TeX, LaTeX and Friends - 2018-01-20

« first day (2638 days earlier) ← previous day next day → last day (2296 days later) »

1:57 AM

@DavidCarlisle -- please reconsider your edit to tex.stackexchange.com/posts/411233/revisions . granted, the original had severe problems, but you've added some new ones. (\rhead{\\Subject} for one.) i was tempted to revert, and then comment that the code is probably not compilable (well, it doesn't have any preamble), but since i'm unable to test, i didn't think that was a good idea.

2:19 AM

@barbarabeeton I didn't edit it other than highlight the code block and hit ^K ? the OP had \\ in rhead

2:40 AM

@DavidCarlisle Perhaps there is some confusion: @barbarabeeton is probably talking about the \\ in the argument of \rhead, not \\rhead. At least superficially it looks as if that got introduced in the edit.

@marmot @barbarabeeton No it was already there; @DavidCarlisle did not introduce any new problems. Click on "side by side markdown" to see what the user had typed. (Earlier, even though the user had typed `\`, the software may not have shown it; that may be why you thought it wasn't there earlier.)

2:57 AM

@ShreevatsaR Thanks for the clarification! That also means that I have to be much more careful when "approving" edits. Is there any way to show all \\ as they were typed?

2 hours later…

4:53 AM

user image

@marmot to see what was typed exactly, you can click on "Side by side markdown" as above

instead of "inline" or "side-by-side", as below:

user image

user image

4 hours later…

8:29 AM

@PauloCereda Two of the germans don't drink beer ;-). (My husband says there is brasilian wine and he would like to try it -- he had only one bottle in his whole life.)

@UlrikeFischer And one of the Italians doesn't drink coffee :)

@JosephWright @PauloCereda we will probably try to stay a bit longer. For a reason that I don't understand flights are much cheaper if there are more days between outward and return flight.

@UlrikeFischer It's the day you fly, I think: some are busier than others

@UlrikeFischer Like I said, I will see what the team plan to do: I might stay the Monday (until the evening)

@CarLaTeX I never drank coffee until we were in italy. Now I like an espresso after diner ;-)

@UlrikeFischer Everybody here like espresso, I don't!

8:39 AM

@CarLaTeX @UlrikeFischer by coincidence bbc news this morning: bbc.co.uk/news/av/world-europe-42750584/…

@DavidCarlisle LOL However 40 seconds without taking into account the chatting :)

8:58 AM

@UlrikeFischer @JosephWright: I think it might be the day, but it also involves the logistics in arranging the flights. :)

@UlrikeFischer To be honest, I don't like alcoholic beverages! :) I can tolerate beer because I don't want to feel an outcast when hanging out with my friends. :)

@PauloCereda If you don't like alcoholic beverages, don't go to Veneto!

@CarLaTeX ooh why? :)

@PauloCereda Ask @egreg

I may be monochrome but I am apparently the only one bringing colour into people's lves

tex.stackexchange.com/help/badges/611/color?userid=1090

9:14 AM

@DavidCarlisle :):):) sun always shines on pineapples

@DavidCarlisle oooooh

9:30 AM

@JosephWright no one reads your blog texdev.net/tag/lua

@DavidCarlisle :)

@DavidCarlisle I do feel a bit like a stuck record

@JosephWright one for TUG 2018. :)

@PauloCereda Possibly: I suspect what I want might not align with what the Lua people do :)

@JosephWright Beware, for they will be there. A pack of them. :D

@JosephWright On a serious note, I have the same feelings. I was mid-through some implementations for my thesis and got stuck with Unicode...

@JosephWright I will post a note in the UK-TUG mailing list later on.

@PauloCereda Indeed: if we could do sorting/case changing then we might imagine re-implementing Biber in Lua (and at some stage MakeIndex)

@PauloCereda Cool

9:43 AM

@JosephWright Exactly. That would be great.

@JosephWright: which reminds me, I might poke Hans and Taco.

@JosephWright of compile icu or some other C implementation into luatex....

@PauloCereda I can see why the Lua people don't want to do it

@DavidCarlisle Yes: Hans did mention (vaguely) something about doing their own Unicode support

@JosephWright Me too, it's by design.

@PauloCereda Quite

@PauloCereda I do wonder how long that is realistic, though, in this area: treating text as bytes and expecting every user to 'roll their own' UTF-8 support isn't sustainable as a position (see our changes in the LaTeX kernel ...)

@JosephWright Lua 5.3 introduced support for UTF-8 sequences, but I don't think it's quite enough...

9:51 AM

@PauloCereda It might be enough ... but it depends on your application

@JosephWright I think it's enough from the language support, however the user might need a better way of handling UTF-8 more conveniently.

@JosephWright The project pitfall, basically. See arara, people started using way beyond my original expectations...

@PauloCereda Oh sure

@PauloCereda @DavidCarlisle is likely right: it's probably only LuaTeX that needs more complex Unicode stuff, and that can be handled by including a C library

@JosephWright adding C libraries rather than going through lua seems to be against the luatex philosophy though, hence the impasse over harfbuzz

@JosephWright Indeed. And they can even benefit from an already consolidated Unicode library, since LuaTeX includes additional features like LFS and friends.

@DavidCarlisle Yes, but LuaTeX already includes e.g. slnunicode as-standard

9:58 AM

@DavidCarlisle $5 someone will mention porting the C backend of LuaTeX to Rust. :)

@PauloCereda Exactly: there is a balance

@DavidCarlisle The difference I think is that Hans feels that font handling is something they want flexibility in: I doubt he wants to do what is pretty 'fixed' work to support Unicode fully

@DavidCarlisle In the end we may end up trying to sort something (perhaps trying to recruit other Lua users): I think longer-term we will want sorting in Lua, as both Biber and MakeIndex functionality would be much better done that way

@DavidCarlisle, @PauloCereda One I think we should try to get a 'session' on a TUG2018: we'll have both the Lua devs and Hans there, so can see what agreement is possible

@DavidCarlisle Oh for pity's sake, slnunicode is likely to be removed from LuaTeX ...

@JosephWright I can lock you guys in a room and provide some clubs and medieval weaponry and open 2 hours later. :)

10:14 AM

@DavidCarlisle Looking at ICU, it doesn't mention sorting [we've already done case changing ;)]

10:32 AM

@JosephWright yes I thought I remembered a C library but I think perhaps just icu and harfbuzz is all I've seen in that general area

@JosephWright do you know its status, luatex links to luaforge.net/projects/sln but that just gives sln.luaforge.net which is 404 these days

1 hour later…

11:40 AM

@DavidCarlisle , @ShreevatsaR -- that's not the way it shows up in the side-by-side view on my screen. (moat backslashes are single, even the ones that should be double; only the one within \rhead{\\...} shows up as double.) don't tell me that the wretched double-backslash bug introduced by "the powers" has returned?!?

@ShreevatsaR -- i've just checked the view that i reported, and indeed, the original (shaded in red) shows just \rhead{\...} while the edit (shaded in green) shows \rhead{\\...}. that means what is being sent to my browser isn't the same as what's being sent to others. this is a disaster!

@ShreevatsaR -- the fact that "inline" and "side-by-side" don't show up with identical information to :side-by-side markdown" is definitely indication of a bug on the site. @JosephWright -- can you please look into this? (it is a most unpleasant prospect to wake up to.)

11:59 AM

@barbarabeeton you must be looking at the rendered view not the souce view. If you do not mark the document as a code section \\ just appears as \ but once you indent by four spaces the \\ appear,

@barbarabeeton I din't think it's a but just markdown conventions \` v \\ ` (which has four \ :-)

@barbarabeeton what you describe is "side by side" view, but there is a second "side by side markdown" view. Unfortunately the buttons juts affect the page view they do not change the revision url so you can not link to a specific view

user image

@DavidCarlisle -- okay, but you have to admit that it's damnably confusing ... i'm concluding that (la)tex is rather a stepchild compared to other sx sites. that's a disappointment.

@barbarabeeton it's just showing what is rendered not different than if the op had put _hello_ then in the side by side view you would see hello and in the side by side markdown view you would see _hello_

Harald Hanche-Olsen

@barbarabeeton I think the primary lesson is that escape conventions are evil.

@barbarabeeton what is confusing is not the revision diff view but the general handling of \ in markdown, even here I have to quote backslash space as getting a single quoted backslash in chat is hard (or impossible, I forget)

12:14 PM

@HaraldHanche-Olsen -- perhaps, but without an escape facility, tex wouldn't exist. i've worked with several other "coded" typesetting systems, and believe me, tex is the only one that is at all usable directly by mathematician authors. and now that they are forced to prepare their own manuscripts, ...

@barbarabeeton the mistake in tex is using \ as the default escape character. I find it works better to use j

Harald Hanche-Olsen

@barbarabeeton I only said it was evil, not that it isn't necessary (sometimes). The problem here is that markdown also uses the backslash as an escape character.

@DavidCarlisle -- well, yes. but your name isn't "john" or "jules".

Harald Hanche-Olsen

@barbarabeeton Practically everything under the sun uses the backslash for escape character. That creates huge usability problems. (I see no good solution to it, though.)

@HaraldHanche-Olsen -- which came first, tex or markdown?

Harald Hanche-Olsen

12:19 PM

@barbarabeeton Indeed. But other systems have also used the backslash, long before TeX even. Think of the unix shell and the C programming language. And unfortunately, the list of usable escape characters isn't infinite. Or even very long.

@barbarabeeton NORMAL PEOPLE WRITE John NOT john :-)

@DavidCarlisle -- just so. nonetheless, ...

@HaraldHanche-Olsen -- just be glad the tex character set didn't remain tied to the keyboard on the sail computer -- the circled plus was the original alignment character. doesn't exist on any other computer (or typewriter) keyboard that i know of. yes, the keyboard arrangement is the limiting feature here. which is why japanese tex uses the yen symbol as the escape character.

Harald Hanche-Olsen

@barbarabeeton At the University of Oslo ages ago, the forward slash was commonly used as the escape character in TeX. This was in the 7-bit days, when Norwegian computer users used a bastardized version of ASCII in which you got the Norwegian letters æøå and their uppercase variants ÆØÅ by typing {|} and [\]. (No I don't know what characters they used in TeX instead of the curly braces. This is all hearsay, but from reasonably informed sources.)

@HaraldHanche-Olsen -- \bgroup and \egroup were included in tex at least partly to accommodate "deficient" keyboards. (i think that's documented in the texbook. try appendix b.)

Harald Hanche-Olsen

12:38 PM

@barbarabeeton All I find is this:

Finally,
^|\bgroup| and ^|\egroup| are made to provide ``implicit'' grouping characters
that turn out to be especially useful in macro definitions. \ (See Chapters
24--26 and Appendix~D for information about ^{implicit characters}.)

@barbarabeeton Maybe they just used regular parentheses instead of curly braces. With an escape convention of /(/) to get ().

@HaraldHanche-Olsen -- okay, thanks for looking it up. (these two can be used individually in macro definitions without complaint whereas actual braces can't. but the "internal" meaning is just the same as typed braces. i know that \sp and \sb are provided for use where ^ and _ are lacking from keyboards,)

3 hours later…

3:50 PM

@barbarabeeton I remember seeing somewhere that in the earliest version of TeX, the very first character it encountered would be treated as the escape character. So if you began your file with \input basic then the backslash was the escape as usual, but if you had say ¡ on your keyboard and wanted to use it as the escape character, then you could start with ¡input basic say.

This feature was removed later because people would start TeX, type the name of their file (say paper) at the ** prompt, and get something like:

**paper
! Undefined control sequence.
<*> pa
      per
?

4:27 PM

@ShreevatsaR -- i don't remember this, but somewhere at home i have a copy of the very first tex manual, and when i get back there, will try to check. i can believe that someone as well disciplined as knuth could manage such a narrow restriction, but for mere mortals, ... a bit too dangerous.

4:44 PM

@StrongBad There was an earlier one in linguistics, which actually had a real point, i.e. it wasn't a joke, although obviously humorous. realclearscience.com/blog/2015/12/…

1 hour later…

6:09 PM

@egreg you were a few seconds faster:-) (but only because I'd stopped to answer side questions in a comment under the question)

@DavidCarlisle Rare to see such a big amount of useless tokens

6:35 PM

You do really love each other, don't you? :)

2

6:46 PM

The Odd Couple (film)

The Odd Couple is a 1968 American comedy Technicolor film in Panavision, written by Neil Simon, based on his play of the same name, produced by Howard W. Koch and directed by Gene Saks, and starring Jack Lemmon and Walter Matthau. It is the story of two divorced men - neurotic neat-freak Felix Ungar and fun-loving slob Oscar Madison - who decide to live together, even though their personalities clash. The film was successful with critics and audiences, grossing over $44.5 million, making it the fourth highest-grossing picture of 1968. The success of the film was the basis for the ABC television...

Harald Hanche-Olsen

@AlanMunn There was also a paper a number of years ago in Annals of Improbably Research – not quite as serious obviously – entitled “Recent Advances in Artificial Intelligence”: The paper was an entire blank page (except for the title and list of authors).

@egreg Haha, spot on!

Harald Hanche-Olsen

@AlanMunn The journal's home page is here. Unfortunately, I can't find the paper by googling.

1 hour later…

7:58 PM

@HaraldHanche-Olsen -- the paper was apparently included in a "best of air" collection, as reported here: improbable.com/airchives/press/1997/globeandmail-111597.html that implies that it appeared no later than 1997. (too bad this question didn't come up before the last "first annual ig nobel awards celebration" -- the book table there always has copies of these collections for perusal and sale.)

Harald Hanche-Olsen

8:17 PM

@barbarabeeton Well, I have almost the entire set of AIR issues in my office – until they moved off paper, except the final year and perhaps one or two issues. So I should be able to find the paper. The collection is in my office, though. I'll check it out tomorrow, perhaps.

9:00 PM

@mickep ask @egreg where the tick went

2 hours later…

10:30 PM

@JosephWright speaking of kerns... did you understand what Hans was suggesting we did for the kern box logging?

@DavidCarlisle He said to me he'll adjust the code to always put the type of kern in the tracing output. I think he meant 'use the burning edge LuaTeX'

@JosephWright oh. I'll do an svn update and see what happens (I knew there was a reason to build in experimental rather than in the tagged branch for tl2017:-)

@DavidCarlisle I've not checked if there is anything in the SVN ...

$ svn up
Updating '.':
D    source/libs/lua53/TLpatches/_patch-02-FreeBSD
D    source/libs/lua53/TLpatches/_patch-01-utf-8
D    source/libs/lua53/TLpatches/_patch-03-export
U    source/auxdir/auxsub/Makefile.in
U    source/auxdir/auxsub/aclocal.m4
U    source/auxdir/auxsub/configure
U    source/doc/Makefile.in
U    source/libs/gmp/include/Makefile.in
U    source/libs/gmp/native/Makefile.in
U    source/libs/gmp/native/aclocal.m4
U    source/libs/gmp/native/configure
U    source/libs/gmp/Makefile.in

@JosephWright changes since yesterday ^^^ :-)

@DavidCarlisle No idea then

10:34 PM

@JosephWright I just did that update now, I'll recompile and see what happens.

@DavidCarlisle Cool

10:59 PM

@JosephWright mail coming up...

11:34 PM

r6558 | luigi.scarso@gmail.com | 2018-01-19 11:57:04 +0000 (Fri, 19 Jan 2018) | 1 line

more granularity in kern reporting (HH)

« first day (2638 days earlier) ← previous day next day → last day (2296 days later) »

Transcript for

Jan '1820

$TeX - LaTeX$ TeX, LaTeX and Friends

\begin{chat}

join this room
about this room

00:00

06:00

12:00

18:00

all times are UTC

$TeX - LaTeX$

site design / logo © 2024 Stack Exchange Inc; legal

mobile