/t/ - Technology

Discussion of Technology

Index Catalog Archive Bottom Refresh
Options
Subject
Message

Max message length: 12000

files

Max file size: 32.00 MB

Total max file size: 50.00 MB

Max files: 5

Supported file types: GIF, JPG, PNG, WebM, OGG, and more

E-mail
Password

(used to delete files and posts)

Misc

Remember to follow the Rules

The backup domains are located at 8chan.se and 8chan.cc. TOR access can be found here, or you can access the TOR portal from the clearnet at Redchannit 3.0.

US Election Thread

8chan.moe is a hobby project with no affiliation whatsoever to the administration of any other "8chan" site, past or present.

You may also be interested in: AI

(2.95 KB 230x87 alignment.PNG)

LaTeX Anonymous 06/02/2020 (Tue) 08:35:25 No. 334
Although not entirely in the spirit of /t/, let's talk a bit about this typesetting tool. What do you use it for? Do you like using it? Encountering any problems? What program do you use? I mainly use it for my academic reports, since it works very well with formulae and chemical reactions. Anything else would be an abstract horror, so to say. I have been using TeXworks+JabRef, and have mostly been satisfied with the results - except for the occasional hiccup like the formula above. The only drawback is that there is not a proper dark theme, so it gets a little straining on the eyes sometimes.
>>354 >enjoy >implying You do what you gotta do, the world ain't perfect so we're stuck with PDF as it's the only document format that mostly works everywhere.
>>334 >What do you use it for? For documents with lots of equations and documents that will eventually be printed. Using it for everything is not good idea. In engineering everyone writes specifications and technical documentation in word and nobody cares about how anything looks. It's too easy to get carried away with typography and cool features, but at the end of the day nobody cares if there is too much whitespace in paragraphs and there are no ff and fi ligatures. >Do you like using it? Sometimes yes, sometimes no. Tables are still pain to make, documentation can be intimidating (tikz, pgfplots, ...). >What program do you use? Emacs with auctex + reftex + yas snippets + flyspell mode. This works well for me, since it automatically generates and inserts references and bibliography. It's so easy to add bibliography with emacs - drag and drop paper, emacs reads its DOI or ISBN and inserts bib entry automatically. Bibliography is pain in word in comparison, so I wouldn't use it for academic work. I've come around full circle in my document preparation and started appreciating wabi-sabi nature of word/libreoffice documents. Even though it's crappy from typography point of view, there is something about imperfections nature of those documents that makes them interesting to look at. For example take a look at pic related: pixelated logo, indentation, no justification, no hyphenation, table title too big, spacing in tables... But I kinda like it despite it's flaws. And making that table would be suicide in latex. I even started using gnumeric for calculating various things instead of scripting - it's just so simple and easy. Even if I rice my emacs to do something similar it would be convoluted and clunky. Sometimes it's best to get rid of your OCD and just do the work, sometimes I tend to worry about my tools more than my work.
There's something that still bothers me quite a bit when doing anything remotely serious, tables are tedious to make, like what >>374 says. Isn't there any way this could be automated? Have a package parse a .csv or some shit. >>345 Alright, I'll look into automatic bibliography management, I have a project coming up which is just about that, referencing a bunch of papers. >>346 At my degree, it was a sort of unofficial minicourse, that didn't even give you credits.
(16.72 KB 474x227 latex_table.jpg)

>>388 The main issue with tables is that the notation is bulky and easy to make a mistake with. And you have to type a bunch of stuff in by hand. Tables are just better-suited to visual editing. There's packages to automate pulling in data from a source to some degree, but I usually just use a table formatter. There's at least one available for free online. For a small table, this is an acceptable solution and is faster than formatting a table manually. >unofficial minicourse, that didn't even give you credits I just took some courses that required typed homework with mathematical symbols and they mentioned LaTeX with a wink and a nudge. They didn't want to offer a course, so they bullied us into doing it. I had been meaning to learn it for some time so I jumped right in. A lot of my classmates, including many of my partners, refused. They kept playing with Word. One time, I had to chew a guy out for spending hours re-creating my work in Word and not making any progress on the assignment. It was incredibly frustrating. A lot of students these days just use Overleaf or something as a compromise. Makes tables easy, avoids having to manage a package, and they can typeset math. But it feels like as much of a sellout as Jupyter Notebooks.
>>389 I know how notation is, I've had to do quite a few tables. It's dreadful. I think I'll look into such packages. By the way, Jupyter Notebooks are a sellout? In what sense? I mostly just looked at those in order to have some sort of half decent pseudo-Mathematica handling of symbolic math and such. It doesn't even seem to come close to Mathematica in most respects, though.
(1.74 MB 320x240 dalailaser.gif)

>>388 >>374 >>389 Hey fellas, do you know how I do that? I've got a quite simple and quick solution actually, it's called being a cheating fucking nigger lmao https://www.tablesgenerator.com
(485.85 KB 480x360 groovy.webm)

>>402 > hyperlinks are not censored properly
>>391 My problem wit Jupyter is that it's a backwards solution to a non-problem. I understand why having executable sections is convenient, especially for mathematicians and physicists. But the entire thing runs in a broken Web GUI that strongly encourages you to use their own servers for it. They (and Google) want to charge a premium for additional RAM or CPU time, and they leak memory like a sieve. You can run and connect to it locally or on your own server, but 99% of the people who like Jupyter Notebook will not be able to do this. I've worked with physicists and they were genuinely shocked I ran Python from the terminal and debugged by either printing stuff out or stepping through with a debugger. They cannot debug unless they look at it in Jupyter Notebook and execute all intermediary steps. And their code is absolutely jumbled. Copied from colleagues who copied it from colleagues, with slight tweaks, none of them understanding proper syntax, just how to mess with the statement for what they're trying to computer. Jupypter is a sellout because their goal is to upsell a Python environment to people who need it but don't want to learn it properly. Half their traffic probably comes from "how to Machine Learning" tutorials.
>>404 >people who are experts in physics aren't automatically experts at coding Stop the presses. More seriously, have you tried the vscode debugger? I know it's fantastic for C++.
(253.10 KB 1440x1080 screams geometrically.jpg)

>>404 > shitty code and scientists You know what I had to learn a year ago, and have been using quite regularly since? MatLab. That can be an abstract kind of hell some times.
>>402 I like your solution, Dalalalalalalai-sama, except for the fact that you're basically outsourcing your information to a website owned by some random faggot that could be selling it to whoever. So while this'll do when in a crutch, I'd rather have a solution that a) is FOSS and b) runs in my own 'puter. >>404 I see. Well, I had to forgo Jupyter because honestly it doesn't seem like it could ever come close to Mathematica, I guess I was a fool for believing it ever could, especially if that is not its primary goal, but the flashy GUI somehow made me think otherwise. Talking about scientists and Python, I'm using it as a MATLAB replacement, and since I'm not dealing with anything too complex (linear regressions, doing ODE with an arbitrary function like a gaussian, numerical calculations galore) it's not like I need MATLAB, even though we were encouraged to use it since we were taught it at a course that was about numerical and statistical methods. Plotting is nice but it has some weird defaults, some of which I've still not modified and I want to. I'm running it like you do, on the terminal, and if shit goes awry, I just drop some print statements here and there to see if I'm doing anything retarded. I don't think I need a debugger.
Tables are fine and using an online thing is shitty but an okay bandaid fix in a pinch. What makes me rage with TeX is trying to do page layout. Images have a tendency to not align as specified, either doing nothing or floating to the top of the page. You need to be a wizard with nested subframes to do anything decent. No libraries handle it well, and the few that are okay at it tend to break when combined with other libraries. It shouldn't be that hard to size an image, specify a starting position on the page, give it a margin, and ask text to flow around it. And once tables are involved it's officially a living nightmare. Better hope you don't have to do it on a deadline
(2.76 MB 640x420 1431599250983.gif)

>you have to compile twice to get Table of Contents to work The absolute state of Latex
>>334 I used LaTeX a lot in uni, especially for larger papers with lots of citations, my master's thesis, etc. My resume is still written in LaTeX actually, though I should probably change it. I don't like LaTeX for smaller papers because of the syntax and compile times. I fell for the meme and use groff -mom for that stuff.
>>557 > My resume is still written in LaTeX actually I tried that, but in the end it looks exactly like every other LaTeX resume, except you can change the color a little. If you want your resume to catch an employer's attention it should look like you put actual effort into it, not as if you just dumped your text into the same template everyone else is using. I ended up just asking someone to do it for me in InDesign. I tried Skribus, but the lack of any live preview made it completely unusable. >>435 Unless someone comes up with a better idea of how you can know the number of a page before the pages have been typeset, that's what we are stuck with. You can use latexmk for compilation, it's a wrapper around the latex binary which will automatically recompile your code when necessary. Then you have to run the command only once.
>>563 I would agree, but so many employers now use different shitty application services that a resume is basically just formality. Having a well-written cover letter tends to be far more important in my experience. Technology is also my hobby, not my profession, so it being in LaTeX at all makes it different from other applicants.
>>334 I have never touched raw latex, BUT, I use Lyx all the time for my projects. It's incredible how good shit looks when you take your time to work on it.
>>566 I'm torn on the issue of cover letters. On one hand, I've never needed one and been hired at some pretty nice places. On the other hand, I think Bloomberg literally threw my resume in the trash for not having one, because their recruiter was very impressed. It's not uncommon for HR and Engineering to have totally different ideas of what makes a good resume. You have to pass through HR's shitty filter to get to someone who knows enough to be impressed by your work. That said, my resume was made in Open/Libre Office. I should rewrite it in Latex, but it will take some time to figure out how to replicate it. The spacing is also really tight. My advice for resumes is to make good use of horizontal space to fit extra info in. If you want to cap it at one page, this is a must. >>616 It really is incredible how effortless the typesetting is once you learn to describe what you want. It's a steep learning curve for newcomers, though. But the results really are on-par with professionally-printed books. Sometimes you can match the quality with a Word document, usually you make unmaintainable in the process.
>>334 Wouldn't want to use anything else for scientific writing. When I got started with it I usually just used emacs but at some point I had to write longer and longer documents split into many files so I switched to using texstudio and jabref which both work great on every platform. The languagetool that can be added to texstudio must be one biggest ressource hog I've ever encountered.
TeX is completely superfluous today. It was meant to generate justified text in a printing format. For everything beyond you need packages that need to be ported to your tex system for your OS and CPU architecture. Now it can't really do anything but creating PDFs well. However today we have better alternatives. There is html, epub, odt formats and WYSIWYG editors. Everything supports instantly rendered justified text. And you can print them too. Or write your XHTML by hand. >buuuuuut maaaath and physics Use any MathML editor. If you want to put it in the aforementioned formats, just compile them into an SVG. You can't extract the math formula from it but neither can you from a PDF. Your teacher lied to you. TeX completely pointless and like PDF or Postscript not compatible with anything because the data is actually a script. I learned LaTeX, so don't try the "you can't judge what you don't know" excuse on me.
And you can still put the MathML into the SVG by using a <metadata> tag.
>>793 >Now it can't really do anything but creating PDFs well. You say that like this is something nobody ever needs. LaTeX is pretty much set and forget, a little inital setup and it just generates pretty much perfect documents. WYSIWYG editors always end up being a hassle. I've seen many documents over the years and the ones written in LaTeX always looked a little better. They also tend to support the features of PDF readers better. (X)HTML is fine for the web but it has a pretty bloated syntax and I don't see why anyone would use it for something that doesn't require it.
>>796 Where's your LaTeX e-reader, nerd?
>>796 >They also tend to support the features of PDF readers better. Grandpa. PDF doesn't even support Unicode properly. I would not want to store text in that format if I don't absolutely have to. Further if the reader only extracts the text, the entire positioning TeX did was completely useless.
Using katex. Had to \mathbb some shit, otherwise mostly has what I want out of the box
>>807 >>808 I'd rather use pdfs because when I cite things with regular page numbers even the actual grandpas won't complain about it. Otherwise epub would be more convenient for reading, yeah.
>>820 If you cite it by using copy and paste, I can just copy and paste your citation into the search function and thus find the point, you took it from.
>>814 Most distributions will just automatically install packages you need. They'll do it in the background when you compile. While I think Latex sometimes has an obtuse way of writing libraries, providing everything out-of-the-box in one go is monolithic and a bad idea.
>>334 >What do you use it for? Anything that involves formulae. If you don't use formulae you shouldn't bother. >Do you like using it? No, but nothing else can even compare. >Encountering any problems? Embedding images sucks and drawing graphs with a bunch of external packages does too. >What program do you use? Texmaker. There are undoubtedly better programs but it does everything I need. >>335 Tikz-feynman is a great library.
>>348 *IN WORLD WHERE* >Latex is shit and bloat .... the text file apocalypses is happening
>>334 JSON
>>348 Latex is the least bloat system in the world. You have to run it 3 times so that it generates all the artifacts for your documents because certain macros need pre-generated input files and the only way to create them is by having ran Latex already to created the inputs. Table of Contents is like this also other index like macros do the same thing. The problem stems from Tex not creating a DOM like structure and then creating the indexes in one execution by reading though the DOM and collecting all information required as an input for that macro. Latex macros, built on top of Tex, can't do this because it's a lacking ability of Tex. Arguably HTML is the most successful format over Latex, DVI files Tex creates, PDF, EPUB, propitiate Kindle books, and anything else you can think of, by adoption alone. What sucks about HTML is it's difficult to archive and use offline. By archive I mean keep copies for posterity and history in an easy format, which PDF does much more easily; print to PDF. Save as HTML and a zip of resources is too incontinent; it has to be a self contained format. Latex is pure power, like how WYSIWYG GUI makers like NetBeans are play-school compared to the power and ease you can control the GUI you're writing by doing in code and not in WYSIWYG. Latex is also abiding this truism, only the hobby community of Tex Maintainers have fail to modernise Tex to function and produce documents on the scale of the abilities of InDesine or Corel. Latex and Tex lack vision, but are very nice and easy to use once the user disabuses themselves of a love of cute WYSIWYG in favour of real productivity and programming might. By the way I'm the same guy that said JSON.
>>334 >what do you use it for Publishing papers. I don't really have a need for typesetting beyond that >Do you like using it Kind of. It does what it's supposed to do, and it's easy to pass the source around between authors. It's also easy to revise things even if you're not a latex expert. If you want to know what I hate then ask me about jewgle documents or something. >What program do you use? Yeah that was the entry level hump trying to get over so I could compile it and look at the results before passing the next version along. I'm as much a copy editor as an idea man for what we should be writing. And anybody who publishes something in disgusting microsoft fonts deserves to be mocked in the best way anybody can imagine.
>>10288 >>What program do you use? >Yeah that was the entry level hump trying to get over so I could compile it and look at the results before passing the next version along. I'm as much a copy editor as an idea man for what we should be writing. There is - TEXmacs https://www.texmacs.org/ - vim - emacs has tex-mode Really any text editor will do. For the proprietary vener-lock-in web-app feel maybe <https://www.overleaf.com/>.They do have good documentation for beginners.
>>10298 lol the emacs tex mode sucks as much unwashed cock as most of its other modes, and that's after you try to write hooks in your .emacs or .xemacs or ~/.emacs/.xmacstab or whatever to just force it in to text mode. Yes I do know ML and I don't care. They can't even keep anything consistent between emacs and xemacs, but you wind up falling between either than and vim, and I'll give you an earful about it not being part of a standard distribution DESPITE ALL THIS OTHER BULL SHIT I DIDN'T NEED OR WANT TO KNOW ABOUT but it's the compilation thing. I'ts kind of like making your own .deb or .rpm. I like how you can just trade it around as a pure text file, but you have to do some digging for the right incantation to turn it in to the final product, and that's important if your chart wound up on the wrong page, or if you have to edit down your list of citations to fit it in to the number of pages you paid for.


Forms
Delete
Report
Quick Reply