Unite forums, OSQA, MLs, ... to a single discourse server?

The Belgian community is using Matrix/Riot for quite a while now. Unless we are using it incorrectly, I do not see how it can replace the forum and help sites. The chat rooms we are using do not have the possibility for threads. This is essential in order to replace the existing platforms.

@escada
No, you are using it correctly. Threads are currently not supported in Matrix or in Riot but it is beeing worked on, as far as i’m informed.

@Thread
Looking at the features, it seems possible. Migrating from FLuxBB https://meta.discourse.org/t/importing-from-fluxbb/51322 → i could give it a test drive with a test-fluxBB forum i set up on my webspace for the theme i made for the forum.
It seems like you have some kind of tree-based threads. You can reply to a specific post and it gets displayed directly as reply to the post. See https://meta.discourse.org/t/using-discourse-as-a-sso-provider/32974/5 this thread for example. Sort by likes seems to be possible, but comlicated :confused: https://meta.discourse.org/t/sort-filter-by-likes/26193 but you can mark an answer as “answer” so it gets displayed as that.

Looking through the feature site there are some advantages for users: Trust system, social login, summarize topics, Emojis, Reply with Email (for the mailing-list-users maybe?), maybe Badges as well. Language support: Chinese, Hebrew, Turkish and so on (https://www.transifex.com/discourse/discourse-org/). Quite some languages. Seems like image upload works as well, but cannot be sure.

For the admins: Community moderation (pretty sure better than the current one in FluxBB), Spam Blocking, admin dashboard, plugins, API and oneclick upgrades, autobackups, (and Mailing list support according to the feature site!)

I think discourse combined with matrix for quick communications like gatherings and quick questions would be something awesome. It seems like OSM could profit by that. I contacted CWG about matrix right now.

@Negreheb , @escada As said, it’s about discourse platform, which is more like a forum than a realtime chat / social network / … .

I agree with the suggested steps and thank you for the technical analysis. Unfortunatly I lack of time to do testing etc. and get in touch with CWG to do some tests on useful merge.

I think the fragmentation is ok, everyone can find the channel that works best for them. It means that your choices of technology might exclude you from some channels (me, I refuse to use Slack or Facebook) but there will always be other options.

Also, I think that we would at the very least need one threaded, asynchronous discussion medium (currently forum, mailing lists), a synchronous chat (currently IRC), and a Q&A system with up/downvotable answers (currently help). We will also want to keep the wiki, changeset discussions, and OSM user blogs with their comments.

Which basically leaves forum and mailing lists as candidates for “unifying”. And I don’t see the value. People often dream of one central place in which all discussions could be cleanly held and recorded for posterity, and everyone would be aware of them and they would not have to be repeated. I think that idea is flawed, because

  • We constantly add new members to the project anyway. For a new member, there will no difference between “we have discussed that in this forum before” and “we have discussed that elsewhere before”; the need to reiterate established consensus will always be there.
  • We are social beings and people must be able to choose the social group they want to be with. Some might mainly be on their regional mailing list, some at the pub meet, some on the national list, and some might prefer the international lists or forums. Some prefer email, some prefer web access, and others personal meetings or chats. Apparently some even use Facebook to discuss OSM topics. Many of these will always remain small groups (the very idea of pub meets is to bring local mappers together), and many people prefer discussing things in small groups. Or something gets discussed on a country list and later pops up in an international forum. It is natural to have discussions in many places.

Bye
Frederik

Thank you for your hints.

Concerning the ability to chose social groups, I don’t see any limitation to subtopics / groups / category to filter the (push) messages.
Concerning the need the “we have discussed that before”, isn’t that a good point to a unified information resource and search possibility?

I see the point, that replacing (not just adding new) communication channels is a difficult topic, as the existing community grew up with the ones we have and share. But IMHO we need to consider, that we might want interest new / younger folks, which is currently confronted with a longer process (register, edit, learn tags and examples, look for community channels, …) and esp. to find a easy way to ask the community is IMHO limited by technical burdens (find & signup MLs / forum, bad search possibilities, bad rich content abilities …) while there are modern tools, which might simplify this process? Simple for non tech people which grew up with (centralized media as) whatsapp, youtube, … ?

I very much agree with Tordanik and would look forward to finding a software that could replace the forum, MLs and help system.

I recently wanted to link to help.openstreetmap.org from StreetComplete as one of the places to go to as a beginner but had to find out that the website doesn’t work well on mobile and researching further, that it is not really maintained anymore. Though, there are several FOSS alternatives to the help system out there.

So even if there is not broad agreement to ditch MLs, help and this forum for discourse or another good alternative, at least the help system could be migrated to one that is still maintained.

I have very little experience with discourse, but maybe there is functionality or plugins available for it to have some features of a help system?

(up and downvotable answers if a topic is started as being a “question”-topic?)

I want to add something as well:

asynchronous communications would be Discourse and ML can be integrated there as well (at least thats on their “Features” list). For synchronous communications there would be matrix. Here is the advantage it can be made with a SSO with OSM i’m sure. The advanteage is, it is possible to use bridges and therefore telegram-users can talk with matrix-users together. Even Slack, Facebook and Whatsapp bridges are available and a Mail-Bridge as well.
So this way people can still choose what they want to use and still can communicate together. And if you can login on the osm-matrix server with your osm-login, well, why now? Theres a discussion about https://github.com/openstreetmap/operations/issues/380 matrix already open as well.

For the Discourse (Help/Answer) things you mentioned westnordost, it seems there is at least a “Mark as answer” there as shown here in this thread https://meta.discourse.org/t/sort-filter-by-likes/26193 So this would fulfill the purpose of the help-system, i guess.

It is my impression that many people can’t search properly. It’s not uncommon for people to repeat a question that was only posed a couple of days ago. Think “I made a change, but it is still not visible”, “I try to download a map, but it doesn’t work [they do not mention garmin.openstreetmap.nl]”, “I can no longer edit my map (umap)”.

Unifying won’t help for those people. And if you use a web search you can already find answers from many existing systems.

Discourse is very nice. Moreover Discourse is also mailing list and long-form chat room in addition to discussion forum.
I’m for Discourse and ready to help with migration.

Thanks, do you like to ping the Operations guys? They said they would like to help if anybody tests a server https://github.com/openstreetmap/operations/issues/377

Hi guys,

In Argentina we are moving forward with our project of migrating from Forum to Discurse and we want to ask you if you can send us a forum export to do our respective tests.

Best

What are we waiting for? FluxBB is a 17 years old forum, and has no mobile apps, nor live updates. It’s time to improve online discussion quality through improved software.

I’m not a forum admin, and therefore have nothing to do with this request, but as a volunteer in other areas of OSM I’d feel very tempted to reply by referring the person asking to Arkell vs Pressdram.

Asking in this way is very likely to put you at the back of any queue.

Oh, I am very sorry if you misinterpreted what I wrote, my intention was not to be rude, but to sincerely inquire what it takes to make this change. Unfortunately I think that everything will continue as it is for a long time to come. I find it difficult to implement the changes necessary to achieve a more dynamic community in the short term.

Is this project progressing, and did you get the data you need? While I believe it would be desirable for the OSM Foundation to host an attractive, modern forum for the entire global OSM community, and it would therefore be a bit sad to see the Argentina forum split off, what you learn during such a migration (or even just a test system or proof of concept installation) could be very valuable information for the rest of us. I would love to read about the results of your tests when it comes to importing existing content, integrating OSM logins etc.

As for the overall discussion: Frederik has a point that unification of communication channels is perhaps a bit of an utopian goal, and that there are upsides of people having their own cosy niches as well. My feeling is that a bit less fragmentation would still be desirable. But even if we just look at forum.osm.org in isolation, it could use a software upgrade. There are some relatively basic features, such as image upload or a sensible markup syntax, which we’re currently lacking.

Hi Tordanik,

I still haven’t found the person who could send me a FluxBB dump so we can run the tests. You know who I should be asking?

In Argentina we created a specific working group for the development of the community and at this point we are finishing our studies (with weekly Jitsi meetings):
Telegram > Discord & FluxBB > Discourse

We are in the voting stage of intentions. We haven’t started the tests yet.

Well…
We also consider the inconvenience of having many scattered channels (IRC, Forum, Mailing, Telegram…) and why they are scattered.
Look how curious, Argentina is the fifth most used forum, but for us it is almost nonexistent. We practically don’t use it, and we think we know why.

I understand the resistance of most of the old users (I’m 53 years old and 11 as an OSM member), but I think that ‘the old ones’ should make a greater effort to attract the new younger ones, which would be a vitamin injection for OSM

Unfortunately FluxBB is too old based and does not meet the current needs and does not attract the interest of use: Google Trends FluxBB vs Discourse: https://trends.google.com/trends/explore?date=today%205-y&q=%2Fm%2F043jw35,%2Fm%2F0tkk306

We will be happy to share the experience with you and I am glad that it is of interest to many others

Best

I agree that the upgrade will benefit the project, but not for this reason. Using the same argument we should abandon OSM, See OSM vs Google Maps.

Hi all, I’m working with the Operations Working Group to explore migrating to a Discourse setup.

How important is migrating existing posts to the new Discourse install? What if we kept them here but in an archived state and asked people to start fresh on a Discourse install?

In other words, I can get a Discourse install running fairly quickly, but migrating existing posts will take significantly longer.

No need at all. Why littering a fresh install.

Well, there are quite some people who would like the old stuff available. And some can be quite stubborn. And there are some possibilities, but i guess these are known already https://meta.discourse.org/t/import-migration-tool-from-fluxbb/51168 and https://meta.discourse.org/t/importing-from-fluxbb/51322

I would prefer set the fluxbb to readonly and move to discourse. That way the old stuff is still available. For the how-to i would suggest, setting up a test-instance, get some key-users from some forums to check it out and explain everything and when that is done, let them explain that stuff to ‘their’ subforum with all the advantages. So everybody gets a heads-up and maybe its possible everybody can check out the new instance. I’M not sure if a transition time is actually seriously needed or good, but thats worth thinking about as well.