Re: RFCs: two proposals for change

Lightning flashed, thunder crashed and Nathan Torkington <[email protected]> whispe
red:
| We have two suggestions on the table:
|  1) Rename RFCs to PCRs. (Perl Change Requests).
|  2) Add a Status: metainfo field.

I still think (as I stated in msg 805) that we need three levels.
An RFC, a proposal, and an implementation plan.  I think the particular
names we use are pretty much irrelevent; no matter what we use someone will
get confused unless they are part of the process.

"Request for Comments" makes sense to me.  It tells me that someone wants
(my) input about something.  "Perl Change Request" sounds to me like either
an initial idea, or a final standard: PerlNewbie says "I think it'd be kewl
if perl could do X" or Larry says "Feature X needs implemented".

How about: RFC stays what it is.  It is the initial request for comment on
an idea.  It can have whatever level of detail we've already decided on.
After the RFC phase, the working group returns a PDS, a proposed design
standard.  It would list the main discussion points, what implementation is
recommended and why.  Then, "The Powers That Be" (Larry, Gnat, some
committee, whatever) returns an FDS, final design standard, which is what
is implemented.

|  1) I'm loathe to make unnecessary changes.  RFCs are requests for
|     comment.  I understand the confusion between the Internet RFCs
|     which are nominally standards, but these are Perl RFCs not
|     Internet RFCs.  Is this really such a big deal that we need to
|     invalidate everything we've already written that mentions "RFC"?

No, but we might want/need to make more levels.  The problem I see is that
most people seem to think that an RFC lives from beginning to end.  I think
that at certain milestones, it makes more sense to "freeze" things and move
to a new level.

|  2) Who is going to decide the status?  I picture a working group
|     chair saying "ok, time to nail this RFC down" and handing it
|     off to a working group.  When they return, is that the end of it?
|     I don't want, and I guess nobody wants, indefinite ongoing
|     blather about a topic.

That's why there are three phases.  The initial RFC, the proposed design,
and the final solution.

-spp