Hi

Tomorrow we look at two new projects in technical typsetteing. They both have production of HTML and accessibile outputs as a central goal.

The TeX Hour is Thursday 23 March 6:30 to 7:30pm GMT. The Zoom URL is
https://us02web.zoom.us/j/78551255396?pwd=cHdJN0pTTXRlRCtSd1lCTHpuWmNIUT09.
UK not yet daylight saving time. Here's the UK time now: https://time.is/UK

The first is Typst, which uses a modern input syntax and a Rust core to create a tool that will be easier to use. I learnt about it from a post to the BlindMath mailing list.
https://github.com/typst/typst
https://libreddit.spike.codes/r/rust/comments/11xpg6e/typst_a_modern_latex_alternative_written_in_rust/
http://nfbnet.org/pipermail/blindmath_nfbnet.org/2023-March/010613.html

The second, if time allows, is my own Unlatex project., where I'll report recent progress and discuss emerging goals and opportunities.
https://github.com/arxtex/unlatex
https://texhour.github.io/2023/03/09/unlatex-texbox-access-tree/

It's a big world and there's plenty of room for multiple projects. The more they share, the better things will be.

wishing you happy LaTeXing, Unlatexing and Typsting

Jonathan

Had a look.

How does that improve over Markdown/Pandoc?

el

On 22/03/2023 20:44, Jonathan Fine wrote:
[...]
> The first is Typst, which uses a modern input syntax and a Rust
> core to create a tool that will be easier to use. I learnt
> about it from a post to the BlindMath mailing list.
[...]

Had a deeper look.

While I would have preferred this used straight Markdown instead
of yet another Markup Language, this most certainly looks VERY
interesting, indeed.

el

On 23/03/2023 16:02, Jonathan Fine wrote:
> Hi Eberhard
>
> Typst is, or is intended to be, an app. Your query, I assume, is based
> on markdown to LaTeX to PDF as the route to PDF. If so, then a big
> difference is that it is not necessary to install TeX Live.
>
> I hope this helps.
>
> Jonathan

On 23/03/2023 14:24, Dr Eberhard W Lisse wrote:
> Had a look.
>
> How does that improve over Markdown/Pandoc?
>
> el
>
> On 22/03/2023 20:44, Jonathan Fine wrote:
> [...]
>> The first is Typst, which uses a modern input syntax and a Rust
>> core to create a tool that will be easier to use. I learnt
>> about it from a post to the BlindMath mailing list.
> [...]
>
>

Dr Eberhard W Lisse <nospam@lisse.NA> writes:
>While I would have preferred this used straight Markdown instead
>of yet another Markup Language, this most certainly looks VERY
>interesting, indeed.

I'm in the process of writing a generator for web pages (and maybe
other output languages) that should use an input language that is
"Markdown-like" in the broadest sense. By that I really just mean
that the markup should appear as natural and unobtrusive as possible.

So of course I am interested: What are - in short and abstract
terms - features that you find "VERY interesting" in this system
(Typst?) and/or which features in such a system are particularly
important to you or features that you miss in previous systems?

On 23/03/2023 15:40, Dr Eberhard W Lisse wrote:
> Had a deeper look.
>
> While I would have preferred this used straight Markdown instead of
> yet another Markup Language, this most certainly looks VERY
> interesting, indeed.

I believe this is the original research which generated Typst

https://www.user.tu-berlin.de/laurmaedje/programmable-markup-language-for-typesetting.pdf

Peter

Thank you,

this thesis is quite helpful.

I generate tables from CSV a lot and while generally using
LUALaTeX usually used PDFLaTeX for this as it is faster. So I
created two approximate Typst and LaTex files and ran them through
the hyperfine benchmarking tool using 2 warmup laps, after which I
ran the PDF Squeezer (pdfs) on the outputs

typst t.typ 109.9 ms
pdflatex l.tex 452.8 ms
lualatex l.tex 925.3 ms

File sizes range from 60 to 20 KB without compression.

Verbosity of the input files is similar, but Typst seems a little
easier to write and read.

Source files enclosed (obfuscated, whois records of IP addresses
originating queries). PDFs were not permitted for upload.

If someone is suitably bored one can maybe compare a really large
document :-)-O, it supports #include "file.typ" :-)-O

Interesting...

el

On 23/03/2023 18:48, Peter Flynn wrote:
> On 23/03/2023 15:40, Dr Eberhard W Lisse wrote:
>> Had a deeper look.
>>
>> While I would have preferred this used straight Markdown
>> instead of yet another Markup Language, this most certainly
>> looks VERY interesting, indeed.
>
> I believe this is the original research which generated Typst
>
> https://www.user.tu-berlin.de/laurmaedje/programmable-markup-language-for-typesetting.pdf
>
> Peter

On 24/03/2023 00:14, Dr Eberhard W Lisse wrote:
[...]
> typst t.typ 109.9 ms
> pdflatex l.tex 452.8 ms
> lualatex l.tex 925.3 ms

Very interesting, thank you. TeX systems do of course impose some excise
in the form of required compatibility with the trip and trap tests, so
they would be expected to be slower than a much smaller program written
from scratch.

> Verbosity of the input files is similar,

Also interesting: most proponents of LaTeX-alternatives are very proud
of NOT being verbose (esp Markdown). The penalty, of course, is that
they are usually not as flexible.

> but Typst seems a little easier to write and read.

I think it appeals to programmers because it looks more like the syntax
of a programming language.

Peter