Re: DIS: [cotc] state of the cfj archives

[Gaelan Steele via agora-discussion]

> On Jan 24, 2020, at 2:01 PM, Kerim Aydin via agora-discussion 
>  wrote:
> 
> (with that wonderful currently-missing technology: search)

For what it’s worth, I used grep against a clone of your archive on GitHub 
(AgoraNomic/cases) while researching my last CFJ, and it worked quite well.

Gaelan

Re: DIS: [cotc] state of the cfj archives

[Edward Murphy via agora-discussion]

G. wrote:


omd inherited these from Murphy, who stored these in a mysql database


PostgreSQL.


(with that wonderful currently-missing technology: search).  The
translation from mysql to flat files left some odd quirks, in
particular with people's names.  Murphy had inherited that database
from someone else (forget who), and I don't know the history before
that.


Eris, who AFAIK originally created it, along with web view and CHUI
maintenance scripts (I eventually translated the latter to the web
as well). I also had some flat files for really old CFJs and was
backfilling them into the database, but didn't get that far with it.

Re: DIS: [cotc] state of the cfj archives

[James Cook via agora-discussion]

On Fri, 24 Jan 2020 at 22:04, Kerim Aydin via agora-discussion
 wrote:
> Over time I've managed to backfill in about half the cases in the
> 3400s, but took time off in the 3500-3600s, so there's still gaps
> there.  Getting 10-20 old cases up per month or so, generally working
> backwards.

When I first read about Agora, many years ago, I remember lazily
poking around the CFJ archive a bit to get a sense of what it was all
about. And of course as a player now it's very useful. Thanks for
working on it.

- Falsifian

DIS: [cotc] state of the cfj archives

[Kerim Aydin via agora-discussion]

Recently it's come up about which cases are missing from the CFJ
archives.  Since missing numbers are hidden on the main CFJ index
page, I thought I'd share this:

https://faculty.washington.edu/kerim/nomic/archive_status2020.htm

This is a static snapshot of the tool I use to look for gaps (the live
version runs too long to share).   Bolded entries are any entry where
the last History entry does not start "Judged".  That may mean
incomplete (in recent cases anyway) but not always.  Missing case
numbers are shown.

Some history:  I inherited the cases around 2015 as a set of flat
files from omd, who was cotc for a while but had to stop updating.  It
was a quiet time (note the 5 month gap between cfjs 3443 and 3444!)
omd's updates stopped in 2013 around 3385 (the indicator is the
disappearance of 'GMT' from the time stamps).

omd inherited these from Murphy, who stored these in a mysql database
(with that wonderful currently-missing technology: search).  The
translation from mysql to flat files left some odd quirks, in
particular with people's names.  Murphy had inherited that database
from someone else (forget who), and I don't know the history before
that.

Over time I've managed to backfill in about half the cases in the
3400s, but took time off in the 3500-3600s, so there's still gaps
there.  Getting 10-20 old cases up per month or so, generally working
backwards.

My favorite part is the format consistency - it takes a fairly simple
auto-parser to correctly read case history going back to 2000, and
almost correctly (will need just a tweak or two) back to 1995 (though
with huge missing case gaps back then).

-G.