XMPP + GSoC - 2026-01-07

As mentioned, I'm trying to see if Convo can participate in a way that doesn't conflict with the NLnet project

In the meantime, Prāv is also shortlisting some features that could be turned into GSoC projects

Since I'm a heavy user of Dino, I could try to think of some ideas that could be worked on there. Not a contributor in any way (yet™) but if someone is interested in working on a GTK/Vala project I can help brainstorm ideas for approaching Dino with

> Since I'm a heavy user of Dino, I could try to think of some ideas that could be worked on there. Not a contributor in any way (yet™) but if someone is interested in working on a GTK/Vala project I can help brainstorm ideas for approaching Dino with For this you need to convince Dino developers ↺

larma:

We're generally not hyped with GSoC recently. The amount of people abusing it for the money has significantly increased and I don't want to put 3x the amount of work into reviewing that the student put into prompting some AI.

We do have some project ideas from previous years in the wiki, some of which are likely at least partially valid. If a student actively wants to contribute to Dino under GSoC, I think we'd still be interested to work with them, but the I think the project idea submission system of GSoC just doesn't fit reality anymore.

> We're generally not hyped with GSoC recently. The amount of people abusing it for the money has significantly increased and I don't want to put 3x the amount of work into reviewing that the student put into prompting some AI. I fully understand that. Though, what I have recommended: If you are motivated in general, have you tried to find a good student out of your community? ↺

> We do have some project ideas from previous years in the wiki, some of which are likely at least partially valid. If a student actively wants to contribute to Dino under GSoC, I think we'd still be interested to work with them, but the I think the project idea submission system of GSoC just doesn't fit reality anymore. What do you mean exactly? ↺

I seem to be missing messages from larma here (on Dino, ironically 🙃) but I can see them at https://logs.xmpp.org/gsoc/2026-01-07

> We do have some project ideas from previous years in the wiki, some of which are likely at least partially valid. If a student actively wants to contribute to Dino under GSoC, I think we'd still be interested to work with them, but the I think the project idea submission system of GSoC just doesn't fit reality anymore. Thanks for the info. I'll check out the list and keep it in mind if I come across anyone who wants to participate. (Most people in my circles are very anti-AI so they're unlikely to (ab)use it for GSoC)

emus: Well the regular GSoC flow is: 1) Student picks a project ideas from the list in the wiki 2) Student does a teaser task. 3) Student hands in a proposal for the project idea via GSoC submission system (referencing the teaser task they did) Now the problem with 2 is that teaser tasks are typically trivial enough that one can do them with AI, meaning that it becomes completely useless to assess a student. And 3 is essentially something AI can do fully automatically if you feed it with the project idea from the wiki. So as a project maintainer, I basically have to work completely outside the regular GSoC submission flow. Hence I would prefer to have an active conversation with a potential GSoC student and work out an idea with them on something that makes sense for both of us rather than pushing them into something using ideas in the wiki. And during that process I would also get a much better understanding of if they're fit for the job or not, than I would ever be able to with teaser tasks. ✏

I know at least one organization which follows this flow of selecting candidates through processes outside GSoC suggested flow, and then accepting them through GSoC

(Side note: I can see larma's newer messages now so it must have been a temporary issue)

> I know at least one organization which follows this flow of selecting candidates through processes outside GSoC suggested flow, and then accepting them through GSoC So they work out an idea with the student, add it to the wiki, and then let the student do the teaser task and submit it along with a proposal? ↺

I don't think adding the worked out idea to the wiki or doing teaser tasks is really mandatory. But as far as I know, Google requires that organizations have a page with project ideas.

> So they work out an idea with the student, add it to the wiki, and then let the student do the teaser task and submit it along with a proposal? Yeah, they basically have a dummy project ideas page. Many people take ideas and submit proposals – but they're all ignored. Only the previously decided candidates are selected. This maybe unfair to those others who do not know about this selection. But the state of affairs with "fair" judgement of AI proposals is even worse. Ideally Google themselves should decide a better candidate selection process. I had filled this as my feedback in the last mentor feedback form.

Okay, but isn't the idea to also prepare out of the community? At least we have that opportunity and we have been calling for this, right? So mentors could also push people they trust in the right direction. ↺

larma: but did you consider to promote in your circles?

I would much rather pay a student from my own pocket because they want to contribute to Dino but couldn't otherwise because financial ties would require them to work on something they get paid for than have Google pay a student that isn't really interested in Dino and just seeks an opportunity to earn money.

If I can have Google pay for that student that actually wants to contribute to Dino, that would be nice. We somewhat did that before (knew the student wanted to contribute to Dino before GSoC and made them use GSoC as a means to get money for it), but that was a few years back when for various reasons I would occasionally meet university students.

I understand that - my point is - there might be people out of your Dino community that are motivated and you have or can build a relation with and then apply to the program. Thats what I meant by promoting the opportunity in your community.

Well, what is it that I can promote? XSF is not accepted as a GSoC organization, no? And it won't if I don't play by their rules and create fake project ideas.

Okay, yes. I understand this, but maybe it would make sense to review if there are ways to promote at local universities if you want to do the efforts. They often have places where one can list such things. Maybe a bit too late for this year.

> Well, what is it that I can promote? XSF is not accepted as a GSoC organization, no? And it won't if I don't play by their rules and create fake project ideas. 😄 Yes and no. Yes, this is the reason how we need to apply. No, we because practially I was calling early so projects can consider to promote this in their communities and form ideas with participants and then just apply. The XSF step would just be a formality as a favor to you basically. I can help promote and find good students, but in the end this needs to come from the projects. I think I need to feedback this more clear into the organisation. ↺

> This is not a Google requirement but if you want to make your life easier and significantly increase your chances of finding the right contributor for your org you will be sure to have had some sort of contact with the person you are planning to mentor for 3+ months. https://groups.google.com/g/google-summer-of-code-mentors-list/c/3KaHXWf36_M/m/BJn8ukUeCQAJ

I think all orgs are affected by the AI issue and there are several things orgs are doing to cope. It might be good for XSF admins to go through this thread and others to get ideas and best practices.

Thanks akshay

The ideas list is helpful, it gets people engaged. GSoC is not just students, but also for people new to open source.

Hello all, I read the chat above and thought I should introduce myself. I am untrusem, currently a student, I started contributing to xmpp related project around a year ago or so, it was for prav project, I rebased the fork with conversations with 2 year worth of update, and I currently volunteer there too, I am genuinely interested in xmpp/jabber ecosystem. And I totally agree with larma's concern Sadly here, Gsoc act as a certificate kind of thing too beside the money, so student seek that for that as well. https://codeberg.org/untrusem

I was hoping to do gsoc this year, as was hoping it would be for xmpp related thing, badrihippo told me about convo participating in folks

Teaser tasks depend on the person - for a student with programming skills, a history of previous contributions is helpful.

larma: I use dino daily but I am not familiar with vala/gtk but eager to learn

Benson: I am from a non cs background actually, so actually a self taught 😅

untrusem: see if there is something you want to improve in a FOSS XMPP project. Ask the main developers to list it to help the XSF get selected.

👍 thank you Benson

Most FOSS projects prefer contributors who will be around for some time. It can take a while to get up to speed on a code base, so contributing to a project you will continue to use is nice as main developers benefit from the time investment.

true

Maybe you can also try doing stuff with Dino right away, as a run up to be ready for if you can do a proper GSoC project?

> I would much rather pay a student from my own pocket because they want to contribute to Dino but couldn't otherwise because financial ties would require them to work on something they get paid for than have Google pay a student that isn't really interested in Dino and just seeks an opportunity to earn money. Funny I came here because I wanted to make a very similar point. If someone from my community steps up and say 'hey I'm really passionate about Conversations, my entire family uses it but my grandma really wants feature xyz I need a few euros to implement that' I'll find them the money.

Money (up to a certain degree, obviously) isn't the issue. And gsoc just sets very wrong incentives

I guess gsoc started out as a nice idea two decades ago or whatever. But it has outlived it's usefulness

here it is regarded as more of a cerficate for job opportunities

The only way it kinda maybe works is if you find the student first and then go through gsoc pro forma and ignore all the garbage applications that come in

> here it is regarded as more of a cerficate for job opportunities Yes. I guess that's fine and all from the students POV. But it doesn't work from an open source software project pov

yep true, as it all comes down to as what larma said about finding people with genuine interest

> The only way it kinda maybe works is if you find the student first and then go through gsoc pro forma and ignore all the garbage applications that come in Thats the point of my mail. And if money is not the problem, I still think there could be done way more to promote those options. ↺

I think like you said before emus looking for volunteers has to be an active process. Maybe we should think of GSoC as nothing more than a source of funding at this point, i.e. not expect to find new volunteers etc. from there

I'm trytrying to reach out in my circles with people I know who could be interested. It might be too late for this round though. But I can try to keep promoting it as a constant option anyway that "hey you can help out with XMPP projects, and if you need financial support then GSoC is one option that keeps coming round"

I mean, if we as ecosystem believe it would make sense to promote onboarding of new developer through a small campaign during the year that everyone interested can participate, the XSF Communication Team would be happy to organise this.

Personally, I think even just a few reminders through the year would help. I had this on my mind after we "missed" last GSoC, and then forgot about it till the call in December when I was busy with other work. I started actively thinking about it a week or so ago and before I knew it the deadline was upon us

Maybe a half-yearly "call" in August/September so projects can slowly wake up and start talking to volunteers, who will in turn wake up when they get some free time and start getting to know the code, culminating in a solid GSoC proposal by the time Dec/Jan rolls round

I think its very important for projects to build such a opportunity regardless of GSoC. That I have no influence on unless someone asking for help.

I did not think about it this way before the recent discussion with larma and Daniel about the issues with GSoC and finding serious volunteers

okay ✏

My guess is there are quite a few people in my circles, for whom even that small financial support for a month will be enough push to put in other contributions over time. But maybe I'm mistaken so let's see how it goes this time. If it works, I can help find volunteers from here, for whichever projects want help in that area

Maybe similar initiatives can happen by other people in the community who have potential volunteers in their circles

Fwiw I made a couple of mastodon and Twitter posts in the past where I said: if you are a student and have a good idea, any idea, I will mentor you.

That way the "work" of finding people can be decoupled from those actually maintaining the projects

I think a general problem we have in the network is that onboarding as a developer is not trivial I think. And from there all the struggles continue.

Not gonna dig them up now because mastodon search is shit and my Twitter is gone. But those posts never got any results

> Fwiw I made a couple of mastodon and Twitter posts in the past where I said: if you are a student and have a good idea, any idea, I will mentor you. And I think that is a great thing more project could consider to do. Though, I respect its also an investment of at least time. ↺

> Not gonna dig them up now because mastodon search is shit and my Twitter is gone. But those posts never got any results I think you just said the message other can think of now ↺

I'm hoping to sortof onboard volunteers I find in the sense of making sure they can at least build/deploy/whatever the project and make minor edits before sending them into the hands of project maintainers as it were

> Not gonna dig them up now because mastodon search is shit and my Twitter is gone. But those posts never got any results The cynic in me points to this as an example of how few students are actually interested in coding for its own sake and not for pay 🥲️ ↺

Oh to be clear this was in the context of a potential gsoc so there would have been pay

I guess the cynic in me thinks that capable students with a genuine interest in open source already have their own stuff going on

That...actually doesn't sound so cynical

(Except maybe in the sense that us XMPP projects get left out if all the genuinely interested students are already busy with their own things)

And I think the original idea of gsoc was that you do that instead of a summer job. But at least here in Germany and I think most of Europe summer jobs aren't really a thing. Because there isn't a clearly defined summer holiday where you take a job for 2 month or whatever. The students that don't have rich parents just work part time all year. And then you'd probably end up with your part time job on top of summer of code.

Or you don't have a part time job (rich parents) but then the money part of gsoc isn't a motivation and people rather do their own thing

This is true, but still, XMPP is not limited to Germany. Anyway, I recommend to consider how we can build an environment that attracts newcomers. If GSoC helps here at some point, if not, well then we don't apply. I will feedback this to Google, maybe they change their application program at some point.

I'm not sure what else google could do here. If you want to have video interviews or whatever selection process, you already can. The org application form asks you to put a link to your page where you list that. Google won't do that selection process for you

And then you can ignore proposal submitted through the webapp that did not follow your process

(I wasn't in GSoC last year so I don't know how bad it was with the LLM generated proposals, but generally in Haiku we were never a very popular project)

It worked fine for me as a student, but I was indeed already following Haiku even if I wasn't yet actively contributing to it at the time

PulkoMandy thanks. Yes, I believe it comes along investing energy into interest for the projects outside. Because when you got this, its easy to forward to gsoc at some point

Hey, I've found this plan on the site: https://xmpp.org/extensions/inbox/xep-ai.html I'm very interested about this. Is this an important task for XMPP?

> Hey, I've found this plan on the site: https://xmpp.org/extensions/inbox/xep-ai.html > I'm very interested about this. Is this an important task for XMPP? I don't know, maybe ask in the Dev channel ↺