[tei-council] fine-grained permissions for subversion

[Piotr Bański]

Trust is only one of the factors involved, and the way you formulated it
it might sound like "trust that someone isn't evil", but there is also
trust that a novice will not feel afraid to touch the source given that
they feel safe being restricted to a single area. "Revert" is easily
said but it involves the time that passes before the need to revert is
identified, including the time to actively verify that a revert is
needed, plus possible embarrassment on the part of the person who you'd
rather want to boldly go to work on their little patch of SVN land, plus
delicate issues of someone well-willing but inexpert stepping on an
expert's ground just because they can (this is the Wikipedia effect), etc.

I thought of arguing the other side, building on what Lou said: given
"cabinet responsibility", who assigns the permissions, on what
principle, can the process be "externalized", will it result in some
competition for "power" -- but in fact, I see pretty good answers to all
these questions.

Anyway, in that better taxi, we weren't exactly plotting to get this
system installed. The topic of fine-grained permissions followed rather
naturally from the topic of access permissions for a project as big and
complex as TEI at SF, and I mentioned this issue as being under discussion
at SF, and followed that up here with the link (the more so that
independently of the TEI's strategy, the very idea is pretty
good/natural/*nixy, harmless for projects that don't need it, and, last
but not least, I'd love to have it in for at least one project of mine).

So, no pressure, no rush, but why not turn the idea around in our minds
every now and then, especially in this "transition period".

  P.

On 15.04.2011 21:43, Stuart Yeates wrote:
> There is also the fact that svn and other version control systems are 
> specifically designed to make changes revertible, should someone 
> make a mistake. Not trusting them to do their job, seems silly to me.
> 
> cheers
> stuart
> 
>> Lou Burnard wrote:
>> While I sympathise with this point of view, I feel duty bound to point
>> out that there is a strong tradition of "cabinet responsibility" within
>> the TEI and particularly within the Council. Divvying responsibility for
>> particular sections or topics up would somewhat go against that tradition.
>>
>> On 12/04/11 18:39, Martin Holmes wrote:
>>> For me, the strongest argument for granular permissions would be the
>>> possibility to designate specific people as "maintainers" for particular
>>> areas of the specification in which they happen to be expert. I rather
>>> like the idea of gatekeepers (rather than having lots of people with
>>> commit privileges over the whole repo). There are some areas where I'd
>>> feel qualified to be a gatekeeper, but others where I definitely should
>>> not be; and with designated people for each area, it would be easier to
>>> know who is responsible for making a specific change once the council
>>> has agreed to it.
>>>
>>> On 11-04-12 09:40 AM, Piotr Bański wrote:
>>>> Much as you say, Sebastian, we didn't consider it critical, just
>>>> something potentially nice to have. So no special concern, I was just
>>>> following up on a conversation on SF access and the granularity of
>>>> permissions.
> _______________________________________________
> tei-council mailing list
> tei-council at lists.village.Virginia.EDU
> http://lists.village.Virginia.EDU/mailman/listinfo/tei-council
> 
> PLEASE NOTE: postings to this list are publicly archived
>