• 3 Posts
  • 141 Comments
Joined 10 months ago
cake
Cake day: October 18th, 2025

help-circle




  • Not sure what is the snail mail, but yes; serious. It’s mostly simulations data. I can easily produce 100 TB/GPU*day. I have access to some 500 GPUs. Data can not stay where it is produced in order to free up space for more to be produced and must be shuffled around to be processed, thus the importance of a fast internet connection. Sometimes I need to get part of this data on my local computer, process it and upload it somewhere else, it can be terribly slow on a normal 1 GB/s connection.







  • Markdown is fine. I’ll take a look at protondocs, pricing is not really an issue - but since this is public money I very much prefer to spend it on actual experiments.

    Don’t know about collabora, I had tried it a few years ago and found it very slow and finicky. Moreover maintaining the whole nextcloud setup would be something I’d gladly avoid.


  • I tried the hedgedoc demo.

    As far as I can see it supports latex and has some change revision mechanism. Unfortunately no bib file support, but I guess that’s ok.

    However, I can not find a way to add comments to the rendered text. Is that possible on the self hosted version?

    That would be a deal breaker, most of what I need is actually reading documents written by other people and ask for clarifications or improvements.






  • Distro choice does matter, but to the common user/newcomer is basically irrelevant.

    Anything else that should be pointed out?

    A new user will be fine with basically anything that isn’t Gentoo or with some very very specific applications.

    People who have specific needs will choose a distro which makes the things they need easier. In most cases a newcomer won’t have such specific needs as to have to choose a specific distribution.

    designed for specific uses that conflict with other uses.

    Don’t see much conflict between uses. You can use any distribution to do anything you’d do with any other.

    Do you want a very stable system but rolling release with the latest updates? Not possible. Do you want a system that is flashed on ram but that will be stable with 100% uptime without ever touching it? Not possible. Do you want a super light system with the best DE animations and graphics? Not possible.

    Sure those are conflicting uses, but just because you’re choosing among two opposite things.


  • It all comes down to the repositories after all. Different distros have different update cycles and policies.

    Oh, also some distros apply a little bit different graphics and customisation on the default setup.

    After that, it is all the same. Distro choice does matter, but to the common user/newcomer is basically irrelevant.

    why distro hop?

    Fun waste of time, good way to learn how to setup a Linux system by doing that repeatedly.

    why don’t you use Ubuntu then?

    It’s a good system, go ahead with it. I don’t like very much their customisations, but it is cool system after all.

    why don’t you recommend Archlinux to a newcomer?

    He will have to read through a few guides and webpages in order to get a working system, compared to reading a single webpage which explains how to flash any other distro on a usb and be done with it.

    why don’t you use Kali Linux as a server?

    The advantage of Kali is that it is designed to live in ram and everything you do is destroyed when you switch off the computer, this is a bit of a pain in the ass if you want to run a server.

    why don’t you use Batocera or SteamOS as your daily driver?

    Don’t even know what those are, but pretty much because I don’t care: the system I have is good and I know there’s little difference between distros.

    why do you trust a community distro more than a corporate distro?

    They’re for different purposes (mainly). Redhat provides tech support. Canonical, well I don’t know what canonical does. If you want good support for maybe a large installation with many computers, paying for red hat may very well be worth it.