Adding a Gitea hosting service to IndieHosters

La boite à idées
1

Bonjour,

TL;DR: when an organization asks me about Gitea, I would like to direct them to IndieHoster where they can rent a Gitea instance and just use it, in the same way they currently can get a forum or cloud storage. It would be a new « app » in the IndieHoster garden

I’m working on forge federation in the https://forgefriends.org project and Aravinth is also working on the same topic in ForgeFlux. Two weeks ago we had the idea to create a Gitea hosting service, because it does not exist yet.

Instead of creating a new organization to do it ourselves, with thought it was better to propose to do that within an existing organization. For the most part because we’re technical people with time in our hands but we know close to nothing on how to create a user base. We hope to make a salary with this activity, eventually, but we’re prepared to wait until there actually is profit.

Our first instinct was to propose to Gitea but it did not trigger an enthusiastic reaction. And then to Codeberg (unfortunately in a private thread) but there was no interest either. I’m very familiar with the Chapril forge which is also based on Gitea and we agreed to jump over the language barrier (it is all in French) and give it a try but the proposal was rejected because it conflicts with the funding model of the organization running Chapril.

Even before Chapril I considered discussing the idea with Indiehosters because it has all the good qualities. But the french only website makes it difficult for Aravinth to participate. I took a closer look today and was happy to discover the forum is partially in English and that was all the incentive I needed to start this discussion.

Instead of writing a very long description of what we thought about in the past two weeks, I’d really like to have your first impression and answer any questions you may have.

What do you think?

P.S. I would have added more links for reference but I’m a new user and only allowed two.

1 « J'aime »
2

Ok, you are now a member.

Current Status

Your post resonates with me, in many directions, and I’ll first try to explain where we are, before trying to draft where we can go together, from my point of view.

Organization

People

In term of org, IH changed a lot lately. We are now around 11 people, representing something like 5 full time. So I’ll explain my point of view, but other people have a different point of view :upside_down_face:

French vs English

2 Years ago, we decided to focus on France and French, and today I really think it was a nice thing to do. But indeed, our forum is still both :slight_smile:

Website

Today, our french website is hosted on https://indiehosters.net/ but as we are mainly discussing about the service we are offering: Liiibre, it might be a bit confusing.
In one hand, the IndieHosters collective is trying to offer a unified collaborative experience between various FLOSS tools, and in the other hand, many different orgs ask for specifics.
So the collective is currently doing a bit of custom, trying to make it standard for everybody.
We currently have this tension, and I’m not sure we’ll be able to resolve it one day.

6 month cycle - Strategy

We just spent a week together to plan the next 6 months. And we didn’t discuss about git at all. And I don’t see myself going to the team to discuss the topic you just mentioned. This is a calendar mismatch :slight_smile: Sorry about that.

GitLab

On top of the calendar mismatch, we had this lengthy discussion last 6 month cycle. What do we do with GitLab. And I have the feeling we are not yet entirely happy with the decision. I’ll try to recap here the decision we took, for my colleagues also :slight_smile:
We currently host 3 GitLabs:

We understand that it is too much, too much resources, for not that much repos.
I understand the dedicated instance, IndieWeb style, to own your land, domain name and so on.
We currently do the FLOSS part of our work under the libreho.st umbrella, we think it is a better idea than under our own domain. It might be better for other orgs to come collaborate under a « neutral » area.
(But I’m not 100% percent convinced by this argument)
We couldn’t « resell » individual account under the libreho.st domain, that’s why we have our own GitLab instance.
And Libreho.st needs some git repos as well.

So until now, we didn’t have time, but I think we could get rid of the libreho.st instance.

And we also have this internal debate of GitLab vs Gitea vs sourcehut…

From my point of view, we do need git, issues, PR reviews, maybe kanban, CI, and artifact repositories (container and helm at the moment).
GitLab provides all of this in one place.)

So we currently have an internal agreement for the coming months about this current setup, but it is not frozen for the coming years.

In an ideal world

In my ideal world, and we discussed about this during the camp, IndieHosters could be fluid to accept such collaboration proposals.

We identified 2 main areas in our team. The « tech » and the « contribution ».
The « contribution » team is in « front » with organization asking our service, or asking « special » services.
The tech part is more on the backend.

What I understand and I wasn’t conscious before, is that if we welcome people in our project, this is mental charge for all the people of the current team.
What I also understand, and I didn’t before, it is that it is better to focus, than trying to solve everybody’s problems.

So, as you present your proposal, I see 3 things that are not currently aligned:

  • you want to offer Gitea, and we currently offer GitLab
  • your colleague doesn’t speak French (is he German or interested to develop this in Germany? I live in Germany, and I’d like to meet people there, to eventually start IndieHosters in Germany :wink: )
  • you are 2 new people for the team to get familiar with, and it doesn’t seem you’ll bring us closer to parity (we try hard on this topic as well)

But, yeah, in an ideal world, we dedicate a bit of resources on our side to help you get started, you do your project a bit independently, and little by little we build a new sustainable offer for you and for us.

My proposal

For the next 6 months, I propose the following:

  • if we hear people asking for Gitea, we’ll come to you to discuss
  • if you hear people that need Gitea, and maybe the rest of our offer, you come to us to discuss
  • on the technical stack, we are on kubernetes/minio, we are working on releasing a package (in the coming 6 months I think), so if we want to collaborate on the technical level, it seems to me that you’ll have to learn this stack (and we can discuss how)
  • but anyhow, I’d love to have an hour call with you in the coming months to discuss this and other topics :slight_smile:
  • and I’ll share this with the rest of the team, so that we’ll all be aware of this possibility and we’ll discuss this more seriously in the next 6 months meetings, which will be end of august

Another possibility is to discuss about this with meet.coop, I’m sure there are some people interested to do git.coop!

Cheers!

1 « J'aime »
3

@pierreok I’m very grateful that you took the time to think and write about this idea! And thanks for allowing more links, I’ll amend the original message with references for posterity.

The language barrier is a significant one and I’m glad you confirm this. It is easy to overlook how it can be an obstacle. But it is a fundamental way for people working on the same project to bond together and choosing to embrace a common language makes a lot of sense. I participated in a project last year involving about twenty people and they chose differently: only to speak English although all but one of them spoke another language fluently. It created frictions (not tensions) and made it harder for everyone to communicate. In the end I’m not convinced the benefit was significant.

The GitLab choice is something I did not see and is a sound one. Although I feel closer in spirit to Gitea and very much admire the originality of sourcehut, they currently fill different spaces in the Free Software world. I can imagine how these differences fueled internal debates in indiehosters :stuck_out_tongue: The idea that was proposed does not necessarily imply a change within the organization but it would require learning and adopting another forge, at least at some level and I can understand how this would need to be approved collectively.

The notion of growing a new team of two on a new project, a bit independently still requires work on everyone’s part. Even if it is stripped down to the bare minimum (a « mental charge ») it still is more than nothing. And together with the language barrier and the choice of Gitea versus GitLab… it’s a little much when combined. Although the proposal was turned down for a different combination of factors by Gitea, Codeberg and April, it boils down to something similar: not perfectly aligned and not in the roadmap.

It appears that we failed to convince existing organizations of the value of our proposal, four times, and that should tell us something. Your answer is by far the most heartwarming and kind, I can’t thank you enough for that :+1: I’m not sure how to analyze why we were so convinced it would be accepted. One thing comes to mind: a very significant asymmetry in the perception of the value of the proposal. Maybe a fifth proposal will convince Easter-Eggs… or not :slight_smile: In any case there won’t be a sixth.

I like your proposal to keep the discussion open and will diligently keep this topic updated as you suggest. Although it is unclear what can come out of this, I share your impression that this resonates on many levels.

To be continued!

1 « J'aime »
4

@pierreok I would love to have a discussion with you on this topic. I took the time to write a blog post to collect my thoughts. For now Aravinth and myself are working on technical stuff and set aside our ambitions to do more than that so there is no rush :slight_smile: I’m pretty much available all the time: if you pick two times slots one of them will very likely be good.

What do you think?

5

@pierreok @unteem @hougo thanks a lot of today’s discussion. The recording (in French) was uploaded.

https://bittube.video/w/6PS6cJ626HVQgV9EjYbDNC

6

Since we last discussed a lot has happened at https://hostea.org. I would love the opportunity to debrief whenever you can. Spoiler alert: zero potential customers in sight. And not for lack of looking for them :smiley:

A date pool around July 17th:

And the admin link in case other dates are a better fit:

Cheers

1 « J'aime »
7

Done.

I may have found a customer for you :wink:

2 « J'aime »
8

:eyes:

Let’s discuss too!

Tuesday July 19th, 2022 3pm UTC+2 https://meet.jit.si/TerribleSplitsCollaborateOpenly

Very much looking forward to this discussion.

9

A gentle reminder about tomorrow’s meeting. I’ll rehearse my pitch to be 5 minutes and then we can discuss informally.

10

Thanks for this meeting. The recording is here.

My main takeaways are:

  • Many suggestions to improve https://hostea.org so that it is more appealing to potential customers (the price is difficult to find, etc.)
  • The price of the instances is too low
  • It is very difficult to be sustainable with just renting instances

I said during the meeting that there is no chance to find customers for this product and therefore it does not matter if they are not profitable, as long as they do no loose money.

I then also said that I have an intuition that there might be B2B related income, contracts with governmental agencies who need support, maybe related to federation. And there might be funding for R&D when exploring federation as deployed in Hostea.

But I was not very convincing. In the end Hostea is a service that works from a technical standpoint but has no visible commercial future.

@unteem mentioned a potential customer and I’d be happy to followup with them to figure out if Hostea is a good fit for them.

IndieHosters could become a member of Hostea. I explained the revenue sharing model using this potential customer as an example. And also explained that since Hostea is not incorporated, members invoice each other every months or every three months.

Since Hostea is horizontal and not incorporated, being a member is just a matter of declaring to be a member publicly.

1 « J'aime »
11

@unteem @pierreok if and when IndieHosters decides to be a member of Hostea, please let me know and I’ll invite a representative to the forum group which is pretty much the only place that lists members.

To account for the Hostea related work (including today’s meeting), you would need to create or update an issue with the time spent, in this repository. The next income split is likely to happen in September and you’ll get your share. It is explained in more details in the revenue sharing model and the corresponding section of the governance guide. I’m happy to answer any questions you may have, it is still very fresh.

1 « J'aime »
12

@unteem @pierreok there has been new developments since we last talked. In a nutshell:

  • Hostea is going to be renamed Gna!
  • Gna! will provide both shared hosting and dedicated hosting
  • Gna! will provide both GitLab and Gitea
  • Gna! will exclusively focus on hosting and not provide services (that’s what IndieHosters does)

It now makes intuitive sense to me why IndieHosters would be a perfect member of the collective. I would very much like the opportunity to present this whenever IndieHosters makes plans for the next six months, which should happen in the next few weeks if I’m not mistaken?

1 « J'aime »
13

The renaming happened (very tedious :sweat_smile: ). Would you mind letting me know when IndieHosters members will spend a week together discussing strategy for the upcoming six months? I’d like to make sure I’m available, should you be inclined to discuss how https://gna.org and https://indiehosters.net can cooperate.

14

Hey ! It is an interesting strategy, I hope it will pan out. We haven’t yet fixed a date for our biannual planning meetup, but it’ll most likely be in November. I think we can plan for a call maybe in October ?

1 « J'aime »
15

I’m available whenever you prefer!

16

What about in the following days?

17

When do you prefer :

  • Tomorrow afternoon ?
  • Friday morning ?
  • Friday next week ?

(Tim is off this week).

1 « J'aime »
18

Friday 28th October would be perfect :sparkles: I’m available all day, you can choose whatever hour you prefer!

19

Cool ! 11h on Jitsi Meet ?

1 « J'aime »
20

C’est noté, à vendredi prochain :slight_smile:

1 « J'aime »