Discovery blog has some nice articles but it is not as mature as Arch Wiki. The community usually consults Forums or Telegram for help, but such answers get documented nowhere.
I built a new knowledge-system to fulfill that gap. If a user asked a question, it cannot be directly answered. Instead, you must contribute a self-contained paragraph, called “snippet”, then cite it into the question. A collection of snippets act as a wiki.
Accordingly, as users’ questions get answered, a wiki for Endeavour OS gets progressively built! Moreover, eOS snippets may cite advanced Arch-style foundational snippets like Manual Prefix.
Here is an easy gaming guide for eOS. An eOS user may ask a question here.
If the community found it useful, we may link or even deploy an instance of it, in eOS website.
I LOVE the re-imagining. Bold things come from this.
Too dry and clinical for me personally. I like a human giving context, anecdotal failures, what to watch out for, the whys of elements are important for a gaming rig, why this is the best way, etc etc etc.
Others may love the cut and paste straightforwardness of this so just offering opinion.
AI is useful if you can ask the right question within the right context. In such case, you don’t need to consult any kind of documentation.
However, AI is bad in learning foundations. At some point, you must build-up a background from books and wikis.
Imagine if quick answers do progressively cite foundational pieces, until the Dragons book is reached! so that you read pieces from the book, as you navigate through quick answers.
If like you AI, we may think of RAG on top of the knowledge-base platform; It is too early now, though.
Yeah! If the community loved it, I may use Solid or AT Proto so that Linux foundational snippets could be immediately re-used by other Linux distro communities!
I like Reddit for that, to gain a sense of what the community as a whole thinks of something.
High-level insight snippets could be built on top of foundational snippets. You may see this question for example, for the transition from user’s questions to foundations, through citations.
I wouldn’t say it is designed for quick copy-paste solutions. It is designed to bridge quick solutions with foundations by citing “snippets”. A wiki could be a collection of quick-solution snippets, or a collection of advanced foundational snippets.
I think a little hand-holding is a good thing, especially for rookies/newbs (or common idiots like myself) that may gravitate to your gaming tutorial. I love the example you furnished–that iswhat I’m talking about more or less.
Thanks for the thorough response. Wish you success going forward.
I like the format, easily readable, straight to the point and easy on the eyes.
I’m not sure why you would recommend anyone doing wine prefixes manually now days when you have Heroic, Lutris, Bottles or other like launchers. But if it’s there just for information if someone wants to read up on how wine prefixes work then, a good source for information
Technically spoken, as the solution to an unresolved question could be marked by the original creator of an initial post, there is some documentation within that forum here.
And just based on my activities here, the links I’ve shared in the past, even if it’s pointing to the specific subsection of the Arch Wiki… that seems to help out others.
Therefore, as certain topics in this forum are still accessible via searches that are also indexed by the most common search engines.
Well, you could essentially scrape the forum and taking these aspects into account. At least from a technical perspective. But if the community would appreciated it to compose a knowledge database based upon solved troubleshooting guidelines, that is indeed a different question to take into account.
From my point of view: The Arch Wiki is a splendid resource. And is my first choice to find answers to specific questions and totally applies to EndeavourOS as well. But due to whatever reasons, some simply forgot the fact that EndeavourOS is still arch-based. And neither check the Arch Wiki. Nor do they make use of the package database hosted on the Arch sites.
If you liked the example, I’d be happy to contribute more similar examples, driven by what the community is looking for. This will be win-win for both of us, as I currently need feedback.
I am happy to listen to any other feedback you have.
The motivation is similar to manually setting up vanilla Arch, even though we have eOS.
Having a quick solution is nice, but a Linux Power User needs to see what happens under the hood. That enables figuring novel solutions more productivily.
Yes. A Forum could serve the same purpose. However, it incentivizes a conversational style, where responses are tailored for a particular question. Its user experience or interface design, does not encourage building a reusable knowledge-base.
“Snippet” knowledge-base system could be designed, such that any contributed snippet must cite a page from the Arch Wiki, enforcing quality contributions.
Good idea.
That’s exactly the point I am currently looking for. eOS community may not always rely on Arch Wiki, but it shouldn’t also expect every problem to be solved by googling or LLM. eOS is branded for an intermediary Linux user.
That’s what has motivated me to build Snippet platform, bridging foundational Arch wiki, with quick fixes mentioned in questions.
I tend to disagree. When I had to setup a debian server with docker, hosting docker containers for nextcloud, immich, adguard and jellyfin, perplexity helped me big time with that.
I started with zero knowledge about docker. I started with asking perplexity the basic stuff about docker, how to setup a reverse proxy and how to do geoblocking on that server. This was an amazing experience for me. My first contact with AI and I was really impressed.
And perplexity also gave me the links to the original documentation. That also comes in pretty handy for a deep dive.
And you do not need to have special skills in “asking the right questions”. You can do it on smalltalk level. And the answers sometimes point you in new useful directions. Directions which would never come to your mind when you just read the documentation.
PS
I still prefer perplexity over chatgpt when it comes to IT topics. I have a free account with perplexity. And I had a paid account with chatgpt and tested it for ca. 3 month. The answers of chatgpt are not as good and comprehensive as with perplexity. For IT topics perplexity is my #1 knowledge base.
Perplexity works great when it is based on a solid documentation. Your experience is going to be even much better, if Perplexity is integrated to a knowledge-base of eOS and Arch.
Think of “Snippet” as some kind of documentation, Perplexity would cite. The goal is to envision an intermediary arena between Arch Wiki and eOS forums, so that you could pick-up the foundations more productively, using Perplexity or any other AI.
Well, from my point of view… there is a specific reason why the rather small team of EndeavourOS as well as the community hasn’t tackled to overhaul, extended and improved Discovery yet.
Resources are limited, documentation in the sense of copying the Arch Wiki are more or less pointless. And there is that critical aspect that your snippet-based knowledge base would require active contributors and maintainers. That is a collaborative effort which essentially require some coordination, guidelines and would have to be adopted.
The current forum is a well-established tool which doesn’t require additional administration… and focuses on answering open questions. And in some cases, just pointing novel users towards the facts that there are the manpages. Or on package related issues, that the arch package database already suggests a solution he simpy hasn’t been aware of.
Additionally, there are other tools which aren’t exclusive to EndeavourOS in the first place.
For instance, tealdeer is a simple helper that essentially brakes down the complexity of manpages by composing an tl;dr summary based upon most common tool uses.
Another complication is more or less : If there is a user who is struggling with his hyprland configuration that he is using, this forum might not be the best place as it’s not the official DE/WM EndeavourOS is using.
There is a large diversity of topics. And just a few authors that would compose new “knowledge snippets” - within the scope of their own comfort zone - might not be capable to address more involved topics that require specific expert knowledge.
The simple forum based interaction - of encouraging users to ask for help and “spreading” the support in the replies to active users beyond the scope of the developer team itself, does work. The forum is still indexed by search engines as most of the other resources are. And, for me as an active member of this community, I’ld rather prefer “on-demand” assistance. Instead of working on a knowledge database that would eventually provide a better user-experience.
Something which even larger linux distributions are struggling with.
Sometimes documentation does not provide the answers because in a lot of instances there is some small piece of information missing either not provided or not explained in a way that everyone could understand. Arch Wiki is a good example of this.
The eOS community is already active on Telegram. We don’t need more resources; we need to well-utilize volunteers’ time.
You are right. My premise is the effort required by maintainers is much less than the effort maintainers do already exert in Telegram and Forums. re-using snippets is more productive than re-answering every question.
I am happy to listen to all maintainers’ requirements. The platform could be tuned accordingly.
I like the idea of pointing users to man-pages. “Snippet” project takes that idea a step forward, so that every question gets pointed to some re-usable piece of knowledge.
The forum is full of redundancy. For example, it is very common to ask for system information like the kernel version. If we have a handbook of common system diagnosis commands, then that should save you much time.
“Snippet” knowledge base has layout for eOS in particular, and “Linux” more broadly. Tools which aren’t exclusive to eOS could be in “Linux”. The nice thing is that you may benefit from other distros contributions in “Linux”.
tealdeer is a wonderful project. We could integrate it into “Snippet” knowledge-base system, so that searching and citing it is more accessible. We could even enforce any contributed snippet to be based on some tldr, man-page, or Arch Wiki. We could use AI to suggest citing tldr without enforcing it.
By the way, If “Snippet” knowledge-base attracted some quality foundational contribution, where it got highly cited in many questions, then that snippet could be elected for the Arch Wiki or tldr.
“Snippet” does not aim to replace Arch Wiki, tldr, or any kind of community effort. It aims to bridge wiki projects with users’ high-level questions, as a win-win for everyone.
If you want to isolate some set of questions, we could create a separate knowledge-base, naming it “eOS - unsupported configurations“ for example.
I agree.
“Snippet” Knowledge-base system could accept contributions from both maintainers and any active user. It is easy to pin or highlight some “snippet” or “wiki” created by maintainers, over “snippets” created by any user.
The knowledge-base system has a section for questions. Once someone asks a question, on-demand you may contribute “snippets” then immediately cite them in the question. See this question as an example.
The whole motivation of “Snippet” is to incrementally contribute to a knowledge-base, on-demand of newly asked questions, saving the burden over maintainers to build a wiki.
I am happy to learn from your feedbacks.
P.S. “Snippet” knowledge-base neither aims to replace Telegram for real-time interaction, nor forums for discussions tailored for a particular question.
Writing about very technical topics is difficult, takes a specific skill set that not all people tasked with that kind of authorship are capable of, takes a lot of time and mindfulness, also putting yourself in the other’s shoes (using this metaphor is surely a good example of how one should not write for an international audience)
Sometimes its not about the effort to be active on telegram and forums. It’s about staying active and keeping the community engaged. If you read the history of it started with forum first and then the distro. An active and engaged community is the main reason why most of us are here.
it started with forum first and then the distro. An active and engaged community is the main reason why most of us are here.