A proposal: split the team, launch quicker?

Quoting some things that were said and linked in Discord:

CG&CC Meta.SE: What’s the future of CGCC?

[TL;DR: Code Golf is also starting to die on SE, and they may be looking to move out.]

weegee — I really wish we could’ve gone improving art’s implementation but we have decided something else

ArtOfCode — I’m finding that I’m not able to contribute code meaningfully at the moment, because while I have some C# experience, I have no idea how a .NET Core web app is supposed to be structured.
I doubt I’m the only one - and those without C# experience even more so.
What if we split - so that those who can’t meaningfully contribute to Codidact right now help build qpixel instead, while those who can continue working on the MVP?

We decided a while back that we don’t want to just use an existing implementation straight off. However, I find us in the situation I can’t contribute to Codidact until the code is significantly further along, and others will be in the same boat. If we have folks who are in that situation but could contribute to a Rails project, what if we let them loose on QPixel (the software that’s currently running Writing) instead, and offer to host dying SE communities on that until Codidact is ready?

I’m not suggesting we develop QPixel on the same spec we’ve hammered out for Codidact - QPixel would remain more of a clone of SE than a new format. Codidact will still be our new Q&A-based community-knowledge-sharing software when we finish it, and any existing QPixel communities would be given the chance to migrate to it.

What say we?


What if Codidact is just the specification, with multiple implementations, some of which can possibly be officially encouraged? I, for instance, might want to “roll my own”, but certainly not using the current tech stack…

I would at least provide the future Codidact Identity Provider and integrate that into QPixel. At that stage I wouldn’t mind launching life boats QPixel instances for SE communities.

Switching Identity Providers is often a major user loss I think we should really avoid that.


Don’t forget about TopAnswers in this mix.

Codidact and TopAnswers are both permanent solutions; QPixel is a stopgap. Writing desperately needed something and the users who reviewed the options found some barriers to entry with TopAnswer as it existed at the time, so we went with the third option, QPixel. TopAnswers has since matured, though still a work in progress, and is hosting a few communities now. They have also made an offer to Code Golf.

The long-term picture will involve some communities being on Codidact and some on TopAnswers. We’ve agreed to ease migration between the two, should any community change its mind. I’d like us to (eventually!) build an interface that allows us to extend our instance’s network to TopAnswers and vice-versa, so your Codidact and TopAnswers communities will show up side by side in your account list. We can and should cooperate.

If TopAnswers meets Code Golf’s needs, we should encourage them to go there. If it doesn’t and they want to stage a site on QPixel while waiting for Codidact, and we can give them that, then we should.

We also need to bear in mind that communities that set up on QPixel, like Writing has, will need to go through two transitions and two learning curves. QPixel is not Codidact nor is it SE. While the data can move freely (including new activity on QPixel), the feature set and UI are going to change out from under people. If we don’t have to put users through that, if TopAnswers suffices or a community is content to wait for Codidact, then that’s better. But I very much appreciate QPixel being available as an option; it’s saved one community already!


BTW, is this a bug of QPixel? See the link

That appears to be just a user profile…? If you’ve found a bug, feel free to report it on Meta over there, or on the GitHub issue tracker.

1 Like

I think we should also grow improving qpixel because it is running a community right now.
While codidact is going on, If I can manage to contribute to qpixel, I will. The first instance of codidact is very far right now.

1 Like

What I’m reading is not “we need job creation measures”.
What I read is “there’s an issue with the codebase and a good reason to fix it, priority 1”.

I see 5 issues, and all but one (“Add a C# linter”) are on the same abstract level as the posts in this forum (which is the correct place for those, IMHO). Heyoo, we just need to resolve 5 issues and we have MVP, apparently!

As an example, look at a question search in SO, where every question only shows a small part of the text.
I would expect something like an IGistCreator interface, and an accompanying issue “Implement IGistCreator”. You don’t really need any knowledge but C# itself for that, but including writing unit tests, that would be at least an hour of work.

However, for the people who can’t write C#, I see no issue letting them work on the stopgap.


Moving forward: Great Plans Abound: or how we're going to launch a Q&A site quicker