@barbarabeeton How about using the term "Digital Typography systems" or "Desktop Publishing systems"?

@DavidCarlisle I'm not sure I follow. Are you saying that if the inner spacing is not removed, you get double the column separation?

@FaheemMitha yes the outer table adds column separation so you want the inner one to go right to the edge of the cell, just try it, you'll see

@DavidCarlisle Understood

@MarcelKrüger what did you use to uncompress the pdf's?

Hello!

I want to ask, what should I do so that when I use latex -> dvipng, the output is just the content in its bounding box?

@UlrikeFischer qpdf

@MarcelKrüger many options ;-). Which one did you use?

@soupless Did you already try dvipng -T tight?

or -T bbox

I haven't tried it prior to this message, but I think that should solve the problem. Thanks

@soupless You're welcome!

Wait, sorry. Last question for now, from dvipng --help, -O c refers to image offset. To be sure, is this the offset from the bounding box or from the top-left of the document?

> Move the origin by x-offset,y-offset, a comma-separated pair of dimensions such as ‘.1in,-.3cm’. The origin of the page is shifted from the default position (of one inch down, one inch to the right from the upper left corner of the paper) by this amount

(from texdoc dvipng)

@raf -- I question whether systems like LibreOffice can really be considered "typography". And LaTeX is also used as the basis of production publishing (which at one time required a mainframe), although it is obviously available in the desktop environment. But I'm probably an outlier in this area, too finicky.

@barbarabeeton ooh twitter.com/shouldhaveaduck/status/1433810990450151429

@barbarabeeton In case you are interested, I have posted this issue as a question on EL&U: english.stackexchange.com/q/574246/388396

@UlrikeFischer Nothing fancy, only the decompression related ones. The full zsh command line was for n in */*.pdf; do mv "$n" dummy.pdf && qpdf --stream-data=uncompress --object-streams=disable dummy.pdf "$n" && rm dummy.pdf; done

@MarcelKrüger that are the options I thought would be the best too. I just made me a small batch qpdf --stream-data=uncompress --object-streams=disable %1 uncompressed-%1. That works very nicely, and it doesn't crash with pdf 2.0 ;-)

@raf -- Yes, interested. Thanks. Looks like it's back to "document processing" or "document preparation". Ignores any implication of quality and concentrates on the target object. While I may have a distinct preference regarding quality (and I do), I think most people don't, but the knowledge of how to prepare documents is a highly useful skill.

@DonHosek Can't wait for your books to drop! I had a bad experience on Kickstarter and left, but I'm otherwise keeping up at your site. Just gave you a shoutout on Twitter.

@UlrikeFischer For such things qpdf is pretty great. Sadly it's --check option is much less useful since it only detects very obvious syntax errors.

@MarcelKrüger I miss pdftk and migrated to qpdf in... 2017, I guess. :)

@DavidCarlisle kotaku.com/…

@PauloCereda I still have pdftk... it one of that snap wizardry, but it works

@Rmano ooh :) Fedora deprecated it a long time ago, so I moved on. :)

@PauloCereda snapcraft.io/install/pdftk/….

@Rmano oh no, snap :)

@Rmano I am not a fan of snaps, I'd rather use flatpak. :) Let me check if it's available...

Yes... I know. But ....

Appimages for this kind of small utility is the best, I think

@Rmano ooh flathub.org/apps/details/net.sourceforge.pdfchain

Latest version of this software is very old and so it may contain bugs, so to be on the safe side make sure to NOT open any untrusted pdf files with it. For this same reason we have limited this application to only save files to standard locations: "Downloads", "Documents" and "Pictures". You can also save to external devices (like USB pendrive).

@Rmano ^^ love the disclaimer :)

It seems there is a java version around

@Rmano A Docker image running an old Ubuntu with pdftk. Wow! :)

I have a 16.04 virtual machine around, maybe I could try to make an appimage... But I'm unsure about licensing

@PauloCereda using a sledgehammer to crack nuts ;-)

@Rmano :)

@Rmano I (forcefully) migrated to qpdf. It's a fantastic piece of software, although the command line syntax seems a bit unusual... :)

@PauloCereda I'll have a look!

@Rmano yay!

New research indicates that over 80% of American adults now have antibodies to the coronavirus, earned either through infection or vaccination. The study, based on data collected from blood donors, estimates that over 80% of Americans over the age of 16 had these antibodies as of May 2021. The authors do caution, however, that their results may not be generalizable to the entire U.S. population.

@PauloCereda trs 80 flatmate had one of those when i was an undergrad

@DavidCarlisle ooh you could tell him the program now works :)

Does anyone know of examples of Lua parsing LaTeX tables, i.e. tabular, longtable and so forth? The only example I've found is tex.stackexchange.com/a/592856/3406

@FaheemMitha wouldn't it be simpler to use lua input? Why parse tex with lua when you have a tex parser to hand?

@DavidCarlisle I'm not sure what you mean by "use lua input".

@FaheemMitha whatever Lua structure you had in mind constructing by parsing a latex table you could construct directly in Lua

@DavidCarlisle That's true. I could. But the use case I'm imagining is processing a TeX file with a LaTeX table in it.

@FaheemMitha it's much easier to parse such a file with tex than with Lua.

@DavidCarlisle OK. Is it possible to hand off the results of the TeX parsing to Lua at a later stage? Or not?

@FaheemMitha depends what you want to do, you can traverse the node tree generated by typesetting the table.

@DavidCarlisle I see. I'm imagining spreadsheet like functionality like a formula adding A2+b2 together and putting the results in C3.

@FaheemMitha that's what I guessed but for such a thing why not just start with a Lua table, do the calculations then write out the tex table. Likely to be far simper

@DavidCarlisle I haven't really thought of such an option. But I'd like all the data to be in the TeX file. Though I suppose such an approach would be simpler.

Also, representing a spreadsheet/table in Lua doesn't sound appealing. I have not tried it, though.

@FaheemMitha it's just a 2d array so a nested Lua table.

\documentclass{article}

\begin{document}


\directlua{
a= % initial data
{{1,  2, 3},
 {10,11,12}
}
a[3]={} % new row
for i =1,3 do
a[3][i]=a[1][i]+a[2][i] % calculation
end
labs={'x','y','x+y'}
}

\begin{tabular}{lrrr}
\directlua{
for i,v in ipairs(a) do
print('\string\n== ' .. i)
for ii,vv in ipairs(v) do
tex.print('&' .. vv)
end
tex.print('\string\\\string\\')
end
}
\end{tabular}
\end{document}

@DavidCarlisle Yes, I see. But it doesn't exactly reproduce the spreadsheet experience.

Not that I've got anything against programming in Lua.

@FaheemMitha and doing it in tex does? you could of course add helper functions to sum rows or other spreadsheet functions. Probably such a Lua library exists already.

@DavidCarlisle A LaTeX table feels more like a spreadsheet, yes. At least, IMO. There is something like that for ConTeXt, I think. But not for LaTeX, as far as I know.

@FaheemMitha I'd say the opposite a Lua table is far more like a spreadsheet is a 2d array you can index any entry by two integers, the data is numeric. A latex table doesn't (normally) allow access to individual cells, and the data is not stored as numeric data spreadtab tries to hide that but ....

@DavidCarlisle I meant in the sense that Lua doesn't support representation of this 2D array in a 2D data frame sort of way. Of course, you could write a function to display it that way, I suppose. But if you want the data integrated with the text, having it directly in your TeX file is an advantage. You don't have to go and look for it somewhere else.

I.e. like R and Pandas dataframes.

Of course, treating a LaTeX table as a spreadsheet is distinctly awkward, so there's that tradeoff.

@DavidCarlisle Actually, it's been a while since I used a spreadsheet, so I can't quite remember how they work, but don't they allow you to mix both text and numerical values? Though presumably not in the same cell.