[RC5] Just be patient about stats....
gindrup at okway.okstate.edu
gindrup at okway.okstate.edu
Tue Nov 18 14:55:51 EST 1997
The mail list, like the buff-in files, must be organized as a LIFO.
:) (Not really, that wouldn't work... (There was a time I didn't feel
like I had to add that...))
I have no ecomment on the patience. Many complained about how long
RC5-64 was going to take. RC5-64 (optimistically) will take as long
as did RC5-56 (and probably a day or two longer). Getting all
time-pressured now is probably foolish.
Alternatively, many wonderful things were promised when we started
RC5-64. Most of that still doesn't work.
The Bovine effort was much more satisfying before RC5-56 ended. This
was because there were very few outstanding promises and because there
was a gradual change in the "system." The sudden changes are
disruptive and disheartening. RC5-56 ended suddenly. Big surprise.
It was going to. Bovine had a couple of days to start people on the
recoding efforts. And still it took up to three weeks to convert
RC5-56 clients to RC5-64.
Bovine has promised too much on this list. Some of those promises are
being broken. Perhaps some of the promises should be rescinded and
effort on the core group's part more *visibly* directed. I *feel* as
though the group has overcommitted itself to too many improvements in
the past month. Some (oh, lord) reorganization and re-commitment to
the core process [ :-\ ] might be in order.
v3 is several weeks away. Pushing it back a bit farther might be in
order so that immediate problems can be addressed. I know that
working on v3 is probably the neat, exciting, geek thing to do,
especially more interesting than fiddlng with stats or porting the odd
CLI option, but (and I am speaking by analogy) it is important to have
solid foundations before building a house.
v3 is very much a house. It will be a big and complicated structure.
However, the foundation upon which it sits, as far as the organization
is concerned, is that of a dedicated and (moderately) happy group of
determined users who will run the clients. Allowing the confidence in
that group to erode will do more to destroy the effectiveness of the
post-v3 effort than any amount of horn-tooting about v3 before-hand.
The problems need to be categorized into time frames. Proxy and
perproxy errors need to be fixed *now*. Client errors need to be
fixed immediately after that. Feeping creaturism needs to be
addressed and made uniform after that. Stats are important for the
"week" timeframe, but the clients and proxies are on the "now"
timeframe. Remember, more people get their first (direct) impression
of the Bovine effort from the clients. That impression should be
painless and smooth. RC5-56 had this (with v2). RC5-64 hasn't
regained this because it (apparently) hasn't been a priority.
I say that this hasn't been an apparent priority for several reasons.
There was slow movement to implement highly successful clients from
RC5-56, although the implementation of completely new clients went
swiftly. Client interfaces are not uniform. Client version numbers
are not uniform. In fact, v2 clients have been poorly represented
since they are "stop-gap" on the way to v3.
v3 is (at best) several weeks away. Perhaps to the RC5-64 effort,
several weeks is irrelevant. To the *people* involved in the effort,
several weeks is much more important. "People" seem to have no skill
at taking a long view. The larger the RC5-64 effort gets, the more
this will be reflected in the effort's population. "People" want
quick results. This is why I listed the priorities as I did.
Erroneous proxies result in instantaneous trouble for multiple
clients. Erroneous clients result in instantaneous trouble for
multiple clients. Erroneous stats result in "day-shaped" trouble for
all users. v3 does nothing on this time-scale.
It appears that the Bovine crew promised more than it can deliver. A
prudent general would fall back to a place where he or she could
regroup and reapply force where it would be most efficacious. I
espouse a similar strategy.
But just as an army does not turn on a dime, neither will the
ever-growing "inner circle". This "circle" is small enough, though,
that it can do a very good approximation.
-- Eric Gindrup ! gindrup at okway.okstate.edu
______________________________ Reply Separator _________________________________
Subject: [RC5] Just be patient about stats....
Author: <rc5 at llamas.net > at SMTP
Date: 11/18/97 12:29 PM
Anyone else find it amusing to get rebuttals before the original message
gets to you?
Anyways, how about everybody just be patient a bit? The world, and this
project, could use some more patience in my opinion.
To unsubcribe, send 'unsubscribe rc5' to majordomo at llamas.net
rc5-digest subscribers replace rc5 with rc5-digest
More information about the rc5