TeX, LaTeX and Friends - 2019-02-17 (page 1 of 2)

manooooh

00:09

@DavidCarlisle ok, thanks! So, is there any solution?

marmot

03:58

@barbarabeeton Yes, I do not doubt that their intention is good. Yet I had errors introduced in my papers, they sent us the proofs with 24h time to make changes while all of us were at a conference. And the changes are subtle, and they do not always specify all the changes, or the important changes are buried under a pile of unimportant ones.

Faheem Mitha

06:37

@DavidCarlisle You mean zzz as in the literal ascii string? No, I mean actual non-ascii Unicode. Though my memory is blurry.

@marmot Yes, the kind of maths involved in that is certainly not easy or trivial.

According to a friend I knew, it is, as he put it, so complicated that you become wearied of it. Or something like that.

I took a course in algebraic number theory once. I could sort of see why people liked that stuff, though it didn't particularly appeal to me.

But it was also enough number theory for me.

David Carlisle

08:49

@FaheemMitha yes I know that's what you meant but really it's the same. No font covers the whole of Unicode, so if you have a font that covers the ascii range,zzz "just works" and if you have a font that covers Chinese then 你好，世界 "just works" and if your font doesn't cover either range, then neither works. For pdftex you need some macro setup in addition to the fonts but the same principle applies.

Sebastiano

@DavidCarlisle Good morning and a good sunday.

@Kurt Dear Kurt, I simply deleted my question, with all sincerity, because I did not want it to be a duplicate of some previous question. If you go to my favorites I added the two links suggested in the comments. Thank you always and for me you are always welcome.

Kurt

09:04

@Sebastiano Ah, I see. I started to write an answer and suddenly the question was gone ...

egreg

8

CarLaTeX

@egreg … and also a correct comma position!

Faheem Mitha

09:24

@DavidCarlisle Yes, I see. I suppose Unicode is too big to be covered by a single font.

Though I've never understood really what makes a font. As opposed to a bunch of different letters. An overarching similar style?

David Carlisle

@FaheemMitha it's a work of human construction, Knuth made a font that just had the letters in MetFont for the metafont logo that needed to be extended when metapost came along. A font designer can put any characters he or she cares to design into a font. Although for pdftex no font can have more than 256 characters so it takes a lot of fonts to cover unicode even in a single style.

Faheem Mitha

09:47

@DavidCarlisle Is that 256 character limit for a pdftex font hardwired? And I suppose that must be inherited from TeX.

David Carlisle

@FaheemMitha yes completely pervasive throughout the entire system and syntax of tex.

Faheem Mitha

@DavidCarlisle I see. Do the newer variations like xelatex have a higher limit?

David Carlisle

@FaheemMitha yes to 1114111

actually one more that that, as counting starts at 0, and arbitrarily many glyphs if you count unencoded glyphs that can be accessed not by number.

Joseph Wright

19

Q: How can (La)TeX read UTF-8?

As described in The TeXbook, TeX reads files byte by byte, regardless of the particular format -- as I understand, this is just how IniTeX is set up. I also understand that LaTeX is just a collection of macros built on top of IniTeX, described in most distributions of TeX by the file latex.ltx....

unicode

Paulo Cereda

10:21

@egreg ooh

@TeXnician expect a lot of news today... sorry for the delay. :)

TeXnician

@PauloCereda Ooh :)

Paulo Cereda

@TeXnician ooh

David Carlisle

@PauloCereda don't encourage him, he's on a 7-2 cycle until someone downvotes him so he'll get a palindrome a week, ignore it unless he gets 777,777

Paulo Cereda

@DavidCarlisle ooh a plan

David Carlisle

@PauloCereda which reminds me, I need to answer some vim questions with emacs answers if I'm going to get 555,555

Faheem Mitha

10:39

@DavidCarlisle Hopefully that is enough. Though I'm unclear why limits need to be hardwired at all. Some hardwired memory allocation issue?

That's generally not necessary with mainstream languages. At any rate scripting ones. But I realise that TeX is a macro language, so different rules may apply.

Joseph Wright

@FaheemMitha There are only a certain number of valid Unicode code points. Everything else is not allowed. So it makes sense to program that in.

Faheem Mitha

@JosephWright So it's not likely another system will come along in the forseeable future?

Joseph Wright

@FaheemMitha It seems extremely unlikely. There are of course quirks in Unicode, but broadly it does what it was designed to do.

Faheem Mitha

@JosephWright Ok.

Paulo Cereda

@DavidCarlisle ooh

David Carlisle

10:44

@FaheemMitha classic tex does no dynamic memory allocation so everything is limited, but it is also built into the syntax, something like \mathchardef\alpha="010B bit-packs the font to use, whether it is an ordinary symbol or a relation or binary operator, and the character number of the character into a single integer, that only works if you know how many bits are in each field so: at most 16 fonts per math expression and at most 256 characters per font

Faheem Mitha

@DavidCarlisle All the above applies to "classic" tex?

So, a form of hashing?

David Carlisle

@FaheemMitha yes which is why xetex/luatex need new syntax (\Umathchardef) with a different syntax they can not just extend the range (unlike \char where they simply say it takes a 32 bit integer not a 8 bit one)

@FaheemMitha not really hashing that would imply mapping a big set down to a smaller hash) it is simply integer indexing a fixed size array.

CarLaTeX

@marmot I edited my question: tex.stackexchange.com/questions/475298/… is it clearer now?

Paulo Cereda

@DavidCarlisle good old days when people cared about this stuff. :)

David Carlisle

@FaheemMitha the main reason for the limit on Unicode being less than 2^32 is UTF-16, many languages assumed that 16 bits per character would always be enough and when Unicode extended beyond "FFFF by restricting to "10FFFF you can use surrogate pairs to encode most of the original range in 1 16bit integer and the whole range in 2 such units. If Unicode extended back to its original range UTF-16 would need to change to allow three or four unit encodings and Java etc would have problems

Paulo Cereda

10:57

@DavidCarlisle we could have fun with the wobbly variant with unpaired surrogates. :)

David Carlisle

@PauloCereda all input should be xml and there are no unpaired surrogates in xml (it's a fatal error)

Paulo Cereda

@DavidCarlisle ooh that would be good :)

David Carlisle

@PauloCereda xml declares those characters to be illegal so you can't even access them from another encoding such as utf-8.

Paulo Cereda

@DavidCarlisle got it

David Carlisle

@PauloCereda they should have declared java and vim illegal at the same time, and made everyone use xslt and emacs.

Paulo Cereda

11:04

@DavidCarlisle you are mean

David Carlisle

Jun 29 '17 at 16:15, by Paulo Cereda

@DavidCarlisle you are not mean :)

Paulo Cereda

@DavidCarlisle oh

Harald Hanche-Olsen

Apr 10 '18 at 15:43, by egreg

1 min ago, by Harald Hanche-Olsen

Feb 16 at 13:58, by Harald Hanche-Olsen

Jan 17 at 14:43, by David Carlisle

Jan 12 at 13:02, by Harald Hanche-Olsen

Nov 8 '17 at 10:40, by Harald Hanche-Olsen

Oct 17 at 6:51, by David Carlisle

24 hours ago, by Harald Hanche-Olsen

Oct 11 at 16:46, by David Carlisle

Sep 10 at 12:03, by David Carlisle

@HaraldHanche-Olsen do you ever get a feeling of déjà vu in this chat room?

Paulo Cereda

ooh a triangle

Harald Hanche-Olsen

@PauloCereda chat.stackexchange.com/search?q=ooh&user=6485&room=41

Paulo Cereda

11:08

@HaraldHanche-Olsen ooh a counter

Harald Hanche-Olsen

@PauloCereda clicking away as we you speak

marmot

11:28

@CarLaTeX Yes, it is clearer but I would like to argue that the title may not really be 100% appropriate. Try
\begin{tikzpicture}[baseline, xscale=#1]
\draw[black, {Round Cap}-{To[angle=90:.3cm 1,length=9mm, flex'=.86]},
line width=4pt]
(.1,.85) to[bend right]
(1.48,.8)
(1.43,1.4) to[bend right] (2.1,2.5);
\node[red, circle, draw, text width=10pt] at (.1,.85) {};
\node[blue, circle, draw, text width=10pt] at (1.43,1.4) {};
\end{tikzpicture}
This has the precisely same behavior meaning that the arrow heads are attached at the last stretch. Therefore I would say that "does not work" is not an…

@CarLaTeX And @TeXnician provided you with a very nice way to get what you want, and for a standard TikZ path with a gap you can just copy this nice method.

Faheem Mitha

@DavidCarlisle Some of that is over my head. Or at least, I don't follow it. But I don't get the referenc to 2^32. Earlier you said 1114111.

@DavidCarlisle Yes, I see.

David Carlisle

@FaheemMitha the "obvious" UTF32 encoding of Unicode isn't used much and the range is restricted to the range that can be encoded using two UTF-16 units. 1114111 is hex 10FFFF that is 17 plains of 2^16

Faheem Mitha

@DavidCarlisle Hence 1114111?

David Carlisle

@FaheemMitha yes hex 10FFFF

Faheem Mitha

@DavidCarlisle plains?

David Carlisle

11:43

@FaheemMitha just terminology, BMP= "basic multilingual plane" = "Unicode 1 or 2" = Unicode 0-hex FFFF. They still try to fit "normal" characters in that range, with the higher planes being used for more exotic things (originally because many applications could not access them, although that is less of an issue now)

Faheem Mitha

@DavidCarlisle So bins, basically?

David Carlisle

@FaheemMitha so for math, all the symbols and arrows etc are in the BMP, usually around hex 2000 but the styled math alphabets like fraktur, blackboard bold etc are in the first extended plane at hex 1D4xxx

Faheem Mitha

@DavidCarlisle I see.

David Carlisle

@FaheemMitha sorry my spelling it's plane not plain

Faheem Mitha

@DavidCarlisle Yes, plane. Got it.

Jasper

12:00

@egreg I wonder what will happen when you get 1,000,000 points. Maybe SE will give you a million dollars, LOL.

Harald Hanche-Olsen

@Jasper He will get no more palindromes with a correctly placed comma for a very long time.

3

Jasper

@HaraldHanche-Olsen This is the best answer on TeX SE.

vlg

hi all, this doesn't warrant a thread- I'm unsure what if anything i've changed in a document, but \def assignments give me missing = inserted for \ifnum and \missing number, treated as zero although it's there along the lines of \def\myc{73}, log file gives me nothing to go off of

Checked every instance of where the \defs are used, as well all \ifnums in the document and the sty

Harald Hanche-Olsen

@vlg Pretty hard to tell from that description. You need to gather more information. I think the first step is to put \errorcontextlines=99.

vlg

never used it before, in preamble or?

Harald Hanche-Olsen

12:16

@vlg Preamble is fine, yes.

It will give you a traceback of what macros are being expanded at the moment when the error happens.

LaTeX puts it equal to 1, because it is felt that the resulting tracebacks will just confuse most people.

vlg

Well, it gives the same error, just the that line this time. maybe putting it after the def's would be more prudent?

The output is pretty much the same, unless i'm looking in the wrong place in the log file

David Carlisle

@vlg I would assume that you have redefined a count register to be a macro, not a good idea:-)

vlg

the defs are number literals and 2 strings, I understand I've mucked up something, but don't think I know how to redefine a count register

David Carlisle

@vlg your description is wrong, the \def assignment will never give a missing number error, a count assignment will give that error if you have redefined the token not to be a count using \def. I

@vlg \newcount\myc good then \def\myc{2} bad.

vlg

The error is right at the counter assignment in the logbook

but these are constants, not counters, there not changed at any point

David Carlisle

12:25

@vlg impossible to debug riddles, post an example.

@vlg of course \def isn't a latex command. so the problem is never supposed to occur.

vlg

example of the logbook:
`\c@dataSet=\count331
)
! Missing = inserted for \ifnum.
<to be read again>
a
l.13 ^^I\def
\mya{0.75}
I was expecting to see `<', `=', or `>'. Didn't.

! Missing number, treated as zero.
<to be read again>
a
l.13 ^^I\def
\mya{0.75}
A number should have been here; I inserted `0'.
(If you can't figure out why I needed to see a number,
look up `weird error' in the index to The TeXbook.)

\c@yy=\count332
\c@mm=\count333`

David Carlisle

@vlg post a test file that reproduces the error:-)

vlg

making this impossible, but thanks for the effort

David Carlisle

@vlg what is impossible? on line 13 you have a tab and a \def but the def isn't that relevant it is whatever you have not shown on line 12 is an \ifnum so it expected to find a number and it found that def.

vlg

Above it are my package, geomtery and \documentclass

David Carlisle

12:33

@vlg why can you not post an example?????, here for example is a document making exactly that error:

\documentclass{article}


\ifnum
\def\mya{0.75}
\begin{document}

\end{document}

vlg

because that works. so providing a minimal working example would me more effort on my part

`\def` work is the preamble like one'd expect, returns the proper thing too

David Carlisle

@vlg what works?

@vlg if you don't have an explicit \ifnum as I showed in the above example then you most likely have a mal-formed date in the optional argument to \usepackage or related commands.

vlg

mal-formed date?

WHY

I removed the day and the hyphen in the optional arg in the packages, and that has this effect?
Never changing the date ever again

Harald Hanche-Olsen

That argument is supposed to have a very specific syntax: 2019-02-17 or 2019/02/17.

Stick to either of those, and you should be fine.

David Carlisle

@vlg as I say if you use an incorrect date format you get an error parsing the date, but why didn't you show your input which would have made debugging that so much easier?

vlg

12:39

But what exactly is it checking? Why stop the entire build if the optional package date is out of format?

Harald Hanche-Olsen

@vlg Because doing a full parse with error recovery is a lot of hard work?

David Carlisle

@vlg it is checking the date which you optionally may put there to have checked.

Harald Hanche-Olsen

Easier to just use number assignment to get at the component numbers.

@vlg LaTeX is full of places where malformed input leads to strange and confusing errors. Due to the way it is built on top of TeX, that is an extremely hard problem to cure.

vlg

'Component numbers'? could you please explain what you mean

David Carlisle

@vlg actually I can't get that error in usepackage or do you mean you edited the .sty file? This is silently ignored:

\documentclass{article}


\usepackage{geometry}[12 /abad date ]
\def\mya{0.75}

\begin{document}

\end{document}

Harald Hanche-Olsen

12:42

@vlg Today's date, 2019-02-17, consists of three numbers: 2019, 02, and 17. The code will just assign each of them to a counter.

David Carlisle

@vlg where did you put the date in the wrong format?

vlg

my .sty had \ProvidesPackage{dmlb}[2019-01 yaddayadda], i removed the date because I thought i wouldn't want to change it at every little thing

David Carlisle

@vlg Finally you show some code:-) That is a syntax error.

vlg

I remember reading that there's a date format, but it never said it was obligatory to get right, I think

Harald Hanche-Olsen

@vlg Well now you know. 8-)

David Carlisle

12:45

@vlg yes it does if you read clsguide or any other official documentation.

vlg

he knows them by heart T_T

David Carlisle

@vlg well it's my code as it happens:-)

vlg

I'll concede, grandwizard

Harald Hanche-Olsen

@vlg @DavidCarlisle makes a living writing documentation. No wonder he gets a bit sore of nobody reads it.

vlg

well, thank you both

David Carlisle

12:46

@HaraldHanche-Olsen actually they don't trust me to write much (something about not wanting "teh" in the manual) but I do make a living generating code out of documentation:-)

Harald Hanche-Olsen

@DavidCarlisle Good thing that programming languages are fine with variable names beginning with “teh”, then.

David Carlisle

@HaraldHanche-Olsen and compilers helpfully mention if you declare the and use teh....

CarLaTeX

13:05

@marmot I have modified the title. @TeXnician's solution is good but I think it's "an hack". I'll try yours asap.

Skillmon

13:19

where is the site containing the frequently used comments, such as the "Welcome to TeX.SX" message?

Harald Hanche-Olsen

@Skillmon I'm not sure there is one. I guess frequent contributors have their own list of stock phrases that they use.

Skillmon

@HaraldHanche-Olsen there certainly is one on meta, I used to have it in a pinned tab, but lost my firefox session last week :(

CarLaTeX

180

Q: Text building blocks

There are some replies that are used quite often. For example, the first reply to many questions is a demand for a minimal example. These replies should typically include a link with additional information. So I thought that it might be useful to collect some standard replies for quick copy&paste...

discussion comments new-users big-list

Kazi bácsi

Hi guys! I'd need to add a few packages, but my MikTeX Portable can't connect to the server. Do you have similar issues? Just don't know where to start...

Skillmon

@CarLaTeX thank you very much, that's the one I was searching!

CarLaTeX

13:23

@Skillmon You're welcome!

vlg

13:49

I'm unsure as to whether \global\count400=\dimexpr\wd0\relax this is correct, or of the global should be before the assignment, i.e., the =, or elsewhere entirely?
It works fine before \advance.

egreg

@vlg The position is right, but you should never use an explicit counter number: you can't know whether that counter has been allocated by some package.

vlg

Can I use all the \count functions with counters? Like the assignment was really easy, advancement and the stuff like '\divide\count400 by\count401'
I know only of stepcounter, and another package let you use arithmetic expression to assign new values or add, but i thought they were integers only; the divide is floor division, so it's not really dealing with floats

barbara beeton

14:06

@StefanKottwitz -- That's attractive. But "greatfully"? They could be helped by a spelling/grammar checker (if a good one really exists).

barbara beeton

14:18

@DavidCarlisle -- But potentially a font could contain more characters than Unicode. In particular, there are "variant selectors" that support alternate shapes for a character with the "same meaning". And this is not just for math characters.

David Carlisle

@barbarabeeton yes I think I mentioned that somewhere in the thread above, even without the variant selectors they can (and do) also have unencoded glyph variants that you can select by name not by unicode number

@vlg it's syntactically correct but rather weird. You should not access count registers by number and you are assigning a length to a count so it will take the value in sp. the \dimexpr is doing nothing as there is no expression.

barbara beeton

@DavidCarlisle -- And there's also the "unpoliced" PUA. (Yes, you did mention the access by name.) So potentially a "full Unicode font" can be an unspecified number of times larger than Unicode itself. (Granted, unlikely, but ...)

David Carlisle

@vlg you are talking about different things, \stepcounter takes the name of a latex counter allocated by \newcounter

vlg

forgot about \newcount\<name>; i think that would solve possible allocation/override issues;

David Carlisle

@barbarabeeton no, the PUA are unicode slots (just with undefined by Unicode meaning) so they count in the 10FFFF

@vlg \newcounter isn't the same as \newcount.

vlg

14:25

I need to access and modify the values, and the default latex counters seemed unwieldy at the time, so i read how \count works and experimented until it worked, currently am having an issue where something isn't working because it's not global, is my assumption at least

\advance doesn't work inside a longtable env, so i'm straight up putting global everywhere to see what works, but up no luck so far

David Carlisle

@vlg it depends if you want to fit in with latex conventions or cut across them. for example latex allocated counters will work correctly with \include or in a tabularx and other places. But a primitive counter allocated via \newcount will not (and direct access via \count300 is almost certain to break something.

@vlg \advance will work in longtable just as well as anywhere else. It is a local assignment though like \setcounter or \nescommand so restored at the end of a group.

vlg

I made a MWE, showing how how it didn't work

David Carlisle

@vlg it may not work, but not because of longtable for example \count0=1 {\advance\count0 2 } \the\count0 will produce 1 not 3 as the adding 2 is local and restored at the }

vlg

counter created outside, and a command the takes 3 arguments and makes a row in a longtable out of them; #1 & #2 & #3 &
and in the fourth column was a small thing that checked if #3 is blank, if yes, increment counter, if not don't

David Carlisle

@vlg table columns are groups so work like the {...} in the comment above (this applies to tabular or align or anything else using & not just longtable)

vlg

14:32

I wasn't blaming your glorious package, I merely meant it was an env.

David Carlisle

@vlg nothing to do with it being an environment it is table (\halign) cells that are groups.

Jasper

Hello @DavidCarlisle I am wondering whether you still look like the man in the picture. =)

David Carlisle

@Jasper possibly more so than you look like your avatar

vlg

burned

Jasper

It is very strange that when I visited the texniccenter website a few days ago, I got some popup advertisement. Seems something wrong happened there.

David Carlisle

14:36

@Jasper but @PauloCereda looks exactly like his picture.

Jasper

But there has been no news on texniccenter website since 2014. Hmm, I guess development has stopped.

David Carlisle

@Jasper it is not emacs, what other news do you need about an editor?

2

Jasper

@DavidCarlisle Haha, I have no idea what is so good about emacs. I have never used it.

Paulo Cereda

@DavidCarlisle quack <3

Jasper

Hello @PauloCereda quack! I no longer use Linux. I am using Windows 10 only now.

David Carlisle

14:40

@Jasper I have used it most days since 1987, on a variety of operating systems.

Paulo Cereda

@Jasper Hi! Oh no

Jasper

@DavidCarlisle Ah, I see. That makes you a very old man. But I am also an old man now, though not as old as you.

David Carlisle

@PauloCereda windows 10 isn't so bad as long as you install X to get a real window system:-)

Jasper

@PauloCereda Maybe I will switch to macOS in a decade.

barbara beeton

@DavidCarlisle -- What i meant is that different users assign the same PUA locations to multiple different meanings that conflict with one another. So it's not guaranteed that trying to use multiple PUA-encoded fonts will yield the desired results. (In fact, almost guaranteed that it won't. That spelled disaster for the desired cleanup of the original STIX table, which I regret.)

Paulo Cereda

14:43

@DavidCarlisle :)

David Carlisle

@barbarabeeton yes sure there is that.

Jonas Stein

15:21

Hi. Why does acro use - \c_one instead of -1? bitbucket.org/cgnieder/acro/commits/94c582b#Lacro.styT2998

If I understand tex.stackexchange.com/a/470147/16227 correct, we could simply write -1 too.

Jasper

@DavidCarlisle I just visited the emacs website. It is frightening it has so many features.

Jonas Stein

@Jasper yes, but it lacks of the default features a good editor has. If you want line numbers, you need slow and broken lisp scripts.

Harald Hanche-Olsen

@DavidCarlisle Looks like you beat me by a couple years. The earliest solid evidence I can find that I was using emacs, dates from 1990. I remember using the beta release of version 19, could it have been around 1993? Emacs 18.57 was such a bad X11 client. Version 19 was a giant step forward.

@Jasper But you only ever need a few of those features. Your repertoire will grow with time. It's a bit like latex: Who ever needs to use thousands of style files anyhow?

David Carlisle

15:42

@HaraldHanche-Olsen that early I wasn't using X so much I had sunview and the emacsview sunview windows client, it took a while for the X version to catch up:-)

@HaraldHanche-Olsen really you only really need M-x doctor

Harald Hanche-Olsen

@DavidCarlisle What about M-x dissociated-press ?

David Carlisle

@HaraldHanche-Olsen that has its uses, but I think the doctor is most useful for people in this chat.

Harald Hanche-Olsen

@DavidCarlisle Maybe we could make a doctor bot and turn it loose on psmith?

manooooh

16:56

Hello!!

I saw this first on a spanish page and then the original reference in english: An Efficiency Comparison of Document Preparation Systems Used in Academic Research and Development

What do you think about the conclusions?

Paulo Cereda

@manooooh I don't like bold. :)

Joseph Wright

@manooooh Oh, that one

Paulo Cereda

@JosephWright We've been there, don't we :)

TeXnician

@manooooh More importantly: What does this say about German researchers? Especially about those claiming to be LaTeX experts?

Paulo Cereda

World would be a lot better if Brazilians migrate from abntex2 to @Joseph's abnt, which has a toucan!

vlg

17:10

Is having the same \numexpr consisting of simple addition/subtraction of a 2 counters and one number, three times in a single command unsightly? It it best practice to use an \edef once at the start?

TeXnician

@PauloCereda But as we just mentioned this nice paper: Does a toucan outweigh that "LaTeX does not offer an internal text editor" :D

marmot

@CarLaTeX Thanks for editing the title of the question!

Paulo Cereda

@TeXnician hmmm I see what you did there

Ulrike Fischer

17:44

8

@marmot ^^^ The marmot sees the Alps for the first time.

marmot

@UlrikeFischer She must be very happy! Did you have honey liquor to celebrate the moment? ;-)

Paulo Cereda

@UlrikeFischer ooh

Jasper

@manooooh Sounds about right to me. LaTeX is better for lots of math, otherwise Word is better. I am a big fan of Word, LOL.

I realise that I am saying this in the TeX chat and not the Microsoft chat, LOL.

CarLaTeX

@marmot You're welcome! However, also your answer is a workaround. I don't think the TikZ behavior is entirely correct. The draw is only one, {Round Cap}-{...} should round the very beginning of the path, whereas line cap=round acts correctly, rounding everything.

Jasper

@UlrikeFischer It is also the first time I am seeing the Alps and knowing I am seeing it, even if it is just a photograph.

Haha, everyone uses tikz these days. So few people use pstricks...

marmot

18:00

@CarLaTeX Well, this is the predefined tikz behavior. I can see why it is what it is and also why the author of tikz designed it that way: in standard tikz, you are supposed to use edges and the like if you really want to multiple arrow heads in a path.

Jasper

I am glad that tikz are not ticks!

CarLaTeX

@marmot So would the correct solution be to create two paths?

@Jasper Try to manage a relatively big Word document with a couple of math formulae and you'll change your mind

marmot

@CarLaTeX I do not think this would fly here because the cool thing about this Hobby feature is that if just skips a part from an otherwise smooth path, isn't it?

CarLaTeX

@marmot Yes, that's a cool feature but you can't solve the problem otherwise, without hacking I mean (also my/@JouleV's original solution is a workaround)

marmot

18:16

@CarLaTeX I just look at this post right now. I don't know if there is another solution, most likely there is. I took your new question as a sort of a bug report (because it sounded like one) and all I want to say is that IMHO there is no bug.

CarLaTeX

@marmot It depends if you intend the two parts (before and after the gap) as a unique path (as I do) or as two paths (as you and @TeXnician do).

marmot

@CarLaTeX TikZ does not know our intentions. There is just a standard behavior of TikZ how it adds arrows. You may now like it or dislike it, this is the standard behavior. And luckily TikZ is powerful enough to grant you the desired outcome with very little effort. My only point is that the standard TikZ behavior is not a bug, this is how it behaves and there is no place in the manual (AFAIK) that suggests otherwise.

David Carlisle

19:28

@vlg why edef not setting a counter?

manooooh

20:11

@Jasper HAHAHA

Supposedly they have proven that much more time is required to learn LaTeX than MS Word (perhaps because of inheritance, it does not matter), and this is more important than the quality of the documents. In fact, a mathematician should not be interested in the conventions or the format of his document, but should be concerned about doing mathematics

@PauloCereda @JosephWright @TeXnician ^^^^^^ what do you have to say about it?

mickep

@manooooh Well, independent of Word/LaTeX, it is difficult to read the text if it is not well set.

manooooh

It really goes against all of us, even me, that I have spent more time learning LaTeX than improving my knowledge in MS Word hahaha. It sounds hard, but that document made me think

@mickep just see the graphs

I think that document concludes that LaTeX is, in short, for publishers, not for "ordinary people" (I am not an editor)

@barbarabeeton would you mind to comment on this document, please? I am interested in knowing your opinion, I want you to make me forget MS Word but that document seems to be right!!

David Carlisle

@manooooh it's the same document that was posted here at the time isn't it? Not really anything to say about it.

manooooh

@DavidCarlisle oh, I did not know that it was discussed here, I am sorry. Why do you say that?

I think it hurts us all to see those statistics, but we have to face them

David Carlisle

@manooooh why? what's the point in being more efficient at generating something you don't want?

Joseph Wright

20:22

@manooooh They ar pretty meaningless

David Carlisle

@manooooh even if you accept the (rather dubious) notion of efficiency.

manooooh

@DavidCarlisle I am not sure I understand you well, but the point is that for people who know Word (which are practically all, at least here in Argentina) it is much more practical and economical to make the documents that instead of learning to do the same document but with LaTeX (for example, the page format with colors and all that is much more difficult to modify in LaTeX than in MS Word)

For example, we should not inculcate almost anything about formatting a mathematician, since he must deal with his trade and not with what an editor does. The same for a "normal person", who wants to make a document for his work, his university etc.

@JosephWright could I know why?

David Carlisle

@manooooh I've only tried to use word a couple of times but how long you take is rather missing the point. If you have a word document you have a binary blob of hard to manipulate data that is essentially useless, try reformatting the whole document to a different format, what's the equivalent of changing the documentclass to a journal style, or usimg tex4ht to generate html or processing a document from 20 years ago or ....

mickep

Best cutting-edge visualizations in world = scout around open source http://arXiv.org: math physics astronomy electrical engineering compsci quantitative biology statistics mathematical-finance. Amazing typography at "Feynman diagrams" Then to http://biorxiv.org

Tweeted by EdwardTufte on February 17, 2019 at 7:45 PM

As a funny coincidence this was just posted on twitter :)

manooooh

@DavidCarlisle maybe I'm still too young, but I've never had to use anything that is outside MS Word. I do not use external files, only images that stick as easily as CTRL+V

Joseph Wright

20:37

@manooooh The task given was to create an arbitary layout. Most LaTeX users don't do that very often. Neither do most Word users. But it's easier to hack a layout together in Word than LaTeX. I'm not at all surprised.

Paulo Cereda

@JosephWright Exactly.

Joseph Wright

@manooooh But, at least for publication, most end users should not worry about this: they should just author the text

@manooooh They could have picked a task that is much easier in LaTeX than Word, for example anything involving floats (in Word you have to do that by hand)

Paulo Cereda

@JosephWright And even that, there are people with the mindset of having full control of float positioning and resort to H, which defeats the very purpose of using floats in the first place. :)

manooooh

@JosephWright ok, I agree with you. What is the problem of MS Word users hacking a design (which by the way 90% uses the default ones)? That saves a lot of time, and you do not have to configure big things, since it's a very large assembly

What I'm complaining about is that MS Word is slow when it involves pages with many photographs or hand-made graphics. On the other hand, in LaTeX it only happens that the time that can be long only happens when you compile it, and not during the entire duration of the creation of the document. In that aspect, +1 for LaTeX

barbara beeton

20:57

@manooooh -- I wrote a critique of that article at the time, but unfortunately, Ive lost easy access to where I might have put it since my retirement a week ago. However, I believe it is highly biased, and based on phony premises. Serious mathematicians need facilities to express their work for publication, and Word simply cannot handle the notation and layout of much math past college level. Yes, (:a)TeX has a steep learning curve, but the handling of math expressions (cont'd)

Joseph Wright

@manooooh Well, it depends on what you are aiming to do. Mot of the time, authors should not really be messing with layout, at least for anything that goes to a publisher. For stuff you do yourself, it's very easy to waste time on layout.

barbara beeton

(cont'd) was designed by mathematicians for mathematicians. Using this properly is a way for mathematicians to record their ideas in a way that decreases the likelihood that a dedicated math publisher will have a harder time making hash of it. So yes, one can say that (La)TeX is for publishers, but its primary audience is mathematicians and other scientists who use that notation.

manooooh

@JosephWright yes, that is right

@barbarabeeton I agree with you in almost everything, but "Word simply can not handle the notation and design of many mathematics at the previous college level" is not true. I myself have used Word, and the equations are very easy to write. Also, all the symbols are there, there is no need to resort to any external source

barbara beeton

@manooooh-- If you wish, you can consider me biased. However, it takes me twice as long (or longer) to prepare a simple memo in Word than it does in TeX, and I make fewer mistakes in TeX, because it's based on plain text, and I can keep my fingers on the keyboard.

manooooh

@barbarabeeton wow, that is true

I did not noticed that. However, when I started using LaTeX around 1 year ago I spent more time in TeX.SE asking, reading and understanding that making the document. Surely this happens whenever you start using a new language, but I tell it to you because it is really the first time I spend so many hours to understand how to graph a function with TikZ or how to insert an image, etc.

barbara beeton

21:12

@manooooh -- Take a look at some AMS publications (ams.org/journals); I suggest the Proceedings or Transactions -- older than 5 years, all are "open". You will find many math displays that would be nightmares in Word. (Some of them are even nightmares on LaTeX, but they can be accomplished in finite time.)

Joseph Wright

@manooooh Er, that very much depends on what you work on

@manooooh My experience of doing simple 16-18 teaching of maths is that it's much harder to get right in Word

manooooh

@JosephWright yes of course. A lot of stuff in university can be achieved only using Word (we should change this)

@JosephWright in university I study integrals, derivatives, sequences... not much. You probably are in a math university, which is awesome

barbara beeton

@manooooh -- Graphics are often hard, and there may be tools that can achieve the desired results more quickly than TikZ, But many such tools are "professional", and require licenses. Mathematicians tend to be a "low-budget" crowd. On the past, it was sometimes said that their most expensive resource was chalk. However, their output was always considered "penalty copy" by compositors.

manooooh

@barbarabeeton I do not know why "their output was always considered "penalty copy" by compositors"

Paulo Cereda

@manooooh why we should change? If it works, it works.

manooooh

21:20

@barbarabeeton thank you! It seems very perfect <3. How can I enter to a journal? I clicked on e.g. "Transactions of the American Mathematical Society" but I do not see a link to that journal

Joseph Wright

@manooooh No, I'm a chemist. I teach some basic maths for people who've not done it before. But I've had 'fun' with simple stuff, getting it to look even vaguely right. For example, Word cannot do a nested itemize and have the markers line up :(

manooooh

@PauloCereda because they do not know the improvements that LaTeX can provide. Even I still can not create a document of much difficulty, that is, I do not take advantage of the full potential that LaTeX can offer me. Of course it works, but I consider documents made with Word old

Paulo Cereda

@manooooh But it works for them, so...

@manooooh Word/(La)TeX are mere tools, they are not part of the objectives.

manooooh

@JosephWright what is the problem of not having markers line up? I am not sure if I understand "Word cannot do a nested itemize" correctly, since there is an option for that:

Joseph Wright

\documentclass{article}
\begin{document}
\begin{itemize}
  \item \begin{itemize}
        \item Hello
        \item World
      \end{itemize}
  \item More
\end{itemize}
\end{document}

@manooooh ^^^ Try doing that

Paulo Cereda

21:27

@JosephWright you are ~~mean~~ naughty :)

barbara beeton

@manooooh -- "Penalty copy" is material that takes much longer and requires much greater skill to compose. Traditionally, much math copy had to be hand-set; it was beyond the capabilities of even the most functional composing machinery (before computers). I have compiled a list of publications dealing with math composition, and intend that it be posted to the web when I can again get access to my notes.

Paulo Cereda

@barbarabeeton Paulo Ney's talk on TUG 2018 about the composition of ICM's proceedings was quite amusing regarding the challenges of composing all math documents altogether (@JosephWright).

manooooh

@JosephWright ^^^^ again, what is the purpose of making · -?

Joseph Wright

@manooooh I'm pretty good at Word: it's part of my day job. So I know all about it's outliner :)

Paulo Cereda

@manooooh Did you see the output of @JosephWright's code, by the way?

They are not alike.

Joseph Wright

21:30

@manooooh See, doesn't line up

@manooooh It's more important when you do enumerates. Try for example doing sub-questions to a question.

Paulo Cereda

@JosephWright now you are being mean. :)

manooooh

@PauloCereda yes ^^^^

Paulo Cereda

@manooooh now you did because we told you. :)

manooooh

@JosephWright would a mathematician/chemist care about that? It seems to me that this is a very large tangent exit ...

@PauloCereda I saw it before I posted the Word's image

Joseph Wright

@manooooh It's pretty annoying when you are writing exam questions

@manooooh Makes it harder to read material

manooooh

21:32

@JosephWright ... for who?

Joseph Wright

@manooooh Undergraduates

manooooh

@JosephWright I do not think so. It is format, someone may like it more or less, but the essence is the content

murray

@ChristianHupfer: Can you help with tex.stackexchange.com/questions/475389/… ?

manooooh

@JosephWright I do not think so. I do not think they look at that. In Argentina nobody cares about the format (or few), everyone needs the information and not the format

barbara beeton

@manooooh -- On the page I mentioned, there is a "radio button" for Transactions. When you press that, a paragraph of text appears. In that paragraph, the name of the journal appears again -- click on that: it is a link. On the journal page, there is a link to "All issues: 1900-present". Choose one at random (2013 or earlier) and explore. Another journal with "interesting" content is Mathematics of Computation".

Joseph Wright

21:36

@manooooh It depends how much you care about formatting. It is important; try reading a paper draft in Word, then the final typeset version (in my case, that will be in 3B2)

\documentclass{article}
\begin{document}
\begin{enumerate}
  \item
    \begin{enumerate}
      \item First part question
      \item Second part question
    \end{enumerate}
  \item Some text which explains
    \begin{enumerate}
      \item Third part question
      \item Fourth part question
    \end{enumerate}
    \item Simple question
\end{enumerate}
\end{document}

manooooh

@JosephWright ^^^^^^ did you mean this? What problem do you think there is?

Joseph Wright

@manooooh When you have a question with sub-parts but no main text: can look like something is missing in Word. See my example above.

Ulrike Fischer

@manooooh you are cheating. Do it without the "Question 1" text.

manooooh

@UlrikeFischer I did not reproduce the comment of Joseph

Now yes

@JosephWright are you referring to that "missing" part?

Ulrike Fischer

@JosephWright can word do descriptions list now?

Joseph Wright

21:42

@manooooh That's the one

manooooh

@JosephWright I never in life would have thought that there could be missing text. And if the doubt is, why can not you ask the teacher?

If I am in a examen and the doubt appears, I will ask to my teacher. That's it

barbara beeton

@manooooh -- Try this Math of Comp article: ams.org/journals/mcom/2012-81-278/S0025-5718-2011-02526-3/… Chances are that his came in in pretty good shape. (AMS provides "author packages" that fashion the content to the style of the publication, so that if an author uses this, it avoids "surprises".)

Joseph Wright

@manooooh You can ask the invigilator, but they'll have to check with the person who set the test ....

@manooooh All sorts of mess-ups happen!

manooooh

@JosephWright the same as in LaTeX!!

@barbarabeeton oh, I will want to see that that compilation!! Thank you, you have been very clear

@barbarabeeton thanks! I can now see the publications. What do you wanted to show me?

@barbarabeeton let me know which section is difficult to reproduce and I will try to reproduce it in Word

P.S. I have never write the References section of a document. I have done it once, but it is slightly different from the AMS requirements

David Carlisle

21:58

@manooooh produce a document on word sharing authoring between co authors on linux, on a mac and on windows, style the preprints on A4 paper then reformat to a custom journal paper size, submit the tex document for submission to a conference proceedings, and then tell me how you'd do that in Word.

Paulo Cereda

@DavidCarlisle Ich sprache kein Englisch

And neither German, it seems :)

Transcript for

$TeX - LaTeX$ TeX, LaTeX and Friends