Mapping SVN usernames to git

[Dirk Stöcker]

Hello,

> The important difference is that if you want to involve many people, 
> you will have more success if you do "what most people do". Even if 
> there were solid reasons why Trac+SVN was technologically superior, or 
> if the 2.5 people working on the source today prefer Trac+SVN, it would 
> STILL make sense to switch to GitHub.

See last mail. I don't want to involve many people. Like I always did I 
want to involve these few that count. And until now that works. Could be 
better, could be worse. Most other projects have an active contributor 
count of 0. JOSM is always in the range of 2-3.

> And Dirk, I can't help but agree with the others that every single time 
> you say "I have enough work already", you are adding momentum to the 
> "switch to GitHub" train. It's not going to make your load lighter by 
> itself, but you *will* get more helping hands in the long run.

Ah, maybe that's a misunderstanding. I don't want to have do maintainer 
work only because to support something fancy which I need to care 
afterwards. And whatever new comes I, I must support it afterwards as 
even if people are involved in setup they will abandon it sooner or 
later (as they are interesting in the setup task, not in maintaining it 
- that are never the same people).

What's there ATM needs nearly no work at all (except when Taylor insists 
that tools should be updated :-). Everything is automated. There are 
weeks where I don't care about JOSM operations at all (Taylor can 
probably tell, these are the weeks when he needs to set me into CC 
explicitly).

> The thing is, the standard volunteer contributor doesn't FIRST decide 
> they want to become a JOSM contributor and then go "ok, let me see what 
> I have to learn to do that". Sure, it would be great if you had people 
> committed like that. But in reality, people will start by making a tiny 
> hit-and-run pull request to scratch an itch. Because they're on GitHub, 
> JOSM is on GitHub, and they don't have to look up how to contact the 
> maintainer and what patch format the maintainer wants - they just fix 
> their thing, make a pull request, and that's it. Some of these will be 
> rubbish and the maintainer will reject them, but some of them will also 
> be good and you'll wave them through. And then some people will make 
> their second and third pull request. And find that it is rewarding to 
> work on the editor used by tens of thousands of OSM contributors. And 
> do more work.

They already can do that and do it. While I insist they open a Trac 
ticket Taylor is more flexible and directly handles the GitHub requests. 
As he does the work it's his decision.

> Yes, SVN is good enough for some things. As I said, I still use it at 
> work (where I don't have to attract volunteers to use it). But SVN is 
> not good enough anymore for a modern Open Source project, and I want 
> JOSM to be a modern Open Source project.

I don't want JOSM to be a modern Open Source project. I want it to be a 
successful project. And that it is.

> If you think that switching to GitHub is too much work for you, let's 
> make a list of what is required, and find volunteers who commit to 
> doing that.

When the discussion comes up about "Switching to GitHub" this means 
everything else must be abandoned. Trac ticket system, Trac wiki and all 
the services relying on that. That would involve to rebuild the whole 
API behind, rebuild the wiki in different form and so on. Several years 
work I'd say.

If we're talking about making it easier for some developers by sending 
patches with GitHub infrastructure. Well, they can do and do that.

Now the problem is still: All these many people e.g. Marcello assumes 
will come and help are nowhere to be seen. They can already help now.

For Freedom In Peace
-- 
http://www.dstoecker.eu/ (PGP key available)