Discussion Forums/Problem Statement
This is a "problem statement" trying to explain what problem our current Discussion Forums are attempting to solve. As the current solution is not awesome, people often get to thinking about how to make it better. However, one must understand the problem trying to be solved, and possible constraints on the solution, first. This page attempts to explain.
- 1 High Level Problem Statement
- 2 Rationale
- 3 Current Solution
- 4 Current Problems
- 5 Possible Improvement Q & A
- 5.1 Would removing the requirement for newsgroup access simplify things and give more solution options?
- 5.2 Would removing the requirement for mailing list access simplify things and give more solution options?
- 5.3 Would removing the requirement for web access simplify things and give more solution options?
- 5.4 Why can't we go news-only? One thing is much easier to maintain than three.
- 5.5 Why can't we go mail-only? One thing is much easier to maintain than three.
- 5.6 Why can't we go web-only? One thing is much easier to maintain than three.
- 5.7 Why don't we bring it all in-house?
- 5.8 Why don't we use a SaaS hosted solution?
- 6 Summary
- 7 Possible Solutions
High Level Problem Statement
Provide a public, approachable, asynchronous, multiple participant, written, archived, searchable, filterable, accessible communications mechanism for the Mozilla project.
- Public: an open project needs public communications
- Approachable: needs to be accessible to potential new community members, ideally without making them learn something new
- Asynchronous: with people all around the world, the primary communications need to be not-real-time
- Multiple participant: we are a community
- Written: written means easy to search, easy to translate
- Archived: we need to find out what people have said in the past
- Searchable: we need to be able to find what they said
- Filterable: no-one can read every project communication; there needs to be a way of splitting by topic
- Accessible: needs to be accessible to new and future disabled community members
The current solution is somewhat documented, and is built around the idea of a set of unified forums, arranged topically, accessible via 3 different mechanisms - email, NNTP (news) and the web. It is read/write via all three mechanisms. It involves partnerships with Giganews (for newsgroups) and Google (for web-based Google Groups).
The hope is that anyone who wants to be involved with Mozilla will be able to cope with either email (mailing lists) or a web-based discussion interface (Google Groups).
The outsourcing is sub-optimal because Google, in particular, are not very responsive to reports of problems - although things have improved in that regard in the last month. Time will tell if this improvement is lasting.
We don't have complete control of the spam problem. A proposal exists to make that better, but requires time from IT which has not been available for some time now.
Also, the gatewaying is not exactly perfect, and threading can break in some circumstances. Patches to Mailman to make this problem better have been proposed.
Possible Improvement Q & A
Would removing the requirement for newsgroup access simplify things and give more solution options?
It's not at all clear that removing this requirement will make any difference. As long as we have mailing lists, we can convert them to news in various ways, including the gmane.org service (where anyone can apply to have it done, not just us). Also, the last time we did a survey, a significant proportion of current users used this access method.
Ditching the news requirement would allow us to just go wholly to Google Groups. We would then, at least, have everything in the hands of one organization. However, John Resig has some thoughts about why that might not be a good idea. Our own experience of the support we receive from Google Groups is poor. They are unlikely to be any more responsive to problem reports than they are now. And they are the source of our current spam problems, which also suggests spam issues would not improve (and we would have no control).
Would removing the requirement for mailing list access simplify things and give more solution options?
Everyone we want to work with has an email account and knows how to use it. The last time we did a survey, a significant proportion of current users used this access method.
Would removing the requirement for web access simplify things and give more solution options?
For an organization which makes a web browser, this would be a... surprising move. The web is also the only way to provide convenient searchable archives, so there needs to be some web presence for the forums. It might be possible to ditch the writeability of the web-based access method (i.e. just have archives, not posting), but removal of the ability to post via the web would raise the barrier to entry of discussions. Also, you may have heard this before, but the last time we did a survey, a significant proportion of current users used this access method.
Why can't we go news-only? One thing is much easier to maintain than three.
Get back to 1983, old-timer.
(Written by someone who uses newsgroups as their primary access method today :-)
Why can't we go mail-only? One thing is much easier to maintain than three.
Presumably you mean mail-only with a web-based archive? Amazingly, it seems the open source community has still not managed to come up with mailing list management software with a decent web-based archiving interface. Mailman 3 looks like a possibility, but it's not there yet. And removal of the ability to post via the web would raise the barrier to entry of discussions.
Why can't we go web-only? One thing is much easier to maintain than three.
The UI of the web-based discussion forums we can find, particularly with relation to threading and unread message markers, still hasn't caught up with email and news (some might add "in the minds of those who like to read information using a mail/news style interface"). It's very hard to consume a large quantity of information this way. Google Groups may be an exception to this... but that takes us back to one of the earlier questions.
Why don't we bring it all in-house?
That's a good idea, and would certainly give us more control; the current blocking problem is a deployable web-based read/write UI (or even, for that matter, a decent read-only UI). The mailing lists part is already in-house; bringing the newsgroups in-house wouldn't solve much of the problem.
Why don't we use a SaaS hosted solution?
This page is not supposed to be "why we can't change anything", but it does set out some of the things which show why the solution is not simple. If we could find decent open-source web-based discussion forum software which we could plug mailing lists into, we'd be in a much better place.
The following software has been noted as potentially helping with this problem:
- Gmane - mail to news/web gateway service. Web interface not awesome.
- DFeed - web interface to news server. Lightning fast and featureful.
- Sympa - mail and web, but no news support
- FUDForum - can be used as a web-based gateway to newsgroups or mailing lists
- Jive - Collaboration suite
- Discourse - Open source discussion platform from some of the folks behind StackOverflow. May not meet all our requirements, but could be a GSoC project away from doing so.
- Groupserver - mailing list manager with web interface