val @val@oc.todon.fr

Comment on a tous pu louper ça il y a 13 jours, je vous le demande.
https://www.youtube.com/watch?v=IB_lTTxYSsg

Are you a person who was around in, or knows a lot about, 1970s/1980s fandoms? I'm working on a project about spoilers and want to learn more about early uses of the term in its modern sense. I'm reading Usenet archives; do you have memories of earlier emergent uses? Please share! Non-fandom memories and leads also welcome, of course.

Comme on dirait qu’il y a des #debian users ici, existe-t-il un équivalent de Spacewalk ou RedHat Satellite pour gère le patching d’un parc de serveurs Debian.

Why Not Mars (Idle Words)

https://idlewords.com/2023/1/why_not_mars.htm

I love this blog, and this is a very interesting post. It changed my (uninformed) opinion on the usefulness of a Mars mission.

Confused by Linux? I made this handy dandy chart to help clear up the confusion! #GNU #Linux #OpenSource

i have a lot of thoughts about the activitypub proxy tool. the tool is not surprising to me and, although the readme unironically uses the phrase “woke admin,” the tool itself does highlight the point of total instance defederation largely being used as a tool for admins to enforce their own personal boundaries, at the inconvenience of their users.

there are far too many incidents of admins defederating other instances over some beef with an admin or user, resulting in people’s social relationships being severed (frequently without their awareness!) over some incident that could have been easily mitigated with other tools.

given this it is unsurprising that somebody would eventually say “fuck you, admin!” and work around the nuclear nature of mastodon instance moderation, if only to make a statement.

the real solution is to shift the role that consent plays in the protocol implementations. activitypub has a large framework for requesting, and affirming, consent. if we had a network of consent, then instance defederation becomes a much less justifiable hammer in the toolbox.

This makes git so much clearer for me.
From https://twitter.com/NikkiSiapno .

My organization (Software Heritage) is using more and more #pads (using #HedgeDoc) for real-time collaborative elaboration of documentation, presentations, etc.

However, once completed, we have a hard time referencing the content, indexing it, archiving it. Completed documentation pages land in our #SphinxDoc project, which is properly indexed and hyperlinked, but not all of the pad content needs to land there.

Do you know of groups that have existing strategies to make sure that the knowledge stored in pads doesn't end up stuck in that black hole forever? I would especially welcome #git or #GitLab based, automated workflows. #Archivists #KnowledgeManagement

I've found the CERN CodiMD archive project, which seems promising, but, hilariously, only referenced on their own CodiMD instance. I haven't found any code for it.

Seems like #GitLab has killed its real-time collaborative edition project as well, so that seems to be a dead end.

(cc. @douardda @zacchiro, boosts welcome for reach)

Bonjour à tous,
Je chercher une solution de planning / prise de RDV.
Via un lien, les personnes pourraient réserver sur un créneau libre et entrer des infos, les créneaux déjà pris n'étant pas accessibles et bien sur les informations non visibles pour les autres.
Seul l'admin verrais les infos.
Une idée de vers quoi m'orienter?
Eventuellement chez @zaclys ?

This one makes me laugh every year… because it is so very true.

The Turing Test poisoned the minds of generations of AI enthusiasts, because its criteria is producing text that persuades observers it was written by a human.

The result? Generative AI text products designed to "appear real" rather than produce accurate or ethical outputs.

It *should* be obvious why it's problematic to create a piece of software that excels at persuasion without concern for accuracy, honesty or ethics. But apparently it's not.

Stop saying there are no algorithms. Algorithms don't necessarily involve opaque machine learning-driven decisions. Showing posts in chronological order is an algorithm. Showing a partial ordering of posts coming in from other nodes in a federated distributed system is an algorithm. Anything involving computers uses algorithms.

Even if you can predict what it does, there's an algorithm. ESPECIALLY if you can predict what it does.

Thank you for coming to my Ted talk.

Seriously, multiple times today by very different people : stop sharing posts made only of numerous emojis that have no CW. Don't. Either repost with a CW or don't repost...

Ne partagez pas de murs d'emojis sans content warning, c'est un enfer pour toutes les personnes qui ont un lecteur d'écran...

Juste out of curiosity, how’s your delivery towards MS emails from your instance?
(hotmail, live, outlook, etc.)