[cut-team] CUT discussion summary for -project@

[Lucas Nussbaum]

On 31/08/10 at 11:55 -0400, Joey Hess wrote:
> In my experience, presenting a mass of open questions to -devel is not
> the way to go if you wish to actually accomplish anything.
> 
> My plan for CUT has been to:
> 
> 0. MEME
> 1. PITCH the basic idea to enough people that it should become
>    immediately clear if many people think it is a bad idea.
> 2. HUDDLE with interested people and figure out enough details for the
>    initial stages[1] to be sure we can do something.
> 3. DESIGN a concrete[2] plan, and present to the project, and make sure
>    they don't think that is a bad idea either. Note that the plan is sure
>    to evolve both as work on it happens, and due to experience with the
>    results. Don't be too invested in the details; do try to start with
>    something that can evolve in any of a number of directions as experience is
>    gained.
> 4. GATHER enough interested people to be sure we can implement a first er,
>    cut of the plan and to commit to supporting users who use it for some
>    reasonable period[1].
> 5. IMPLEMENT
> 6. REVIEW
> 7. REPLAN
> 
> Note that the above features rough consensus, working code, and *no polls*.
> And er, that's not just my plan for CUT, I've used that plan before. :-)
> 
> At the moment, I feel that steps #0, #1 are done, #2 is basically complete,
> and so far we have one candidate concrete plan for #3 (by aj), but it
> needs to be broadened out and some things like security support are not
> filled in yet.

I don't have the impression that I fully understand your concrete plan
yet. (at least I lack the global picture)

How do you plan to address security support for snapshots? It sounds
difficult and not motivating work, but necessary if we want users to use
the snapshots.

I think that, whatever the outcome of a poll, the fact that there is
consensus on the winning proposal makes it easier to recruit people to
help with the gory details.

Lucas