| DocScrutinizer05 | chem|st: sorry, no idea. Isn't sicelo one of council? or is board even sth different than council? no idea, I only know the wiki page as linked in topic hasn't seen updates since afes | 02:15 |
|---|---|---|
| DocScrutinizer05 | is wiki.maemo.org down? | 02:24 |
| DocScrutinizer05 | aah nope, just slow | 02:24 |
| DocScrutinizer05 | council, Current: January 2018 - (current): Eetu Kahelin (eekkelund) Jussi Ohenoja (juiceme) Timo Könnecke (mosen) | 02:26 |
| DocScrutinizer05 | juiceme_: do you know about status of coucil? | 02:27 |
| ceene | i'm just going to use you guys to rant a bit | 10:58 |
| ceene | I hate latex with the same passion I love it | 10:58 |
| ceene | I've spent 2 hours to find the issue that was breaking my document, and it was an unescaped _ in text | 10:58 |
| bencoh | :) | 11:12 |
| ceene | and it's not even fixed yet | 11:19 |
| ceene | i'm gonna end up rewriting this shit on word | 11:19 |
| bencoh | aww, noo :/ | 11:21 |
| ceene | i have this thing to document a few things at work | 11:21 |
| ceene | and some of the sections are autogenerated from code | 11:21 |
| ceene | it's pretty and automatic | 11:21 |
| ceene | but since a couple of days it doesn't work | 11:22 |
| ceene | and I don't know fking why | 11:22 |
| ceene | there's no reason, no reasonable error, no explanation | 11:22 |
| Maxdamantus | git bisect? | 11:31 |
| ceene | I already know which is the thing that makes it fail | 11:32 |
| ceene | but I didn't know the exact underlying reason | 11:33 |
| ceene | i think it's the underscores | 11:33 |
| * Maxdamantus has wondered when TeX will be replaced. | 11:34 | |
| ceene | i think it's pretty when it works | 11:35 |
| ceene | but getting it to really work is a pain in the ass | 11:35 |
| ceene | i don't think it's worth the effort | 11:35 |
| Maxdamantus | I think the basic principle behind TeX is desirable: having a programming system for concisely creating printable documents, where various stylistic things are just handled by normal abstractions you have in programming (eg, functions) | 11:37 |
| Maxdamantus | The way it works just happens to be essentially magical. | 11:37 |
| Maxdamantus | because it was designed a long time ago, when a lot of programming concepts were either still being developed, or were just difficult to implement. | 11:39 |
| ceene | the language is ugly, you always need to load and undefined number of packages, but sometimes they conflict one with the other or must be included in a very specific order, but that order may conflict with some other package you need | 11:39 |
| Maxdamantus | so instead of figuring out how to express everything cleanly using something like a lazy programming language that just fundamentally does function application (which is basically all you need), it's based on macro expansion, which is very hard to reason about. | 11:40 |
| Maxdamantus | really, all you need is some usable syntax for expressing function application, function expressions, and variable binding, and you basically have the same level of power as the TeX language. | 11:41 |
| Maxdamantus | just need to then provide some typesetting primitives (which TeX does in its own system), and some higher-level typesetting library (like LaTeX), and some module system for referring to third-party libraries. | 11:42 |
| ceene | most of the time you don't really need something like that | 11:43 |
| ceene | you only need a sensible way to write a table | 11:43 |
| Maxdamantus | Yes, but unless you have "make a table" as a primitive operation, you need some sort of abstraction to do that. | 11:43 |
| Maxdamantus | and having "make a table" as a primitive operation would be a very inflexible system. | 11:44 |
| ceene | yep, but latex/tex is supposedly about how you can separate content from presentation, but that is basically false | 11:44 |
| Maxdamantus | since practically everything requires special support from the base system. | 11:44 |
| ceene | given that you need to splatter all your text with arcane commands | 11:44 |
| Maxdamantus | Having it without the ability to write abstractions is basically like writing TeX, but without being able to write things like \def | 11:46 |
| Maxdamantus | Practically noone writes plain TeX, and the ones that do will probably be using \def | 11:47 |
| KotCzarny | so, basically html+css but with more up-in-the-ass granularity? | 11:47 |
| Maxdamantus | LaTeX is just that common library on top of TeX, which depends on being able to write abstractions. | 11:48 |
| KotCzarny | well, xml+css | 11:48 |
| ceene | well, yes, that seems more sensible than latex to be honest | 11:48 |
| ceene | most of the time you want to write paragraph after paragraph of text, why the hell do I have to add \paragraph{} ? | 11:48 |
| ceene | markdown is even better | 11:48 |
| ceene | because the inevitable commands are basically non code, so it doesn't interfere with the textual visualization | 11:49 |
| KotCzarny | last time i had to create a folder out of products, i just created html that i later converted to pdf and set it to printing company | 11:49 |
| KotCzarny | so basically: sql+files -> php -> html -> pdf -> printing | 11:49 |
| Maxdamantus | So PHP is where you get your abstraction from. | 11:50 |
| Maxdamantus | I don't think HTML/CSS alone is sufficient for replacing LaTeX as it's used, eg, in academia. | 11:50 |
| KotCzarny | still, it was enough, and flexible at that | 11:51 |
| KotCzarny | any changes were easy to add and were seen instantly | 11:51 |
| Maxdamantus | Right, it's very flexible, because you've got an actual programming language in the pipeline: PHP | 11:51 |
| KotCzarny | some say php isnt programming language | 11:52 |
| KotCzarny | ;) | 11:52 |
| KotCzarny | but preprocessor | 11:52 |
| Maxdamantus | So you're fully capable of separating the "content" from the "display" | 11:52 |
| Maxdamantus | ie, you're not having to write something like "indent 5px" at the beginning of every paragraph. | 11:52 |
| KotCzarny | that's css's work | 11:53 |
| Maxdamantus | or denote classes that effectively say the same thing. | 11:53 |
| Maxdamantus | anyway, here's something I was working on at one point (I'm not the primary author), very dubious of trying to write this in pure HTML/CSS: https://maxdamantus.eu.org/42.pdf | 11:55 |
| KotCzarny | imo it's very doable with html+css | 11:56 |
| Maxdamantus | It's obviously *do*able | 11:57 |
| Maxdamantus | but is it maintainable? | 11:57 |
| KotCzarny | yes | 11:57 |
| KotCzarny | if you throw php at it too | 11:57 |
| KotCzarny | to generate things in proper places | 11:57 |
| Maxdamantus | You can obviously write Linux in x86 assembly | 11:57 |
| Maxdamantus | but is it maintainable? | 11:57 |
| KotCzarny | but yes, i think you can write it quite cleanly in pure html+css | 11:57 |
| Maxdamantus | Sure, if you use PHP instead of LaTeX, you can do whatever you want with it. | 11:57 |
| Maxdamantus | You can write a layout system in PHP that just figures out where every character goes. | 11:58 |
| Maxdamantus | you could even just write a font renderer in PHP, so you don't have to rely on font handling in the HTML being accurate. | 11:58 |
| KotCzarny | any preprocessing makes it much easier. that i agree | 11:58 |
| ceene | the thing is that latex doesn't really does a good job in the end | 11:59 |
| Maxdamantus | but for all of that, you need libraries, so you need to either write it yourself, or rely on someone else having written it. That's the part that you actually get with LaTeX: someone else has already written a bunch of reusable layout combinators. | 11:59 |
| KotCzarny | btw. what's that 'FOO' on next-to-last page? | 11:59 |
| KotCzarny | something left behind? | 11:59 |
| ceene | https://i.imgur.com/JnpCI6d.png | 11:59 |
| Maxdamantus | Can't remember. That whole document is a WIP. | 11:59 |
| ceene | solve this | 11:59 |
| KotCzarny | or spacer put there so big cells go on next page | 11:59 |
| Maxdamantus | can probably look at it in git | 11:59 |
| KotCzarny | nah, got own work to do, eh | 12:01 |
| ceene | why is that <we over the border of the box? | 12:01 |
| Maxdamantus | ceene: because we weren't caring about exact positioning issues at the time. | 12:02 |
| ceene | why do you have to manually tweak that? is it because < is somehow a special character? | 12:02 |
| ceene | but that's the thing, you should have to care about that | 12:02 |
| ceene | s/should have/shouldn't have/ | 12:02 |
| infobot | ceene meant: but that's the thing, you shouldn't have to care about that | 12:02 |
| Maxdamantus | ceene: the document is mostly defined semantically, positioning issues can be fixed by mostly just fixing the combinators. | 12:02 |
| Maxdamantus | ceene: it's really not important for a WIP. | 12:02 |
| Maxdamantus | ceene: should the document really be purple? | 12:02 |
| KotCzarny | :) | 12:02 |
| KotCzarny | 80/20 rule | 12:03 |
| KotCzarny | 80% of work takes 20% of time. remaining 20% takes the remaining 80% of time | 12:03 |
| KotCzarny | 'hey, can you put that dot 3px higher and make it glow?' | 12:03 |
| ceene | you know what I mean, the presentation is wrong and it's not your fault that latex somehow decided to move that <we a few pixels to the left | 12:04 |
| ceene | and thw whole document is astonishing, but why does it fail in something so trivial? | 12:04 |
| Maxdamantus | ceene: because with power comes responsibility. | 12:06 |
| ceene | you're justifying a bug | 12:07 |
| Maxdamantus | ceene: if the system gives you the ability to have fine-grained control over layout (instead of saying, eg, "I'll put whatever you output inside a box which is strictly bigger"), then you have to be careful about things like that. | 12:07 |
| ceene | that whole string fits in that box | 12:07 |
| KotCzarny | i remember when i was looking for work, trying some big newspaper company | 12:07 |
| ceene | the text above and below is alineated | 12:08 |
| KotCzarny | they did their webpages PIXEL EXACT | 12:08 |
| KotCzarny | with few supported browsers and html+css | 12:08 |
| KotCzarny | and css was primitive back then | 12:08 |
| ceene | but that line, that doesn't begin with a math character somehow is broken, and there's no reason for it to happen | 12:08 |
| KotCzarny | and ie6 still ruling the world | 12:08 |
| KotCzarny | luckily i didnt get that job | 12:08 |
| ceene | alineated doesn't exist in english? | 12:09 |
| ceene | lol | 12:09 |
| ceene | aligned | 12:09 |
| ceene | i'm an idiot | 12:09 |
| KotCzarny | :) | 12:09 |
| ceene | that whole string fits in that box, the text above and below is aligned, but that line that doesn't begin with a math character somehow is broken and there's no reason for it to happen | 12:10 |
| ceene | much better | 12:10 |
| Maxdamantus | The actual reason is because there's a random `\hspace{-2ex}` in front of it. | 12:12 |
| KotCzarny | :) | 12:12 |
| ceene | lol | 12:12 |
| ceene | maybe this time it's just that | 12:12 |
| Maxdamantus | for the most part, the TeX code is free of positioning information, that's one of the few exceptions. | 12:12 |
| KotCzarny | which means it was a workaround for something else that didnt quite get bug matching rule set | 12:12 |
| ceene | but i'm really fed up with that kind of thing that happens just because | 12:12 |
| Maxdamantus | \produzione{\XvP\!\!}{ \Xt{}{\x}{\T}{ \valP } }{}\\ | 12:15 |
| Maxdamantus | \seguitoProduzione{ \hspace{-2ex} \mbox{<we write just $\val$ and $\vs$ when $\p$ is clear from the context>} \hspace{-10ex} }{}\\ | 12:15 |
| Maxdamantus | That first line corresponds to the production rule just above the positioning error pointed to. | 12:16 |
| Maxdamantus | (the actual source code has additional newlines/indentation, so it's a bit easier to read than what I posted here) | 12:16 |
| Maxdamantus | (though I often did get annoyed at the main author's occasional indentation inconsistency (occurs in Java code, not just TeX)) | 12:17 |
| Maxdamantus | he's also Italian, so the TeX is partly in Italian. | 12:17 |
| Maxdamantus | I think the occasional negative spacing is just to force stuff to fit onto one line, but since the document was being semantically updated (not just updated to improve styling), those negative space hacks might've become inconsintent. They should obviously be fixed at some point. | 12:20 |
| ceene | i'm reading this http://www.danielallington.net/2016/09/the-latex-fetish/ | 12:28 |
| Maxdamantus | and as for the "FOO", apparently one of the source files was just replaced entirely with the content "FOO" | 12:29 |
| ceene | well, that's obvious, I thought it was yet to be written | 12:30 |
| Maxdamantus | presumably because the content was moved to a different place, and I guess he didn't feel like removing/unreferencing that file at the time. | 12:30 |
| ceene | neither that nor the guy writing italian is latex' fault :) | 12:30 |
| KotCzarny | or that didnt feel like fixing broken page layout | 12:31 |
| KotCzarny | ;) | 12:31 |
Generated by irclog2html.py 2.17.0 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!
