In a possible moment of insanity, in 


I volunteered to try to help solve a problem Tom Lane noted: "The
hard part of this problem is finding a convenient way to capture
status data out of the community's conversations."  I observed
that companies who do this well actually employ people to do that
sort of thing, and that this might be a way for code morons like
yours truly to make a contribution to development.

I've been struggling since then, trying to figure out where to start. 
There are a _lot_ of discussions on -hackers, and many of them are
blind alleys.  Moreover, I can't summarise everything, I don't think,
and still make any of those summaries sufficiently detailed to allow
them to be useful.  So I have a proposal.

I was thinking of tracking 3 or 4 such discussions in the next
release cycle, as a kind of proof of concept.  I'm willing to do
that, but I'd need guidance from those who are trying to produce a
complicated feature, telling me that they need the support. 
Therefore, if someone involved in some such discussion pokes me
saying, "Follow this thread, please", I'll follow the thread in
question (as well as follow-up discussions that come of it), and
produce regular (weekly?) summaries of what I take to be the state of
the collective mind, until such time as the code supporting the
feature is checked in and agreed to.  Then, at release time, the
developers can evaluate whether the tracking produced few surprises
at the end (and, perhaps, less thrash), or whether the experiment did
not provide any benefit.  If it does, we can see whether we can make
this sort of thing scale by adding some additional volunteers to do a
similar job in future.

Does that seem worth doing?  


Andrew Sullivan  | [EMAIL PROTECTED]
"The year's penultimate month" is not in truth a good way of saying
                --H.W. Fowler

---------------------------(end of broadcast)---------------------------
TIP 9: In versions below 8.0, the planner will ignore your desire to
       choose an index scan if your joining column's datatypes do not

Reply via email to