Summary: the attribute formerly known as "name"

I asked for responses suitable to IRV-style voting.  Not everyone
responded to that, but it's still doable.

Of the first-choice votes, the breakdown is as follows:

role: 1
messageName: 4
messageReference: 3*
message: 1

identifier and messageType received no votes (at any preference). 
reference and messageRole each received two votes, in both cases only
second or third preferences.  We therefore drop all but role,
messageName, messageReference, and message.

Dropping role, and selecting the second preference for that vote,
changes nothing: the second choice is messageRole, and there is no third
preference.  Dropping message, and selecting the second preference for
that vote, adds another to messageReference, for a total of 4:4.

Overall, viewing the matrix, messageName seems to have attracted the
most approval (four first-choice, one second, two third), followed by
messageReference (3, 3, 0); none of the other suggestions really seem to
be in the game.

Since the results are hardly convincing, however, I would suggest that
they lead to a proposal for a quick headcount vote: those preferring
"messageName" versus those preferring "messageReference" versus those
preferring "other".

Note: messageReference is also referred to as "messageRef", and some
seem to prefer a shorter form.  All who suggested one suggested the
other, though.

Voting matrix (if columns don't line up, cut and paste into a text
editor that uses a monospaced font)

     1 x         x role
     2     2       messageRole
     x             messageType
 2 3   1 1   1 3 1 messageName
 1 1   2   1   2 2 messageReference [and messageRef]
             2 1   message
 3 2               reference
                   identifier
 A B C D E F G H I

Voters did not all use IRV-style preference, but these have been
formatted so.  The "x" entries represent a vote *against*, which wasn't
really what was asked for, but here it is.  One of the nice bits about
this style of voting is that it's visually not difficult to identify
where the compromises can be achieved (somewhere in
messageName/messageReference, in this case) (it's clearer without the
"anti" votes, actually).

Voters A: AAL, B: JS, C: SW, D: GD, E: SP, F: PY, G: TJ, H: RC, I: JJM

Amy!
-- 
Amelia A. Lewis
Architect, TIBCO/Extensibility, Inc.
alewis@tibco.com

Received on Wednesday, 10 September 2003 11:20:57 UTC