@tauli A problem that is probably not impossible to solve.
Without wanting to go into details here, because there are probably smarter minds involved with federation who could come up with a better strategy, this is a simple idea to approach this:
- instances are often specific to topics, e.g. space, biology etc, similar to how the StackExchange network already works.
@tauli Instances with related topics could follow each other to also search their index (or replicate the remote content in a local search index) to provide useful results from other instances.
This allows to keep the overall index small enough, because you don't have to index everything but only instances relevant to your's focus.
Not just topics, but related questions. That feature of SO is extremely useful when you initially find questions that are almost, but not quite your question. That would be fairly easy to add to a federated system.
It would take ages… Firstly, you would need a well structured website. You would need many people filing entries. And then you have to push it through social media. In Germany, I always use selfhtml.org for the basics of HTML, CSS and JS. For PHP, I mostly use php.net and only for specific questions I have to look somewhere else. Of course my answer is to be found on stackoverflow -.-
Actually since yesterday I'm pondering about the idea to build a #federated version of stackoverflow, nothing written yet, I'm reading, researching.
Also, right now I was checking this stack exchange sqlite db under CC BY-SA 4.0 to check how useful and doable would be import this data and using as a base for the federated version.
Also wondering if we could use this data somehow to train our own opensource AI to help the community, but I'm do not have knowledge on LLM/AI things. Please if there is any expert I would appreciate the opinion on that.
A better place to download the dump content, with more interesting tables, like the one with the Votes; the other link the dumped data only contains two tables Users and Posts. Right now downloading the whole data related with stack overflow, and will take some time due my humble home internet connection, so I didn't have the chance to take a look at the data, but I guess that's the interesting thing.
I disagree, unless search engines change significantely a Federated alternative to Stackoverflow wouldn't really work. Their reliance on domains is tough enough for a social network like Mastodon which doesn't even rely that much on search engines and (at least in my opinion) a death sentence for alternatives to many other platforms.
the data is actually dumped and stored somewhere accessible. The license for each post is some creative commons thing that requires attribution. Given the latter, it would be possible/legal to move everything to a new place, but I gleaned from discussions on SO meta that the codidact people ran into tall technical hurdles.
@Sibshops Yeah, LLMs are a bit different than Stack Overflow. Sure, they don't insult you for asking questions, but they can't refuse to answer, so will happily give you nonsense if they have no idea. Not entirely unlike some people... ;)
@Sibshops We asked about a knowledge base. LLMs will never replace this, because they can only reproduce, not answer new questions. They can be your resource intensive search feature for a knowledge base, but that's not what we are interested in.
I'm not sure that a decentralized knowledge base is a good idea. It should look more like Wikipedia — run by a nonprofit organization, on open source software, allowing exporting dumps
I wonder how a truly Fediverse alternative to #StackOverflow would look like, given that #AskFedi is a thing. If the questions would be normal ActivityPub messages (everyone can read them) with a common hashtag for subscribing/muting. And (public) answers from everywhere would show up one the page, but could be curated and amended on the target server, maybe we would already be close.
blogs. They often have answers not found on stack overflow. And they can have answers from stack overflow(it appears this is legal now). Copied answers can have attribution,” and alternative/better presentation.
If search on Mastodon improves, bloggers who post here will have their answers discovered and commented on. Google will still rank them.
But the main point is that blog owners will continue to own their content and may even choose to block or even subvert ML spiders.
Themes are expansive customizations that change multiple elements of the style of your forum design, and often also include additional front-end features.
A Q&A platform software for teams at any scales. Whether it’s a community forum, help center, or knowledge management platform, you can always count on Answer.
Thiago, the one called 'Jedi'
Als Antwort auf Codeberg.org • • •Codeberg.org
Unbekannter Ursprungsbeitrag • • •Sensitiver Inhalt
@evil_puniko The problem is to build a knowledge base. The amount of useful data StackOverflow accumulated in their realm is a major vendor lock in.
Any serious alternative would need to go through a long way to fill its database to become useful.
It's likely that a federated alternative would make it much easier to succeed here.
Michael Baumgartner
Als Antwort auf Codeberg.org • • •Samuel Philipp
Als Antwort auf Codeberg.org • • •Codidact - Helping each other learn.
www.codidact.orgtauli
Als Antwort auf Codeberg.org • • •Codeberg.org
Als Antwort auf tauli • • •Sensitiver Inhalt
@tauli A problem that is probably not impossible to solve.
Without wanting to go into details here, because there are probably smarter minds involved with federation who could come up with a better strategy, this is a simple idea to approach this:
- instances are often specific to topics, e.g. space, biology etc, similar to how the StackExchange network already works.
Codeberg.org
Als Antwort auf Codeberg.org • • •Sensitiver Inhalt
@tauli Instances with related topics could follow each other to also search their index (or replicate the remote content in a local search index) to provide useful results from other instances.
This allows to keep the overall index small enough, because you don't have to index everything but only instances relevant to your's focus.
naught101
Als Antwort auf Codeberg.org • • •Sensitiver Inhalt
@tauli
Not just topics, but related questions. That feature of SO is extremely useful when you initially find questions that are almost, but not quite your question. That would be fairly easy to add to a federated system.
Zeronior
Als Antwort auf Codeberg.org • • •In Germany, I always use selfhtml.org for the basics of HTML, CSS and JS. For PHP, I mostly use php.net and only for specific questions I have to look somewhere else. Of course my answer is to be found on stackoverflow -.-
char
Als Antwort auf Codeberg.org • • •Actually since yesterday I'm pondering about the idea to build a #federated version of stackoverflow, nothing written yet, I'm reading, researching.
Also, right now I was checking this stack exchange sqlite db under CC BY-SA 4.0 to check how useful and doable would be import this data and using as a base for the federated version.
Also wondering if we could use this data somehow to train our own opensource AI to help the community, but I'm do not have knowledge on LLM/AI things. Please if there is any expert I would appreciate the opinion on that.
seqlite.puny.engineering
EDIT:
A better place to download the dump content, with more interesting tables, like the one with the Votes; the other link the dumped data only contains two tables Users and Posts. Right now downloading the whole data related with stack overflow, and will take some time due my humble home internet connection, so I didn't have the chance to take a look at the data, but I guess that's the interesting thing.
here:
archive.org/download/stackexch…
SEqlite
seqlite.puny.engineeringGamey :thisisfine: :antifa:
Als Antwort auf Codeberg.org • • •Sensitiver Inhalt
Henning Deters
Als Antwort auf Codeberg.org • • •Henning Deters
Als Antwort auf Codeberg.org • • •Sensitiver Inhalt
Anthony Rabine
Als Antwort auf Codeberg.org • • •Baggypants
Als Antwort auf Codeberg.org • • •Sibshops
Als Antwort auf Codeberg.org • • •unfa🇺🇦
Als Antwort auf Sibshops • • •Sibshops
Als Antwort auf unfa🇺🇦 • • •@unfa
Yeah, I know they give nonsense sometimes. The nice thing about using it to code is you can easily check their nonsense to see if it is right.
A lot of times you can tell them that this is nonsense and how to fix it and they will.
LLM's have not been very helpful with visual things like CSS or asking them to create an SVG. But with backend code they are pretty good.
Codeberg.org
Als Antwort auf Sibshops • • •WerySkok :verified_think:
Als Antwort auf Codeberg.org • • •unfa🇺🇦
Als Antwort auf Codeberg.org • • •Florian Schmidt
Als Antwort auf Codeberg.org • • •If the questions would be normal ActivityPub messages (everyone can read them) with a common hashtag for subscribing/muting. And (public) answers from everywhere would show up one the page, but could be curated and amended on the target server, maybe we would already be close.
Iizuki
Als Antwort auf Codeberg.org • • •Codidact
www.codidact.comDmitri ☕️
Als Antwort auf Codeberg.org • • •blogs. They often have answers not found on stack overflow. And they can have answers from stack overflow(it appears this is legal now). Copied answers can have attribution,” and alternative/better presentation.
If search on Mastodon improves, bloggers who post here will have their answers discovered and commented on. Google will still rank them.
But the main point is that blog owners will continue to own their content and may even choose to block or even subvert ML spiders.
kir0ul
Als Antwort auf Codeberg.org • • •There's topanswers.xyz/ as alternative, but I don't know of any #federated alternative yet.
Maybe @samcarter knows if that's something that have been discussed?
nieebel
Als Antwort auf Codeberg.org • • •what about dev.to as #StackOverflow alternative?
#AGPL #GPL #FOSS
DEV Community
DEV Community6543
Als Antwort auf Codeberg.org • • •Well #lemmy with a dedicated decentraliced indexer like yacy.net/ but pirpose build and integrated could fill the gap.
So no there is project on it.
Well smal forges have a similar discovery disadvantage if compared to e.g
github.
So a P2P indexer that could be integrated per network and would get content by that network injected instead of stupid web crawl could solve that.
Similar as to how you can use milisearch and integrate it purpose build into projects.
Home - YaCy
yacy.netBlaise Pabón - controlpl4n3
Als Antwort auf Codeberg.org • • •#Discourse is really sweet to host and it has a vast plugin community.
discourse.org/plugins
And themes to control what fields to expose
meta.discourse.org/c/theme/61
Fwiw, some of the discourse devs were in the original stack overflow team.
theme
Discourse Meta💡𝚂𝗆𝖺𝗋𝗍𝗆𝖺𝗇 𝙰𝗉𝗉𝗌📱
Als Antwort auf Codeberg.org • • •programming.dev - A collection of programming communities
programming.devgrindhold
Als Antwort auf Codeberg.org • • •Sensitiver Inhalt
Daniel Opitz
Als Antwort auf Codeberg.org • • •Apache Answer
answer.apache.orgKaleb 🚫
Als Antwort auf Codeberg.org • • •lots of mentions of #Codidact, but it doesn’t look like anyone tagged @codidact Maybe they’ll have something to add.
#StackOverflow
Frank Ring
Als Antwort auf Codeberg.org • • •Elephant (free 🇵🇸)
Als Antwort auf Codeberg.org • • •https://programming.dev
programming.dev