Migrating content from old forums

I would consider a conversation as “active” is there has been replies within the last 2-3 month. In order to have a easy possibility to continue a ongoing discussion/topic.

The old fluxbb forum with its bespoke Basic Auth integration can’t
handle their user name: ᚛ᚐᚋᚐᚅᚇᚐ᚜ :rainbow_flag: | OpenStreetMap
https://www.openstreetmap.org/user/᚛ᚐᚋᚐᚅᚇᚐ᚜%20🏳️‍🌈

I think nobody volunteers to fix this in the few remaining {weeks, months}.

Correct. I think it’s because the forum uses MySQL which doesn’t handle the :rainbow_flag: character

It happens with the OSM Wiki too, it deletes everything after you save it. :rofl:

If I’m not wrong, option 1 is clearly in favor from the replies posted here.

How do we reduce the “cons” that were mentionned in the initial post ?

A migration test will answer many “cons” listed for option 1, the only blocking one I see is “We’ll require from an Ops person to handle the whole process. AFAIK, currently we don’t have anyone with this kind of bandwidth.”

I’ve proposed to spend some time on migration tests, asked about the “beta” migration that seems to have been done (is it documented somewhere ?).
Anybody else ready to spend time to give the migration a try ?

4 Likes

When I asked about this in the Forum thread where we were discussing consolidating forum/QA/mailing lists, the majority of people responding there preferred not migrating content as long as the content was still available as an archive (and existing links worked). This would allow the new forum to decide on a new and different organizational structure and encourage new topics instead of listing a bunch of older topics. Think of it like the TIGER import on an empty OSM.

Can someone interested in migrating the old content to this new forum talk about why they prefer that option? What is gained beyond an archival snapshot?

To be clear, I’m not suggesting that we delete the old forum. I want to maintain that content and keep links working as much as possible.

1 Like

What about the reasoning in posts 10, 11, 15, 32 and 33 above?

1 Like

Hmm, maybe I’m misunderstanding which posts you’re talking about, but I’ll take a closer look and try to extract the comments above:

This is valid, but I would argue that people searching for problems will come in from an external search engine before searching the forum directly. This might result in them finding archived, read-only forum content and an answer their question. If not, starting a new thread on Discourse is not a problem.

I agree with this. To me, active means conversations within the last few days, not last few months though. As @nukeador suggested, it’s easier for people to participate if a long-running thread is split into smaller ones or summarized and restarted in a completely new thread.

I performed a test migration to a local installation of Discourse. It ran on a local Docker container so I didn’t get a chance to save the results and it wasn’t usable outside my laptop anyway. At the time, there wasn’t enough information to link the old FluxBB user account and the new Discourse user account. I believe the problem was that the FluxBB database didn’t have email address or the Discourse account didn’t have email address because it wasn’t exposed as part of the existing OSM OAuth1 handshake.

This resulted in a bunch of “orphaned topics” that could never be linked to their real owner if they ever signed in to the Discourse instance. These orphaned topics cluttered up the interface quite a bit and would make it pretty difficult to participate in the new Discourse-based forum.

I imagine this account problem could be remedied with more access to the database, but that would require approval and supervision from the sysadmins to protect user information.

1 Like

Right, email addresses have never been exposed before, and the email addresses you would find in fluxbb have been manually entered by a user. The current logon on fluxBB is using Basic Auth, and you would have to log on using your “display_name” (in osm.org terms).

After more thinking about this, we don’t actually need to have Discourse handle the redirects.

The migration script can generate a list of past topic/post id, and corresponding id in Discourse (I’ve seen that in other Discourse import scripts.

Based on this, some quite light code on forum.openstreetmap.org could be only handling the redirects to new URL on community.openstreetmap.org, not having to add 1M redirects in discourse itself.

This also avoids to have a old link to “forum” to be redirected first to “community” to get redirected to the new URL by discourse.

When I did the migration from phpBB to Discourse for the french forum, we had not OAuth2 on OSM.org, and no link has been done with OSM accounts at that time. This was not a problem because phpBB was not linked to osm.org accounts.

When OAuth2 became available on osm.org, I added OAuth2 login to our Discourse, and one nice thing is that you can link your existing discourse account with your OSM account afterwards (no need to access email addresses).

It is a possible path, unless we want a mandatory OSM account linked to each Discourse account.

More migration tests seem necessary from my point of view, possibly on a public test setup to have feedback.

3 Likes

Let’s keep this proposal open for input and feedback until March 29th.

After that date the @forums-governance team will take in consideration all the input and take a decision.

Thanks!

As there might be migration tests necessary to come to an informed decision, I think that a deadline for this topic is not good. (Or at least 1 week is too short)

1 Like
  • looking for the answer if a topic has been already discussed would not bring a result here. So Users not aware about the archive start discussions from scratch and potentially try to reinvent an existing wheel.
  • sooner or later someone for sure will challenge to delete the old forum due to security concerns, maintenance cost/knowledge or all of it
5 Likes

Isn’t that happening most of the time anyway? I mean, looking back at a few years of Overpass related questions, I see very similar questions coming up all the time.

Maybe the poor search on the forum is to blame here, or maybe people just don’t take the time to go through all those old questions, which are even outdated and irrelevant at times. Similar experience with Help OSM, by the way.

Tagging discussions may be different, though. However, there’s so much material out there that you really need to know what to look for. It’s very daunting for newbies trying to find their way through years of history. It’s even worse when there’s no consensus in discussions.

This is why I’m suggesting a static archive. We would generate HTML files for each forum, topic, and page on forum and put them on a hard drive instead of passing it through old/unsupported PHP.

@lkw We can see if there is something missing from the issues and blockers @iandees found in the tests he did a few months back. If you have additional things to test in mind we can create a list and see if there is someone with time to run them.

Yes, please migrate all the content in a new Forum.

Thanks for participating @WST1961

It would be great to get a bit more detail on you feedback, specially on the things that are important for you to solve, so we can better understand how to choose/craft an option that balances pros and cons.

I didn’t do not so much work, but I experienced again and again that suddenly so-called archive forums were lost. In 2022 it is normal for me and many people on the world, that is possible to get old informations. I hate Archive Forums, allways to do a search twice and so on.

1 Like

I see, as @iandees commented, there is a commitment to keep any form of archive always online in the future (also because it won’t need any special software because being HTML files).

Also, there are many ways to secure a few “backups” of that archive just in case there is a temporal downtime at some point (archive.org, IPFS copy, github/gitlab pages…)

Would that solve your concerns about it?

Hi everyone! :wave:

I’m quite new to the participating side of osm, after many years on the using side, so I don’t have a (long) history with the old forum and therefore maybe some fresh and hopefully valuable view on the topic…?!

When I read through the forum, I stumbled into the announcement, for the new discourse based place and I first thought “Cool, everything’s better then this dusty old PHP-pile and discourse has some cool modern features!” (I’m still thinking that way!)

When I heard the first “complaints” for migration, I thought “C’mon, no one’s reading the old posts anyway. Let’s ‘just’ put it in read-only and have a fresh and clean start at the new place!”

But when I read the reasoning pro migration, my opinion shifted step by step.

Furthermore I’m wondering about the handling of this discussion:
Why rushing towards a (kind of) point of no return, by populating this instance with new categories?
Why not even testing the migration, as @cquest offered his help?
I think some testing is crucial for any decision making.

Note, that the corresponding script seems at least to be maintained until 2020.

  • You’re lowering the bar for users from the old forum. Software change instead of communication channel change as mentioned above.

Link isn’t working. Could you please say more about these tests?

might! the script includes a translation from bbcode to md.

Not with some proper redirection. But that’s something, someone has to implement, yes.

DB yes, but why is that a problem? Discourse should handle this?!

Not with some proper redirection. see above. And the wording “penalty” is kind of misleading imho.

Tests will give more insight…

best Daniel

2 Likes

The links points to a previous message on this topic, it’s working here.