What happened to all the posts from the old Askbot-based Ask Fedora site in 2019?

[ankursinha (FranciscoD)]

@fasulia

These are all matters of attitude and prioritization, not necessarily resource or technical feasibility.

Maybe resource and technical feasibility inform our attitudes and priorities?

@fasulia

I’m running out of ways to phrase the frustration…

So what are we doing here? Trying to change the team’s attitude? Should we apologise? Should we move back to askbot to prevent broken links?

I dont understand the goal of this discussion anymore

@fasulia

If the issue is resources or technical limitations, these could be put as challenges to the community to find solutions. It’s not clear to me that either of these are insurmountable, but if that is truly the case, surely most would be understanding. At the very least there would be warning and transparency in the matter.

Why do you assume we didnt put them to the community? Seen the community blog?

OK fine, these arent insurmountable, so how are they to be surmounted? I’m not asking for discussion of attitude here. I’m requesting implementable solutions.

[mattdm (Matthew Miller)]

For what it’s worth from my point of view: we really did look at possibilities. There just aren’t any that aren’t a project unto themselves, and part of the whole reason for this shift was that we don’t have a lot of spare resources for such projects. No exaggeration: with the capacity we have for projects like this, preserving existing links even in a degraded form would have delayed launching the the new Ask by one or two years.

I know there is value in the old links. This is not a problem of not understanding that. It’s making a judgment call balancing that against the other realities.

[fasulia]

@FranciscoD

The solution being to drop the ideology?

Ideally

@FranciscoD

Why do you assume we didnt put them to the community? Seen the community blog?

No I have not seen the community blog. I don’t know where or what that is. I also did not frequent AskFedora regularly. After upgrading to Fedora 30 I had some questions and was suddenly faced with broken links and a totally new platform without much info on what happened. Based on this and other threads, it’s safe to say that’s a common experience.

@FranciscoD

OK fine, these arent insurmountable, so how are they to be surmounted? I’m not asking for discussion of attitude here. I’m requesting implementable solutions.

Is the user responsible to come up with implementable solutions? Is it not sufficient to point out usability issues? It’s great to consult the community on issues and suggestions, but let’s not dump the responsibility on the wrong party.

@mattdm

For what it’s worth from my point of view: we really did look at possibilities. There just aren’t any that aren’t a project unto themselves, and part of the whole reason for this shift was that we don’t have a lot of spare resources for such projects. No exaggeration: with the capacity we have for projects like this, preserving existing links even in a degraded form would have delayed launching the the new Ask by one or two years .

I know there is value in the old links. This is not a problem of not understanding that. It’s making a judgment call balancing that against the other realities.

This is the best response and explanation I’ve seen so far. During this transition period, I would suggest making a sticky with such a statement, optionally with possibilities considered and reasons for their rejection.

Some may still question the judgment call, but at least recognizing the issues created, the reasons why, and possible remedies, may provide some solace and help move the discussion on to more productive topics.

[ankursinha (FranciscoD)]

@fasulia

Is the user responsible to come up with implementable solutions? Is it not sufficient to point out usability issues? It’s great to consult the community on issues and suggestions, but let’s not dump the responsibility on the wrong party.

This is the crux of the issue.

FOSS is not “vendors” providing solutions to “users”. It is “communities” working together to achieve a goal—even if it is merely to develop an OS. When users engage with the community, they become part of them.

So, you are not a “user” in the sense of corporate software, and we are not providing “support” in the same way. You are a community member that has identified a problem and is discussing it with the rest of the community to find ways of improving the situation. It is why everything is discussed openly with the community—this doesnt happen in the “vendor” “user” model. The “vendor” identifies a service that will maximise their returns and focusses efforts on developing it. There is no community that together decides what the best way forward is. If the "vendor"s plan works, they get rewards, if it doesnt, they dont.

So, it isn’t dumping responsibilty—it is about sharing it so that even with our personal resource limitations, the work can still be done. It cannot be done if a task that requires 130 hours of work is asked from individuals that only have 15 hours of time to volunteer (15 hours a week is a made up number to illustrate the issue. It is already too high a number for most volunteers who do full 40 hours a week jobs, and squeeze out time for their families and their volunteer work).

The sum total of community resources only increases when more individuals come together to pool in their personal resources.

@fasulia

making a sticky with such a statement, optionally with possibilities considered and reasons for their rejection.

This:

Is there a benefit to using the registry.fedoraproject.org/fedora image vs Docker Hub's fedora image?

IMO this is something that should be documented somewhere…

@fasulia

Some may still question the judgment call, but at least recognizing the issues created, the reasons why, and possible remedies, may provide some solace and help move the discussion on to more productive topics.

I’ll summarise to make this easier to follow:

• recognised the issues

  • this topic, and the other one

• explained the reasons why

  • again, on this and the other thread
  • here too

• implemented possibile reemedies:

  • Moved common questions to quick-docs and added a list here
  • HTML only read-only archive of askbot, also here
  • attempt redirects
  • updated 404 and attempted redirects from Discourse

• requested ideas on other possible remedies:

  • attempt 1 of N here

• More information on the planning

Please, continue to suggest solutions—if it is within the resources of the community (which increase if you join it and help ) to implement them, we will do so.

[noda (Daniel Aleksandersen)]

@FranciscoD, I’ve come up with a real solution. It just requires one redirect to be configured. Any chance you could poke the right people to get this done?

I’ve submitted all the 26 500 questions (and their answers and comments) from Askbot to the Internet Archive. Which I could do with Askbot as everything was Creative Commons licensed. (The new site should be as well.) It’s a shame they were submitted from askbot.fedoraproject.org instead of their original locations but it’s better than nothing. (Note that it can take a few hours or days before all the submitted URLs become available in the Wayback Machine.)

Any chance you could setup a redirect from https://discussion.fedoraproject.org/en/(.*) to https://web.archive.org/web/20190517/https://askbot.fedoraproject.org/en/$1 Where $1 is a back-reference to the .* expression. This URL schema should work for every single English language question.

This should help keep the 25 000 English language posts available. Other languages have also been archived and can be configured in the same way.

Note that non-canonical locations and non-question URLs won’t work. The old content management system had dynamic URLs so there are probably a billion combinations that won’t work. The primary canonical questions and answers references will work just fine, however.

Two URL forms have been archived:
https://askbot.fedoraproject.org/$lang/question/$question_id/$question_string and https://askbot.fedoraproject.org/$lang/question/$question_id/?.

[ankursinha (FranciscoD)]

@noda

I’ve submitted all the 26 500 questions (and their answers and comments) from Askbot to the Internet Archive. It only took 2 hours. Which I could do with Askbot as everything was Creative Commons licensed. (The new site should be as well.) It’s a shame they were submitted from askbot.fedoraproject.org instead of their original locations but it’s better than nothing. (Note that it can take a few hours or days before all the submitted URLs become available in the Wayback Machine.)

Thanks very much for taking the time to do this!

@noda

Any chance you could setup a redirect from https://discussion.fedoraproject.org/en/(.*) to https://web.archive.org/web/20190517/https://askbot.fedoraproject.org/en/$1 Where $1 is a back-reference to the .* expression. This URL schema should work for every single English language question.

Unfortunately, it cannot be done from discourse. We tried that already and the redirects that we can do from the admin panel are basically permalinks to individual links and so cannot be used for this. I’ve asked our Fedora infrastructure team if they can look into it. Hopefully, they’ll be able to find a way.

@haghighi: could you re-open the pagure ticket please (since you filed it, I can’t reopen it).

[noda (Daniel Aleksandersen)]

@FranciscoD

Thanks very much for taking the time to do this!

It took less than a minute to start the process and it then took my computer 6 hours to slowly work its way through all the old posts. (Had to go real slow and nice to not put any undue strain on IA’s or Fedora infrastructure.)

@FranciscoD

Unfortunately, it cannot be done from discourse. We tried that already and the redirects that we can do from the admin panel are basically permalinks to individual links and so cannot be used for this. I’ve asked our Fedora infrastructure team if they can look into it . Hopefully, they’ll be able to find a way.

—and I guess you don’t want to add 26 500 redirects by hand?

Here is the entire individual list of redirects in Apache HTTPD configuration language:
https://gist.githubusercontent.com/da2x/57193594996bbddb6e0e27f052b06596/raw/a7a624b0a7bdce5d134353657f6c12318447c5cf/fedora-askbot-redirects.conf

[ankursinha (FranciscoD)]

The infra team provided an update, turns out it isn’t that simple either. Have a look at the response on the pagure ticket

I’ll ask Discourse and see how that works out.

@noda

—and I guess you don’t want to add 26 500 redirects by hand?

[noda (Daniel Aleksandersen)]

@FranciscoD From some quick searches it seems that you can use provide simple redirect patterns in the Discourse administration interface. There are at least several people asking for help getting their redirect patterns correct, so I think this suggests that patterns should be supported. I don’t know the pattern expression language as it isn’t documented, however. (I couldn’t really find much documentation for Discourse at all.)

This should work (assuming I’ve guessed the expression language correctly):

from: "^/([a-zA-Z]{2}(-[a-zA-Z]{2})?)/([0-9]+)"
to: "https://web.archive.org/web/20190517/https://askbot.fedoraproject.org/$1/$3/"

(Reference 1 is the language code and reference 3 is the question id number. These references may be off by one as I don’t know what matching engine is being used.)

[ankursinha (FranciscoD)]

So, the admin interface has nothing about redirects. What we have access to here, as part of our pricing plan are permalinks: What do Permalinks under Customize do? - support - Discourse Meta

It’s explained here: Redirect old forum URLs to new Discourse URLs - sysadmin - Discourse Meta and looks like a one to one mapping, not a pattern based one.

(I.e., there’s no way to access the nginx or whatever server from the admin interface to add these there)

[powergame (Mehdi Haghgoo)]

This is very bad and discouraging. Makes one tend to ask Fedora questions in other networks like Superuser and Unix Stack Exchange, because they are more stable. With the old AskFedora, I hoped we’d be building a comprehensive reference for Fedora users. Really sad they are wasted

[mattdm (Matthew Miller)]

I understand the frustration, but we’re just not resourced for that kind of thing.

Plus, I think the reality didn’t match “building a comprehensive reference”, because the tools for search, deduplication, moderation, editing, and continuous improvement in AskBot just aren’t there and despite our investment are realistically never going to materialize. So while there’s definitely some real loss, there’s also a big historical mess which would be a many-person multi-year project to digest in a useful way.

So, yeah, we’re starting fresh, but hopefully better this time!