</b><br/>baz then how can I write that back without having <br/> change into <br></br>?
@DavidCarlisle Yea it's probably better, then someone can get rep ;-)
 
@StephanLehmke how are you writing anything, don't you have templates matching elements and then write to a file? If so you just want a template matching br
@StephanLehmke but what about my retire rich plan? I can't do that on rep alone.
 
9:34 AM
@DavidCarlisle Better charge for longtable ;-)
@DavidCarlisle I still don't get that part. If I get something like foo<br/>bar into a macro with \xmlgrab, I can write that macro out with \protected@write without interpreting <br> at all. How would I do it with interpreting <br>? How would foo and bar get written out?
@DavidCarlisle If you need a new job we could hire you :-)
 
speaking of html, I have new toy, html to LaTeX convertor in javascript, with support for CSS (not yet):
 
@michal.h21 Tell me when you can convert canvas stuff to TikZ ;-)
 
@michal.h21 cool! Blog post? :)
 
@StephanLehmke I was thinking about converting svg to tikz, which I had done before using xmltex :D, but if it is possible to read canvas data using dom, it may be possible :D
@PauloCereda I hope so, after it is done
 
9:56 AM
@DavidCarlisle Looking at the code, actually it shouldn't be that hard to do: put something like \XML@startemptyelement there and differentiate in \XML@doelement. I'll cook something up...
 
10:16 AM
Out of curiosity, I believe LaTeX sets a limit of floats per page? :)
 
10:35 AM
@PauloCereda You do not know Frank's famous answer?
301
A: How to influence the position of float environments like figure and table in LaTeX?

Frank MittelbachTo answer this question one has to first understand the basic rules that govern LaTeX's standard placement of floats. Once these are understood, adjustments can be made, for example, by modifying float parameters, or by adding certain packages that modify or extend the basic functionality. LaTe...

 
@HeikoOberdiek Oh my, I completely missed that one!
 
11:21 AM
@StephanLehmke ok I may look later if if needed you give me a clues where to look,
 
@Mico Anything new on this matter? tex.stackexchange.com/a/145619
@Mico, it would be great to enable the replacement of fl by \char"E0E0}\-\hspace*{0pt}l if one of the selnolig rules applies.
@Mico I see that you haven't commited to selnolig since september. But maybe you made some thoughts on this topic. Libertine would be great but you mentioned that you will take a look on EB first. Shouln't that be the same approach, if you replace by an unicode-char?
 
11:57 AM
@DavidCarlisle It worked :-) Look at my beautiful <br /> :-)
While I was at it, I also tried to fix the annoying fact that you'd get a spurious name space in things like <a 0:href=" which validators or tools like xml_pp would complain about (not tested yet).
The hardest part was not coding it, but dealing with catcode madness.
XMLTeX might be a shrew, but in the end it can achieve everything :-)
 
@StephanLehmke I wonder who wrote that brilliant code. :)
 
@PauloCereda It works so well, I can even fix it without understanding the least bit of it :-)
Maybe we should port LaTeX3 input syntax to be based on XMLTeX, that would teach them :-)
 
@StephanLehmke You know that the user syntax isn't actually decided on yet, I guess ?
 
12:13 PM
@JosephWright boo.
 
@JosephWright XML has a lot of advantages :-)
 
Hello @Paulo, what do you think of my new cover? dickimaw-books.com/fiction/crime/the-private-enemy/… (No ducks or chickens, but there is a hat.)
 
@NicolaTalbot ooh it looks awesome! :)
 
@PauloCereda Magdalene did the painted bits and the rest of it was me playing around with gimp. Oh, I've been asked to do a book reading in August, so I'll bring along the puppets :-) (Although I'm not sure about the age group of the audience!) It's some kind of literary event, so I might get the opportunity to tell everyone about the benefits of typesetting in (La)TeX. :-)
 
12:28 PM
Eliminating the unwanted 0: namespace prefix also worked! Yay!
 
@NicolaTalbot :)
@Nicola: ^^ :)
 
@PauloCereda Oh no! My gangster is about to be run over by a duck!
 
ediff finds 2000 differences now in my 2.7MB HTML file :-)
 
@NicolaTalbot wooohooo
@StephanLehmke Holy cow!
 
@StephanLehmke try appending > /dev/null to the ediff command line that usually gives a more comforting report
 
12:42 PM
@DavidCarlisle Er well I wanted the differences :-) As many less HTML validation errors :-)
 
 
3 hours later…
3:17 PM
TikZ experts around? :)
 
3:28 PM
@PauloCereda me me
 
@DavidCarlisle <3
 
user image
3
@PauloCereda I don't know if I qualify as an expert, but you could try me and find out.
 
3:43 PM
@Paulo Cereda: Hi
@PauloCereda: Hi
@PauloCereda: It's about the blog post
 
@ManuelKuehner Hi hi hi hi.
I got the ping.
I guess. :)
 
Ok. So what do I do?
Is it a blog like WordPress?
 
@AlanMunn Thanks Alan, just a minute. :)
@ManuelKuehner First, I think you need to have permission to access the blog infrastructure. @Joseph: are you around?
 
@PauloCereda Yes
 
@Manuel: could you try accessing here first? tex.blogoverflow.com/wp-admin
 
3:46 PM
@JosephWright: Hi
 
@ManuelKuehner Hello
 
@JosephWright Thank you! I think we might need to give permissions to Manuel. :)
 
Yes. I see the Dashboard
 
@ManuelKuehner Try now
 
@JosephWright: Try what?
 
3:48 PM
@ManuelKuehner Re-loading the WordPress dashboard: you should have author privs
 
@JosephWright: Like this (sceenshot?)
 
@JosephWright Apparently now he does. :)
 
@ManuelKuehner Looks good
@PauloCereda How long before we get the first Lua-related bug report for l3build?
 
@JosephWright: So I just start a new post and add puctures, backround information and a bit of TeX code?
 
@JosephWright Tomorrow. :)
 
3:50 PM
@PauloCereda I'm expecting you to volunteer to join the team to support it!
@ManuelKuehner Yup
@ManuelKuehner Normal procedure is to save a draft then ask in chat for some reviewing
 
@JosephWright: Ok. Thanks. I am leaving now. Ok?
 
@ManuelKuehner Cool
 
@PauloCereda: bye
 
@ManuelKuehner Thank you, Manuel. :)
@JosephWright Perhaps I should delve into the project, that way I can promise results to you guys. Fine, sign me in.
 
@PauloCereda Well at the moment what would genuinely be handy is if you'd look over the Lua code in l3build: I'm no expert on these things, and it's very much hacked together
 
3:56 PM
@JosephWright I'll take a closer look today, I already made some notes about it. Don't worry. :)
@Joseph: more info when I see you online. :)
 
HTML is frustrating. It's almost, but not quite, entirely unlike XML :-(
2
Hard to convince XMLTeX of producing it...
 
@StephanLehmke We could blame the author.
 
@PauloCereda Well to be honest, XMLTeX is a small part of my problems ATM.
Well, I discovered that in HTML you can't use &#13; which is exactly what XMLTeX will produce if there is a line break in an arbitrary chunk of text ;-)
 
@StephanLehmke Oh. :(
 
In fact, about half of my problems stem from DocScape being a bit too forgiving when it comes to content structure.
 
4:10 PM
@StephanLehmke Ah, like LaTeX only probably less so
 
For instance, DocScape has a construct for containing "paragraphs" which is consequently output as <p>. But when text content comes from rich text editors (i.e. CMS systems), it can happen that some anonymous chunk of text contains itself some <p> tags which are then part of the "paragraph" content. On the TeX side, this is sorted out at execution time. But HTML does not accept <p> inside <p>, so I have to filter the text content when writing HTML :-(
 
@StephanLehmke I can imagine the issues: nice flexible input is great as a user but there is a cost
 
About the other half of my problems is that I based a lot of my internal constructs (for instance, tables) on HTML 4 structures, for instance <table>, <tr>, <td> with all the relevant attributes, which are now all deprecated for HTML5 :-(
 
@StephanLehmke use div not p (that's what I do, simplifies things enormously:-)
 
@StephanLehmke: Shouldn't you be watching the game? Germany is winning so far! :)
 
4:21 PM
@DavidCarlisle Is <div> really the same as <p>???
@PauloCereda I've got a radio here :-)
But the I am not that much of an enthusiastic fan.
 
@StephanLehmke :)
 
@StephanLehmke it's p with a fixed content model: allows displayed lists etc but doesn't have p's default css adding a \parskip so you need to add some css but <div class="p"> with css of div.p {margin-top: 1em;} gets you running again
@PauloCereda oh the game, that reminds me...
England 	575-9 & 267-8
Sri Lanka 	453 & 189-5
Sri Lanka need another 201 runs to win
 
@DavidCarlisle Is England winning?
 
@PauloCereda possibly
 
@DavidCarlisle Frustration is building up...
 
4:26 PM
@DavidCarlisle Tight
 
@PauloCereda If Sri Lanka score 201 more runs, they win, If England get 5 Sri Lankan batsmen out, they win, if the 5 days are up without either of those happening, it's a draw.
 
@DavidCarlisle If Sri Lanka score exactly 200 and then are all-out, it's a tie
 
Gah! If I set the attribute border at a table element, then all the borders are on, but if I set the CSS atribute, then only the outer border is set!?! :-(
 
@DavidCarlisle No game should end up in a draw. :)
 
@StephanLehmke Isn't CSS great
 
4:27 PM
@JosephWright Too complicated. :)
 
@PauloCereda Seems entirely straight-forward to me :-)
 
@JosephWright :)
 
@PauloCereda you get a "tie" if the game ends but the scores are level. a match is "drawn" if it runs out of time
@PauloCereda on the contrary it keeps things interesting, if you were 100-0 down at half time in football you might be tempted to give up but in cricket, even if you can't win. if you can just survive until end of play, you can still rescue a draw. Surviving when people are hurling rocks at you at 100mph is an interesting spectacle for the spectators:-)
 
@DavidCarlisle And the team who is out needs to be in and vice versa?
 
@PauloCereda exactly
 
4:31 PM
GOOOOOOOOOOAAAAAAAAAAAAAAAAAAL!
Germany!
 
@PauloCereda at Lords?
 
@Stephan: GOAAAAL!
 
@DavidCarlisle Indeed: I tend to describe the ball as 'essentially made of wood'
 
Angela Merkel briefly appeared on television. Spooky.
@DavidCarlisle At where?
 
Lord's Cricket Ground, generally known as Lord's, is a cricket venue in St John's Wood, London. Named after its founder, Thomas Lord, it is owned by Marylebone Cricket Club (MCC) and is the home of Middlesex County Cricket Club, the England and Wales Cricket Board (ECB), the European Cricket Council (ECC) and, until August 2005, the International Cricket Council (ICC). Lord's is widely referred to as the "home of cricket" and is home to the world's oldest sporting museum. Lord's today is not on its original site, being the third of three grounds that Lord established between 1787 and 1...
 
4:35 PM
@DavidCarlisle ooh a cricket stadium!
 
@PauloCereda "ground", lesser sports have stadia :-)
 
@PauloCereda Not really a stadium: that would be the MCG
 
@JosephWright Oh. :(
GOOOOOOAAAAAAAAL!
@Stephan: GOOOAAAAAAL!
Where's @egreg?
 
It seems there is always one team who can reasonably complain about referees; this time it's the Portugese...
 
@StephanLehmke Pepe is an idiot. He gets a red card on every game he plays, even for Real Madrid. :)
 
4:50 PM
@PauloCereda Well on the radio it sounded like he didn't do anything really, just acted "as if".
 
@StephanLehmke if player.getName().equals("Pepe") then referee.apply(player, Card.RED); end
 
@StephanLehmke I don't know what the rules are, but although he definitely didn't really head butt him, it was a definite provocation.
 
@AlanMunn It's maybe too hot, and too strong coffee before the game.
 
5:21 PM
@PauloCereda You should add a translation function to Psmith. For example "referee" -> "umpire", "off-side" -> "LBW", "GOOOAAAL" -> "Howazat!"
 
@NicolaTalbot Goal -> Howzat seems a bit of a stretch...
 
@AlanMunn Possibly. I couldn't think of a better one.
 
@NicolaTalbot ooh! :)
GOOOOAAAAAAAAL!
I mean,
Howazat!
:)
 
5:38 PM
It seems that all the Germanic speakers are beating the crap out of all the Romance speakers...
 
@PauloCereda Erm, out/not out (delete as appropriate) :-)
 
ooh! :)
 
@AlanMunn Which would make English a Romanic language ;-)
 
@StephanLehmke Yes, I guess that's true. Well 2/3. I forgot about the England match.
 
@PauloCereda it went down to wire: Match drawn - SL 201-9

Last ball...edged...just short of the slips. Sri Lanka survive, England are denied. A wonderful final day ends it a draw. That is why we love Test cricket.
 
5:49 PM
@DavidCarlisle Wait a minute. Test?!
 
@AlanMunn I'd have assumed the other way round, with the climate favouring peoples from warmer countries...
 
@StephanLehmke It's that colonialist grit...
 
@PauloCereda Test cricket is the full form played at international level with 5 day matches, other fomats include 1 day cricket or 20-20 or limited overs or various other ways of reducing the time
 
gnashes teeth It just turns out HTML attributes are positional. What a nice surprise! NOT!
 
@PauloCereda 1 day cricket is the one that's colloquially known as "pyjama cricket".
 
6:15 PM
A question over at latex-community:


Alignment

    How do you do these kinds of things, where you have something in the left and then something in the center:

(1) (insert space here) 1/2=0.5
 
@NicolaTalbot You have now propagated a misspelling to poor @PauloCereda...
 
^ That is the whole post.
 
@AlanMunn Oops. I should've consulted the dictionary. (No excuse, really. It's right by my elbow.)
 
@Johannes_B We had a question a few days ago that was basically "How come when I typeset my document, the text doesn't look right."
 
@AlanMunn I think that was my first laught today.
 
6:18 PM
@Johannes_B I must say I really wonder about what is going on in the minds of such posters. Do they believe in telepathy?
 
@AlanMunn I think they are not really aware that the question is soo extremely vague.
@AlanMunn btw: I can't find my car key, do you know where it is?
@AlanMunn I think you saw a car once, must be easy for you ;-)
 
@Johannes_B Yes, I think that must be so, but it still amazes me.
@Johannes_B Look over here, there's more light.
 
@AlanMunn Ah, found it :-)
 
@Johannes_B :D See how helpful I was!
 
»I searched google for a few days now, but i couldn't find any results. What i want is <put any default setting here> but i can't figure out how to do it«
@AlanMunn You are a genius.
 
6:45 PM
well, sh*t, Deutschland just kicked our asses
 
@MarioS.E. You can fraternize with Spain now :-)
 
kan
There was definitely one free kick that man Ronaldo was not given... :-/
 
@NicolaTalbot LOL
@MarioS.E. To be honest, I was expecting a better game from Portugal. But Pepe... dude, when that guy will ever learn.
 
@PauloCereda I'm also spanish :(, I'm feeling it
...why the hell is the Iberia Penninsula taking so much goals?
@PauloCereda Pepe é um bruto
 
@MarioS.E. ... e um idiota.
 
7:10 PM
@MarioS.E. Hubris.
 
@StephanLehmke ?
 
8:03 PM
@AlanMunn This... I... well, at least I have Italy :D
 
8:16 PM
Is this on topic? It's only about gnuplot
1
Q: gnuplot: plotting multiple functions with different ranges

user2869037I would like to plot the inverse of y=xe^x, with dashed lines leading to the point (-exp(-1), -1) set parametric set arrow 1 from -4,0 to 4,0 nohead set arrow 2 from 0,-4 to 0,4 nohead set trange[-4:4] set xrange[-4:4] set yrange[-4:4] set xlabel "x" set ylabel "y" plot [-4:4] t*exp(t), t lt...

 
8:35 PM
Installing TL 2014 on my Xubuntu virtual machine
 
9:22 PM
@egreg I think it is off topic.
@egreg I think it is off topic.
@egreg I think it is off topic.
@egreg Do we have a gnuplot.stackexchange? Or something the like?
It keeps telling me to retry. Internet connection is very bad today :-(
 
@Johannes_B It happens. But the message has arrived. ;-)
 
@egreg Looks like i am crazy :-D
 
@DavidCarlisle That was wrong. CSS attributes are positional. But all the HTML attributes are deprecated now and replaced by CSS attributes, so I'm totally confused.
 
@egreg Did you flag it?
 
@Johannes_B Yes, I did.
 
9:26 PM
@egreg Thanks
 
@StephanLehmke you mean things like align= in tables? well you can use them anyway of course they all work same as ever, but if you want it to pass an html5 validator most of them map directly to css so you just need to generate table class="foo" and then dump the alignment in the css.
 
l3build looks interesting
 
9:46 PM
@Johannes_B Any bugs in it you can blame @PauloCereda
 
@DavidCarlisle I'll have a look at it ;-)
 
9:57 PM
@DavidCarlisle I just didn't realize that border-left: 1.00374pt; border-left-style: solid; is different from border-left-style: solid; border-left: 1.00374pt; :-I
@DavidCarlisle Generating "real" styles and classes and such is certainly the way to go, but I have to implement it all in TeX, you know ;-)
For the time being, I'm simply attaching a bunch of HTML and CSS attributes to every HTML element.
@DavidCarlisle While you're here: Is there a way to refer to something in a HTML file (say, an equation number) from another part of the HTML file? I made a "private extension" of the <a> tag to refer to a page number in DocScape, but this is now invalid HTML of course.
 
@StephanLehmke ah well that's because border-left is a shortcut if you'd used border-left-width the order wouldn't hav emattered
51 secs goal for usa:-)
 
@DavidCarlisle Ok, so it's an intermittent problem while transiting from the HTML border to a half dozen CSS attributes :-)
 
@StephanLehmke you could in theory use css counters but implementation is a bit spotty still, I always generate teh numbers statically and just use html for the hypelinking so <span id="foo">(2)</span>.... <a href="#foo>see (2)</a>
 
@DavidCarlisle Well actually I'd like to keep the DocScape functionality (of giving the page number, i.e. \pageref) intact ;-)
 
@StephanLehmke by page number do you mean a number after an html page is paginated by the html formatter or do you mean one html document per page?
 
10:07 PM
@DavidCarlisle No the page number of the PDF document. After all, DocScape's main functionality is producing PDF, HTML is just a by-product.
 
@StephanLehmke but what do you want it to do in the html? that number or a number that refers to the html context?
 
But whatever "pseudo-HTML" I used to denote a page reference, must now become valid HTML of some sort (while retaining the functionality for PDF production).
@DavidCarlisle At the moment I would be satisfied with not producing a validation error :-) I have to sort out the whole reference business later.
Probably something like <a class="pageref" href="#foobar">xyz</a> would do. For PDF production, I can interpret this any way I want, for HTML I assume it's sufficient to provide a dummy class "pageref".
 
@StephanLehmke well you don't need to define the class at all if you don't want special formatting, referencing undefined class isn't an error
 
@DavidCarlisle Good to know. So nonsense class names are gonna be my swiss army knife for hiding PDF-specific functionality in HTML ;-)
 
@StephanLehmke also any attribute beginning data- is valid in html5 so you can use <span data-docscape-pageref-id="zzz"></span> span and fill it in by some post-process or in teh browser with javascript or...
 
10:19 PM
Hm, the HTML validator constantly reports 1001 errors, no matter how many errors I fix. Well at the end it says "too many errors" :-)
 
@StephanLehmke are you using validator.nu based validator?
 
@DavidCarlisle Ok, another important information :-)
@DavidCarlisle validator.w3.org
 
@StephanLehmke see the link that says "as an alternative non-dtd based" use that one if you are targeting html5, it's a w3c hosted instance of this validator.nu
 
@DavidCarlisle What's the difference?
 
@StephanLehmke many html "features" can't be expressed in dtd (eg any attribute with name starting data- is valid:-) and html5 drops the dtd altogether and expresses the grammar in prose, but validator.nu does a good job of capturing most of it (it's relaxng based at its heart but mostly it is custom java code, not validating to a grammar
 
10:27 PM
@DavidCarlisle So you mean the other validator will report stuff as invalid which is actually valid?
What side are the browsers on?
 
@StephanLehmke yes and the other way round:-)
@StephanLehmke the forces of evil
 
@DavidCarlisle So is there a way to validate the validators?
 
@StephanLehmke for html5 there is only validator.nu in practice:-)
 
I mean, what about formal grammar and such? Certainly there must be a way to ascertain what syntax is really correct???
@DavidCarlisle Hm, what should make me trust it? It doesn't have any certifications on the front page!!1!
 
@StephanLehmke Ian H doesn't believe in grammars they see it as a major feature if the html5 rewrite of html that they got rid of the grammar, the language is defined by the actions of the parser state machine which is defined via thousands of lines of english text and no formal grammar or machine notation.
 
10:33 PM
@DavidCarlisle Oh great just what we like for automated data processing. If that's what they want, why don't they just switch to LaTeX?
 
@StephanLehmke html4 was (officially) defined by dtd so an input stream that was not valid was offically "not html" and had essentially no defined behaviour. Of course browsers did not generate errors and reverse engineered each others recovery. HTML5 turns it on its head and defines a behaviour for every uniocde inpit stream, but some are classed as invalid, you can't easily do that with most grammar languages
@StephanLehmke because they hate latex and xml and me:-)
 
@DavidCarlisle Well actually they probably hate computer science too, sounds like a "strong AI" type attitude.
 
@StephanLehmke but weird as it sounds it does almost make sense, as while html4 was cleanly defined via dtd the implementations were wildly differemt, but the parsing of ie/safari/chrome/firefox is now very close and very close to the spec (support for actually implementing the various features is variable of course but getting the thing to parse the same in all browsers is a big plus)
 
@DavidCarlisle Well you're more experienced of course, but I don't see why this wouldn't have been possible with XML syntax defined by a proper RelaxNG schema (maybe with some Schematron rules). I'd assume the main factor for acceptance is taking a step back and thinking about a sensible semantics which can satisfy the needs today and in the future.
I mean, HTML syntax seems to be quite negligible at a time where all the functionality comes from CSS styling and JavaScript programming (I still can't come to grips with canvas, WTF???)
 
@StephanLehmke HTML5 specifies exactly what parse tree results from <hh<<<&&&< 8888></table> it would be hard to say much about that with schematrion
 

« first day (1324 days earlier)      last day (3608 days later) »