[tei-council] F2F Duke

[Martin Holmes]

Hi Peter,

> A third issue I wonder whether anyone (except me) feels the wish to
> discuss is about Subversion/GitHub policies/best practices. That
> became virulent with the recent Sourceforge admin issue and I wonder
> whether we have (need) any guidelines about roles and functions (e.g.
> admin, developer) and how to commit (e.g. I really like the git pull
> requests because it enforces some sort of code review).

I REALLY don't like the idea of this. For one thing, there is no Linus 
in this Council. Nobody performs the role of gatekeeper on our codebase 
at the moment, and I don't think anyone ever should. All changes which 
are other than trivial typo fixes go through the ticketing system, and 
the implementer is only carrying out the wishes of Council; if they need 
help with the mechanics of it, or if they make a mistake, our normal 
review processes should catch it. Many of us are subscribed to the 
commit notification list and see every commit that goes in; we tend to 
take a quick look at anything interesting if we have a chance, so there 
is someone watching most of the time. Nobody to my knowledge has ever 
abused the commit process, and egregious errors are caught by the build 
system and fixed quite quickly.

I'm not against git and the kind of branching and merging it allows, but 
I don't think we have a need for it in P5, given that we're really in 
maintenance mode. I would support moving to git for something like P6 
development, where we would expect to try out radical new ideas with 
wide repercussions quite frequently.

To my mind, in the TEI SF repo, "admin" means just that: someone who 
does mundane management tasks like updating things, adding users, and 
that sort of thing. I don't think we should have a class of people who 
oversee the work of another class of people.

But maybe my leftie roots are showing a bit there. :-)

Cheers,
Martin


On 14-11-10 02:12 PM, Peter Stadler wrote:
>
> Am 10.11.2014 um 21:16 schrieb James Cummings
> <James.Cummings at it.ox.ac.uk>:
>
>> On 10/11/14 18:27, Lou Burnard wrote:
>>> I assume we'll be trying to progress the correspDesc proposal
>>> further.
>>>
>>> In that context -- is there a write-up anywhere of the paper
>>> presented at the MM last month? Some of us were were unable to
>>> attend the session!
>>
>> I believe it may be linked to from the draft minutes of the SIG
>> meeting:
>> http://wiki.tei-c.org/index.php/SIG:Correspondence/minutes-evanston
>
>>
> The direct link:
> https://docs.google.com/document/d/16DTp789wkrhPXH_79zexSy8wHWZId7W-RZkO4e1BAGM/edit?usp=sharing
>
>  I hope to get some new version done before our meeting. So expect
> some more ODD with requests for comments to come … That said, I also
> have my ODD validation service on my agenda and would be more than
> happy to present/discuss this during our meeting, as well.
>
> A third issue I wonder whether anyone (except me) feels the wish to
> discuss is about Subversion/GitHub policies/best practices. That
> became virulent with the recent Sourceforge admin issue and I wonder
> whether we have (need) any guidelines about roles and functions (e.g.
> admin, developer) and how to commit (e.g. I really like the git pull
> requests because it enforces some sort of code review).
>
> Best Peter
>
>
>