TeX, LaTeX and Friends

egreg

10:27

@DavidCarlisle Re Leslie's report: I guess his expectations about Bézier curves is wrong

\setlength{\unitlength}{5pt}
\begin{picture}(0,50)
%\qbezier(46,37)(30,26)(34,28.8)%bad
\color{red!20}\qbezier(46,37)(30,25)(34,28.8)%
\color{red!30}\qbezier(46,37)(30,25.2)(34,28.8)%
\color{red!40}\qbezier(46,37)(30,25.4)(34,28.8)%
\color{red!50}\qbezier(46,37)(30,25.6)(34,28.8)%
\color{red!60}\qbezier(46,37)(30,25.8)(34,28.8)%
\put(34,28.8){\circle{2}}%
\put(34,28.8){\circle{5}}%
\end{picture}

Joseph Wright

@egreg :)

egreg

@JosephWright That's just because I'm going to lecture and don't know how to answer on gnats

David Carlisle

@egreg Leslie? Oh I better read my email....

Joseph Wright

@DavidCarlisle I thought it was spam until @egreg mentioned it!

@egreg Have fun ;)

David Carlisle

@egreg just reply to the email, and change the CC to say latex-bugs@.. and latex-team@.. not the ones that it gets automatically

Ulrike Fischer

10:36

@DavidCarlisle a glimpse in the luatex future (miktex next has now luatex 1.06): warning (font expansion): autoexpand not supported (I don't think that it is something to worry about but I asked Robert anyway).

Joseph Wright

@UlrikeFischer Oh goody

David Carlisle

@UlrikeFischer I tried to build 1.06 the other day and the makefile failed, so haven't had time to debug why.

Ulrike Fischer

@JosephWright I really don't think that it is serious. There was a discussion in july between Robert and Hans, and if I got it right Robert wrote that "autoexpand" isn't making a difference in luatex anyway and so they probably removed the keyword. But at first I got a shock and thought that all expansion is gone ;-)

David Carlisle

13:23

@egreg welcome to gnats:-)

daleif

13:35

I consider writing a small column for a student magazine/paper about wrong uses of the command \text. A few relevant bad ones: \text{log}(A) \in \text{Mat}_n(R) and R_\text{max} (including the missing outer {}'s. Any other ideas for relevant examples of wrong uses for \text?

egreg

@daleif $A=\{x\in X:P(x)\text{ for some }P\text{ and }Q(x)\text{ whenever it's raining hard}\}$ , meaning long set descriptions with several words.

Joseph Wright

13:52

@DavidCarlisle :)

@DavidCarlisle Just as we are looking to leave ...

@DavidCarlisle, @egreg Enjoying Chris's rumblings?

daleif

14:33

@egreg ahh, yes a classic, put it all inside the \text, forgot that one, seen it a lot.

Harald Hanche-Olsen

@daleif There is also the style issue: Sometimes, it is better to state in words that … $A$ is the set of $x\in X$ so that …

daleif

@HaraldHanche-Olsen exactly, but I think that is a little to broad for the context of this small article.

Ulrike Fischer

14:48

@daleif as \text changes with the surrounding font imho there are not many places where it is semantically correct. You only need to show {\scshape $ a \text{ and } b $ }.

Harald Hanche-Olsen

@daleif Got to admit, I have sometimes written stuff like \begin{cases}42x&\text{for } x>0\\…. It gives me a guilty pleasure of having done something I really shouldn't do, which still seems harmless enough. But apart from ignoring any \mathsurround in effect, does it actually have harmful side effects?

daleif

@HaraldHanche-Olsen I usually rewrite them to \text{for $x>0$} just because I like it better.

@UlrikeFischer exactly. I just adjusted a preprint/working paper for a Centre, where the authors used \text{Var} in the text and \text{\rm Var} if used inside a theorem. One of them actually got a bit mad when I pointed out how bad that was.

Harald Hanche-Olsen

@daleif Unrelated, but if we're swapping war stories: I saw a paper in which the authors wanted to refer to a range of equations, like (3.4)–(3.7). In the rest of the paper, they used \eqref properly, but here, they used instead (sensitive folks should avert their eyes here) $(\ref{foo})-(\ref{bar})$ !

daleif