what it takes to unbundle, in triangle form

Pierre-Yves Chibon pingou at pingoured.fr
Thu Oct 8 12:50:59 UTC 2015


On Thu, Oct 08, 2015 at 08:44:07AM -0400, Matthew Miller wrote:
> An unbundling triangle:
> 
> 
>              inclination
> 
>                  /\
>                 /  \
>                /    \
>               / all  \
>      A       / three→ \         B
>             /   ideal  \
>            /  unbundled \
>           /    package   \
>          /                \
>         /__________________\
> 
>   availablity            expertise
> 
>                   C
> 
> 
> A: Inclination + availability, short on expertise: 
> 
>      *Ideally*, this the packager learns quickly and moves to the
>      center of the triangle, at least for this package. More
>      likely, frustration reduces motivation and the package just
>      gets dropped. Other times, a "okay, this works" job is done,
>      but there may be bugs, including potential security issues,
>      and even in better cases, the package becomes a special case,
>      harder to maintain, forever.
> 
> B: Inclination + expertise, but not available:
> 
>      Any ideas to create more time or more people are welcome, and
>      I don't mean that in a snarky way. My thinking is that we're
>      better off having the people who really care about this
>      problem work on tooling and automation which will do a better
>      job than the "get over the initial high wall" process we have
>      now, by being more thorough and by also applying _after_
>      initial packager review.
> 
> C: Availability + expertise, but no inclination:
> 
>      The problem is: we can't *make* people have this inclination.
>      Fedora just plain doesn't have that weight. I wish we did, but
>      it's *clearly* not so. The only outcome of a hard line on this
>      is less relevance for us. That's why I'm in favor of a softer
>      line, and approaches which educate and encourage rather than
>      demand.

There was a middle ground there that could have been pursued a little more: the
sandbock repo which less strict guidelines keeping the current Fedora repo with
the current policies.


Pierre


More information about the devel mailing list