I hear your arguments, Vaclav, and understand that new generations have other approaches to communication on the net, and use other tools. I have been confronted with students to whom I suggested to ask their questions on a mailing list only to realize a bit later that they had no idea what a mailing list was, but were a bit afraid to ask. SoI agree that if relevant we need to adapt.
I personally am in the same logic as Zoltan: I like way how mailing lists allow to passively receive the info. But I understand the barrier that signing up to a mailing list can represent.
So, the issue here is not so much about specific tools (why not push Twitter, Instagram, etc as tools - I know people who search twitter whenever they have question). My issue is about dispersal of communication. On my side it is probably more related to communication between devs where I witness an increased difficulty of following important discussions because they happen in github PRs and I find it difficult to sort through all github mails and identify the important ones.
Your concern is more with new users and possibly new devs, but probably those coming from a bit further away. So these are probably different questions and should be discussed separately, but the issue of dispersal of forces needs to be discussed even in the context of user support.
I would be curious to get an idea of the numbers and proportions linked to the problem you describe: How much potential interest do we really loose because of the absence of a forum ? Of those, how many actually have a github account ? I feel the discussion to be a bit in the dark.
For the issue you raise, given the discussion since, the ideal would probably a solution in the form of a forum that allows you to also receive and contribute message by sending mails. This would satisfy both worlds.
For the issue I raised, it is probably more a discussion of how to identify and focus important discussions into one channel, and less about the tools.
Moritz
Am 21. Januar 2021 04:31:43 MEZ schrieb Vaclav Petras wenzeslaus@gmail.com:
Let me finally write some arguments for GitHub Discussions.
First of all, I think it is a tradeoff, so I agree that the issues here are valid, at least to a point. My question now is if it is worth enabling GitHub Discussions anyway.
As I mentioned earlier, people are asking for a web-based solution (see e.g. post from November on grass-user [1]). I think emails (and mailing lists) are awesome, but mailing lists are increasingly seen as archaic and not accessible. Nabble does not seem to cut it and it was even demoted on the mailing list for its link instability (which I think is a concern). It seems that if the Nabble situation would be fixable, it would be fixed already. Signup to receive all emails for a specific mailing list before posting a question is a big commitment, especially when people are using multiple software packages or are just trying out GRASS GIS. Is it clear to everybody they need to sign up before posting anyway? When you are already committed to GRASS GIS, they might not show stoppers, but when you are not, they certainly can be. Conclusion: If we want even the uncommitted users to ask questions, we need something which feels light, you already have an account there, and it does not require you to manage email filtering.
There are already web-based forums, namely GIS StackExchange and StackOverflow proper where GRASS-related questions are being asked. This demonstrates the interest in the web-based Q&A platform, however when you look at the posts there, you see that it does not work that great. First, many of the original posts and consequently answers are actually not a good fit for that kind of platform - often a back and forth discussion is required. And perhaps more importantly, there are only a few GRASS power users answering there compared to mailing lists and comparing to how many people from the GRASS community have an account on GitHub. Conclusion: Even if we don’t direct users to a platform and support that platform, people will use it anyway resulting in harm as questions are not properly answered.
GitHub Discussions is a good web-based forum for three reasons, 1) GitHub is a platform we are already committed to, 2) devs, core+addon contributors, and bug-reporting users all have an account there, 3) a lot of potential users already have account there. The last point is especially interesting because not only that a lot of code-aware GIS users or scientists have an account there, but a lot of developers have an account there and we are very very interested in attracting developers. Developers/programmers need to combine multiple projects to create whatever they are creating. Asking them to subscribe to a mailing list in order to ask a question is exactly the reason why they will try their luck with another project. Conclusion: To attract more users, especially those who are developers, a GitHub-related service, such as GitHub Discussions, is needed and we are already on GitHub.
As I mentioned in the initial post, I don’t think enabling GitHub Discussions means closing mailing lists. I think it is important we have there is an option to ask a question, or even report a problem, without signing up for a proprietary third party service (it is bad enough we more or less require that for contributions). However, as there are people who see GitHub Terms and Conditions or a web interface as a barrier to post a question, there are people who see mailing list sign up for more inbox traffic, emailing, or attachment limitations as a barrier. The commitment needed for a mailing list, includes, perhaps surprisingly, an important group to reach that is the developers.
Best,
Vaclav
[1] https://lists.osgeo.org/pipermail/grass-user/2020-November/081842.html
On Mon, Jan 18, 2021 at 8:41 AM Markus Neteler <neteler@osgeo.org> wrote:
On Sun, Jan 17, 2021 at 6:35 PM Brendan <brendan.harmon@gmail.com> wrote:
Could posts on the mailing list automatically be posted on GitHub Discussions and vice versa? That’s how the GRASS Nabble forums work right? Those look great.
If that would work, perhaps yes.
Otherwise I’m with Moritz and the others who see the risk of fragmentation etc.
Markus
grass-dev mailing list
grass-dev@lists.osgeo.org
https://lists.osgeo.org/mailman/listinfo/grass-dev
