• Members 4 posts
    April 6, 2019, 12:32 p.m.

    I just found that we have a new forum on discuss.python.org. It’s a shame that the forum is not powered by a python built tool like misago although discourse is a pretty good choice.

    Having said I may start an experimental forum at my new school (probably on intranet only to avoid hassles) and I’ll definitely use Misago.

  • April 8, 2019, 9:41 a.m.

    Can't say I am surprised by the software choice (even if I am surprised by PSF deciding to start their own forum finally).

    Discourse has big names and big money backing it, as well as some awesome brands already adopting it. Those make it a very popular first choice for new communities by big orgs.

    There's no way Misago could compete with them.

  • Members 29 posts
    April 8, 2019, 2:41 p.m.

    Discourse might be the leading choice but doesn't compare with misago at all. Discourse is slow and bloated check out page load/speed tests compared to misago. I considered both and went with misago. I can understand why they went with discourse though.

  • Members 5 posts
    April 10, 2019, 2:37 a.m.

    It's never to late. Misago is fascinating as it is; as long as it keeps that up, maintainers from discuss.python.org may change their mind.

    And I agree that Discourse is somewhat bloated.

  • July 21, 2019, 2:08 a.m.

    Misago has plenty of features but also plenty of rough edges that turn people off after knowing it better. :(

    I feel that 0.20 sets us on right path with theme system and customization options in admin panel, but there is still bunch of things that need improving or adding in next versions, of which two biggest issues I wish to tackle on next are:

    • UI is currently a weird mixture of static HTML and React.js components repeating this HTML but adding interaction. We should go the way of Facebook, and only use HTML for navbar and footer, and always use React.js between those. This will save me the need to write HTML twice, and save users the confusion of having their template customizations disappear, as well as remove the "flash of interactive content".
    • There's no plugin system of any sort in Misago. Instead Misago relies on pipelines approach where user can configure how parts of Misago should work by adding paths to their python functions or classes in settings.py. And frontend gives no option for plugins, which is super annoying if you want to inject some content to default pages (like advertisement or announcements). This causes massive issues for people using Misago, so we should move to something else, likely replacing current pipelines with proper plugin system. Now that misago-docker is recommended way for installing Misago, we could also use Discourse's approach for plugin setup, and give users option to install plugins by creating custom plugins.txt file with names of plugins they wish to have installed (but also support private plugins).

    In long run I would like to reduce amount of features in Misago, so there's solid core forum package that I can sanely maintain myself without large time investment, and plugin system leaving path open for others to fill in the missing parts, allowing more customization and taking some burden off me.

    I also know I've promised GraphQL API some time ago, and its still on the list, but it may have to wait to after UI refresh/plugins.

  • Members 1 post
    July 22, 2019, 11:58 a.m.

    As far as I'm aware there is extensive email integration baked into discourse. I really like misago but there's really no comparison when it comes to features.
    @rafalp how do you feel about mailman/hyperkitty ? that's one group of software that could make it more competitive with discourse and still keep the numer of things that need to be developed in-house to the minimum.
    I'm not that well versed in django to suggest anything only that mailman suite could also be used with docker.

  • July 22, 2019, 2:38 p.m.

    Thanks for taking moment to share, @ojn

    I really like misago but there's really no comparison when it comes to features.

    And there will never be, to be honest. Misago is mostly me doing things in my free time, and Discourse is a team of dedicated of guys working 8 hours a day five days a week improving existing features or adding new ones.

    There is also an unsolvable problem of different people coming with different expectations. For one person good email integration is a must-have, other says that category subscriptions are must-have, even other says that Facebook-like reactions are a must-have.

    Personally, I'm feeling that I've kind of engineered myself into a corner with the current iteration of Misago, which does many things, but has long-standing deficiencies because there always was one more feature that was a must-have to somebody. In future major version I would like to actually cut features and focus on well polished forum package with plugin system - so if somebody is missing something, there will always be a space for plugin that somebody else will hopefully develop and maintain, taking the burden off me, and letting me focus on iterating and improving main "product".