Re: DIS: Re: OFF: CFJ 1608 assigned to Maud

[Michael Slone]

On 3/28/07, Ed Murphy [EMAIL PROTECTED] wrote:

Actually, this may be trivially true due to Rule 2034.


Murphy,

Since I was running out of time in which to judge, I had to submit a
version of the judgement which did not take into account your argument
regarding rule 2034.  You may wish to submit a concurring opinion.

--
C. Maud Image (Michael Slone)
It's your fault, H. Notary!
   -- Manu, in agora-discussion

Re: DIS: collated rules history

[Zefram]

Here's the current state of my historical rule text work:

http://www.fysh.org/~zefram/agora/rules_text.txt

(Note the file is 2.9 MB in size.)  This file is specifically concerned
with the text of each rule, and does not indicate mutability or titles.
Problems with the dataset are flagged by comments in square brackets.

I'm pondering what to do, in the long term, about rule amendments that
weren't counted properly.  For example, there's one that I've just come
across in R2071.  In the last FLR that that rule appeared in, just before
its repeal, its history was listed as

Created by Proposal 4574 (Wes), 14 May 2004
Amended(1) by Proposal 4607 (Goethe), 8 August 2004
Amended(2) by Proposal 4611 (OscarMeyr), 10 September 2004
Amended(3) by Proposal 4614 (Goethe), 21 September 2004
Amended(4) by Proposal 4627 (root), 4 December 2004
Amended(5) by Proposal 4666 (Quazie), 9 April 2005
Amended(6) by Proposal 4709 (Manu), 18 April 2005
Amended(7) by Proposal 4722 (Quazie), 25 April 2005

This fails to mention that it was also amended by Proposal 4597 on 4
July 2004.  The textual change of this amendment was correctly recorded
at the time, but the history annotation was never updated for it and the
revision number was never incremented.  The forms of the rule immediately
before and after that amendment were both consistently listed as 2071/0,
with the history annotation having only the creation line.

Under the old definition of amendment numbers, which was in legal
effect during the entire time that R2071 existed, the rule after
P4597 was legally 2071/1, and immediately before repeal it was 2071/8.
The amendment numbers being reported were all off by one from 2004-07-04
to 2005-06-24.

I think emending the record to show the true legal amendment number for
each version would be terribly confusing.  The off-by-one numbering
got entrenched in practice, and contradicting it seems a bad idea.
So I'm leaving the numbers 2071/1 to 2071/7 alone.  But that means that
2071/0 is ambiguous.  I now have the rule's history recorded as

Created by Proposal 4574 (Wes), 14 May 2004
Amended by Proposal 4597 (Sir Toby), 4 July 2004
Amended(1) by Proposal 4607 (Goethe), 8 August 2004
Amended(2) by Proposal 4611 (OscarMeyr), 10 September 2004
Amended(3) by Proposal 4614 (Goethe), 21 September 2004
Amended(4) by Proposal 4627 (root), 4 December 2004
Amended(5) by Proposal 4666 (Quazie), 9 April 2005
Amended(6) by Proposal 4709 (Manu), 18 April 2005
Amended(7) by Proposal 4722 (Quazie), 25 April 2005

The case of R2071 coincidentally resembles a number of other cases where
rules have a R750 automatic amendment.  (I don't have the text of R750
or before-and-after of the amendments; they're just in the history logs.)
For example, R1043's history goes thus:

Created by Proposal 474 (Alexx), Sep. 17 1993
Amended by Proposal 1043, Sep. 21 1994
Amended by Rule 750, Sep. 21 1994
Amended(1) by Proposal 1305, Nov. 4 1994
Amended(2) by Proposal 2599, May 11 1996
Amended(3) by Proposal 2697, Oct. 10 1996
Amended(4) by Proposal 3829 (Steve), Feb. 8 1999
Amended(5) by Proposal 4011 (Wes), Jun. 1 2000

The amendment by Proposal 1043 changes the rule's number from 474 to 1043.
R750 then, AFAICT, amends R1043.  P1305 then makes the second amendment
to R1043, and so on.  Legally the amendment number should be 2 after
P1305, and more generally all the amendment numbers ever reported for
R1043 are off by one.  I'm certainly not going to change all of those
revision numbers in the record.

The coincidence is that R2071's uncounted amendment happens to precede
all its counted amendments.  For the moment I'm making use of that: I'm
using the same code that handles regular pre-1069 amendments to handle
R2071's irregularity.  But it's quite likely that at some point I'll find
an uncounted amendment between Amended(5) and Amended(6) of some rule,
and then the same pattern won't apply.

I'm considering extending the revision number with a letter
in ambiguous cases.  For R2071 we'd have 2071/0a (the originally
created version), 2071/0b (post-P4597), and then back to the regular
2071/1 (post-P4607).  The term 2071/0 would then be deprecated due
to ambiguity: it is already de facto ambiguous.  R1043, meanwhile, can
have revision numbers 0a, 0b, and 0c preceding 1; these revision
numbers could be unambiguously applied to either rule numbers 474 or 1043.

-zefram

DIS: collated rules history

[Kerim Aydin]

Zefram wrote:

Here's the current state of my historical rule text work:


Nice, this goes back farther than mine which ends with FLRs
in the current archive.  One thing: I've been trying to figure
out a non-tedious way to get Repeal dates in; repeal dates
would be needed to reconstruct an FLR for any given date.
Any thoughts?

-Goethe

DIS: test

[Kerim Aydin]

This is a test to see if my mailer is wrapping text correctly or not.  No, THIS 
is a test to see if my mailer is wrapping text correctly or not.  No, this is.  
Or rather, this.

Re: DIS: collated rules history

[Zefram]

Kerim Aydin wrote:
out a non-tedious way to get Repeal dates in; repeal dates
would be needed to reconstruct an FLR for any given date.
Any thoughts?

I'm afraid that doing it properly is just going to be tedious.  Not only
do we have to examine the adopted proposals to see which one repealed
each rule, it's also necessary to check for any last-minute amendments
between the prior ruleset publication and the repealing proposal.

However, it's possible to get approximate repeal dates by comparison
of consecutive ruleset versions.  At the moment I don't do that sort
of analysis: my code takes the rulesets all mixed up (in a bucket,
with the eggs on top), and works from the individual rule versions that
they contain.  Absence of a rule is never noted, only positive evidence
is used.  (The input consists of 351 rulesets so far, plus an extra file
for short-lived rules that were never published but which I've recovered
from Michael's archive of voting reports.)

-zefram

Re: DIS: Re: OFF: CFJ 1608 assigned to Maud

[Ed Murphy]

Maud wrote:


On 3/28/07, Ed Murphy [EMAIL PROTECTED] wrote:

Actually, this may be trivially true due to Rule 2034.


If I understand the argument you intend, this would depend on Agorans
agreeing not to challenge the vote collector's announcement of
results.


Yes, that's what I had in mind.

Re: DIS: web pages

[Manuel Lanctot]

On 3/29/07, Zefram wrote:

Who controls the website at http://www.agoranomic.org?  It needs
updating.  It points at Michael's copy of the ruleset, which
is now out of date.  I'm henceforth hosting the current ruleset
at http://www.fysh.org/~zefram/agora/current_flr.txt.  The SLR is
available in the same directory.

While we're at it, it might be nice to put some actual explanatory text
on the main website.  And link to other people's Agora pages, such as
the CFJ database.


There's the seldom-used wiki, too.
http://agora.lendemaindeveille.com

~Manu

Re: DIS: collated rules history

[Zefram]

You may all be interested in this old ruleset from 1993-08:

http://groups.google.com/group/rec.games.abstract/msg/86da94d8a3ba0ba6

Wes pointed this out to me, as an initial response after I contacted
him about archives.  I'll be adding it to my database.

-zefram