[tryton-debian] Future of the Tryton suite in Debian

Wed Apr 11 10:46:51 BST 2018

* Mathias Behrle  [2018-04-09 18:42 +0200]: 
>* Raphael Hertzog: " Re: Future of the Tryton suite in Debian" (Mon, 9 Apr 2018
>  16:27:04 +0200):
>
>Hello Raphael,

Hello Mathias, 

I came here wondering if people had stepped up to propose themselves
to maintain the packages.

One of my goal was also to get to know how I could maybe help in this
project (do some NMU, I don't know).

I realize now that it might not be a good idea as it could be seen as
yet another way in which B2CK is controlling Tryton.

>> My own experience is limited and I have the feeling
>> that it's largely operated by C?dric Krier acting like a benevolent
>> dictator for life.
>
>Indeed the project is operated by B2CK, a two person company. One of
>them being the leader in SABDFL style (Cédric Krier), the other being
>the head of the Tryton foundation (Nicolas Evrard).

I am indeed the head of the foundation, much against my will I have to
say.

Nobody wanted to be the president and as the foundation is a Belgian
entity I thought that I might as well be the president (in fact I
wanted to quit the foundation but with the same reasoning we (Cédric
and I) thought that there should be one of us to be there for
administrative related tasks).

>While the project is claimed to be a community project, it is setup
>in narrow confines to suit the needs of the de-facto editors B2CK.
>That said you will search in vain for some sort of public and
>transparent project governance. As a direct consequence the project
>loses a lot of potential contributors.

As member of the board I would be interested to understand what you
mean by 'transparent project governance'. On which level?

>Beginners are often straightaway tired after having taken (or
>eventually not) the initial ridiciously high barrier of code
>contribution just to correct a typo.

I disagree on this point: a typo require usually only a patch.

>Others (me included) just refuse to be forced to register an account
>with google to be able to upload reviews. OTOH the contribution of
>just patches is commented with the formal hint to 'How to
>contribute'.

We know for a absurdly long time that we need to get away from google.
Yet there are few progress made on this point.

I'll explain a bit why we're doing so many hint to the "How to
contribute" documentation.

We think that there should be a review process for all non trivial
patches, this is to ensure the quality of the software (which I think
is the main selling point of Tryton). So requesting people to create a
codereview is ensuring that the people submitting the patch is willing
to discuss the possible shortcomings of his contribution. If he
doesn't want to be part of this process then so be it. Its patch
might some day be the basis of another patch fixing the issue, or it
might be included as-is, but this will happen later.

>Despite the constant lack of ressources there was until now never a
>serious move to open up the project to e.g. spread the load.
>Spreading the load means spreading responsibility and trust in
>co-workers.

Sometime I wonder if we shouldn't do something like what Pieter
Hintjens said: allow anybody to commit anything and revert afterwards
the bad code. But on the other hand, we're proud of the quite high
quality of the code and it seems contradictory or difficult to keep
the same level with this process.

>But even so said core committers are waiting for the LGTM
>of the leader, other contributors have obviously to wait, until the
>leader finds time to do personally the commit.

I don't see it that way. The whole purpose of codereview is to wait
for a LGTM.

The issue of this process is that people don't review Cédric's patches
enough and thus in the end he commits them because otherwise they
would stall. Sergi is reviewing way more patches than I do and discuss
on Cédric's patches for example.

But I agree that this is sometimes a frustrating process because
your code is criticized and usually people don't like that. And some
choices are sometimes arbitrary.

>Those are just some few examples of my perception of bottlenecks in the
>project, that after all seem to be volitional to keep the project under the
>tight control of one person resp. one company.
>
>> But still it's much better than any other similar project in this scope
>> that I know of.
>
>You know Debian, don't you? ;)
>
>I know a number of projects, that don't hesitate to work and publish a
>transparent project governance. Projects that have comprehensible, documented
>and transparent guidelines and procedures to be followed to e.g. judge
>about the meritocracy of a contributor. 

How does debian judge the meritocracy of a contributor?

What we do is the following: Is someone contributing codereviews
out of the blue (meaning it's not related to its own issues). If yes
then he will become a core committer.

>> I have been able to contribute my
>> changes and I even got commit rights on a module (account_fr).
>
>Yes, I know. That was certainly due to the fact that you provided at
>that time excellent patches with exemplary commit messages (finally
>you showed to ..hmmm.. how really exhaustive commit messages can look
>like), but also due to the fact that you are a very well known and
>reputable person in Debian.

The name of Raphael rang some bells to me (as I am a debian user) but
Cédric didn't know him at all. So no, his reputation did not matter.

It's only the quality of the discussions and of the patches.

>It's great that it is that way. There could be many more.

I couldn't agree more.

>> >   Furthermore the work of maintaining Tryton in Debian is
>> >   quite low valued in Tryton.
>>
>> But it is highly valued here in Debian and by the many tryton users out
>> there.
>
>I am pleased to here that. At least for the Debian part. I can not
>follow for the *many* Tryton users out there. If that should be the
>case they miss clearly to make their voice heard on the project.

And I agree on that too.

But I wonder if it's not the fate of every software project: the vast
majority of people using it just want to get something that works for
them but don't really want to get involved.

>> >   Constitutional questions and differing opinions
>> >   are mostly considered annoying. There is a clear tendance to prefer
>> > exclusive instead of inclusive behavior (which e.g. recently led due to
>> > poor and repudiative communication to the removal of the OpenSUSE packages
>> > from the website).
>>
>> Again, can you elaborate and be more specific?
>
>Just some pointers to form your own opinion:
>
>The issue
>https://bugs.tryton.org/issue7111
>
>History of the issue
>https://bugs.tryton.org/issue5375
>
>Issue on OpenSUSE
>https://bugzilla.opensuse.org/show_bug.cgi?id=1078111
>
>And a discussion thread on tryton-dev
>https://groups.google.com/forum/#!topic/tryton-dev/4fWWO0C5hvA

I wonder what haven't been told about this issue yet.

You'll maybe be happy to know that Axel has decided to remove the
patch and that OpenSUSE should be again on the download page.

-- 
(°> Nicolas Évrard
( ) Liège
 `¯