It's not "beautiful" (in fact, the author seems to have terrible layout/style sense) and that github page is just...terrible. I'm not sure why the author chose to embed 10 pages of PNGs on it.
But
I think there's definitely a need for well-curated collections of "common things people want or need to do documents"...like the "draft" watermarking, confidential flag, QR codes, org charts/family trees, and so on - especially if they are well-commented.
I would love to see such a collection for doing various common page layouts, and also for leveraging many of the features of PDFs.
And no, the various sites that host lots of example LaTeX documents don't feel like they are really a substitute. I've found them poorly curated and as a LaTeX novice, it's difficult to figure out how to extract the one thing the author of the document is doing that you want to pull and replicate. So often, I've tried to do so and ended up with incomprehensible 'compile' errors.
Even after repeated attempts to learn LaTeX and do useful things with it, I still find it almost impenetrable. Debugging is nearly impossible because the "compiler" gives completely shit error messages, and looking at LaTeX code as a beginner, I just see meaningless ASCII vomit. LaTeX seems to use a ton of "this one particular punctuation mark or letter tacked on to this command or function means LaTeX will actually interpret that as..." bullshit that makes the code impossible to read or understand.
I understand why that sort of shorthand was done given LaTeX was written in the Every Byte Is Sacred Era - but some evolution of LaTeX to be more parseable to the human eye/brain in the last three decades would have been nice.
The third edition is just out of the presses. I have the 2nd one, and one of the appendixes is an explanation of LaTeX and TeX infamously cryptic error messages, and it is really useful.
As for using PDF features, try PDFLaTeX or even better, XeLaTeX or luatex. The latter use utf-8 and advanced PDF kerning techniques, among other useful stuff.
Is there any software that can take handwritten equations (say on an ipad or ocr) and convert them to latex reliably because there’s no way I’m going to try and learn those incantations.
But I found that once I learned the LaTeX syntax, I can type equations faster than I can write them by hand, especially if they're complex. This is because most equations have lots of repeated elements (for example, when say, you're rearranging an equation), and I can copy and paste faster than I can rewrite them. I could take notes in LaTeX faster than a professor could write them on the chalk board.
I wrote my dissertation in LaTeX, and while I hate the debugging (errors are arcane), learning the "incantations" are really the least of one's worries. I wouldn't be put off by that.
Snip,TeXpad,Ink to LaTeX, Mathpix, MyScript MathPad,... I used Mathpix ones (3 years ago it wasn't really good ocr from handwriting, but I have heard that they have made a lot of improvements since this time). I tried some of them, and I hope it will help.
Why do you think there's no way? I never put any concerted effort into learning TeX maths syntax but by the end of writing my thesis it was as easy as writing it by hand. I would never dream of scribbling maths into a computer. Complete waste of time.
I work on maths using pencil and paper. I only use TeX for typesetting. Only a tiny fraction of my scribbles would ever get typeset, and it's usually going to be completely rewritten in the process anyway. I recommend just learning TeX for typesetting maths. It's simply the best tool.
This is giving me flashbacks to writing my thesis back in 2005 in lab in Germany. I remember spending almost a full working day on getting the logos of the two institutes I was working for aligned horizontally in LaTeX.
You can create amazingly beautiful publications in LaTeX and I'm not aware of anything equal for writing mathematical formulas, but I would probably not recommend anyone use LaTeX for a publication in 2023.
For one, graphics in typst is no match for LaTeX with tikz. Typst is trying to integrate Penrose as an alternative. We definitely need a modern general-purpose diagram and illustration DSL. It's to be seen whether Penrose can fill that role.
Another matter is that there are many LaTeX templates for many academic journals. LaTeX will reign supreme until typst templates are created and adopted as an alternative.
Finally, there's on matter I don't know about. I've heard that the TeX engine solves some tough optimization problems on word spacing, line breaking and layout. I wonder how typst compares, given it's young age.
But
I think there's definitely a need for well-curated collections of "common things people want or need to do documents"...like the "draft" watermarking, confidential flag, QR codes, org charts/family trees, and so on - especially if they are well-commented.
I would love to see such a collection for doing various common page layouts, and also for leveraging many of the features of PDFs.
And no, the various sites that host lots of example LaTeX documents don't feel like they are really a substitute. I've found them poorly curated and as a LaTeX novice, it's difficult to figure out how to extract the one thing the author of the document is doing that you want to pull and replicate. So often, I've tried to do so and ended up with incomprehensible 'compile' errors.
Even after repeated attempts to learn LaTeX and do useful things with it, I still find it almost impenetrable. Debugging is nearly impossible because the "compiler" gives completely shit error messages, and looking at LaTeX code as a beginner, I just see meaningless ASCII vomit. LaTeX seems to use a ton of "this one particular punctuation mark or letter tacked on to this command or function means LaTeX will actually interpret that as..." bullshit that makes the code impossible to read or understand.
I understand why that sort of shorthand was done given LaTeX was written in the Every Byte Is Sacred Era - but some evolution of LaTeX to be more parseable to the human eye/brain in the last three decades would have been nice.