List Info

Thread: minutes 20-2




minutes 20-2
user name
2008-02-21 18:22:38
Hi everyone,

Hereby the minutes of the meeting we had om Wednesday the
20th of February.
It was not really a structured meeting, so the minutes
turned a bit out that
way as well.

Roel.


------------------------------------------------------------
-------------
This SF.net email is sponsored by: Microsoft
Defy all challenges. Microsoft(R) Visual Studio 2008.
http://clk.atdmt.com/MRT/go/vse0120000070mrt/direct/01/
_______________________________________________
LogiLogi-list mailing list
LogiLogi-listlists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/logilogi-
list

  
Re: minutes 20-2
user name
2008-02-22 06:58:59
Thanks for the minutes. I have some small corrections. They
don't need
to be changed in the pdf, but if they're here for the record
it's
fine.

One confirmed some conference is the Digital Humanities 2008
Conference:
http://www.ekl.oulu.fi
/dh2008/

The bug-hunting will be organized like a small competition.

IE6,7 and FF 2 (& 3) covers not 70%, but about 90% of
the browser
usage (
http://www.w3schools.com/browsers/browsers_stats.asp,
http://www.w
3counter.com/globalstats.php).

For the layout especially coloring has to be done; Bruno
might like to
do this, or someone of the Nijmegen team...

About Feng's question; there is a central db in Rails, but
it is
defined in migrations (db/migrate/), and can be seen in it's
entire
definition in the db/schema.rb file (should not be changed
there as
it's overwritten on rake db:migrate...)

Sorry; my mistake; the diff (& only the LCS algorithm
used in it)
ended up being written in C, not C++... (can be found in
doc/lcs.c
now...)

The voting-power formula follows in my next mail,

(rest of the minutes are cool),

Wybo

---

::Student:
- Informatiekunde (computer linguistics, IR, webtech),
History and Philosophy
- Member of the Center for Metahistory Groningen (http://www.rug.nl/let/cmg)

::Free Software and Open Source Developer:
- Active in the LogiLogi Foundation (http://foundation.logi
logi.org)
- http://www.LogiLogi.org,
Cumulative, shared commenting, publication and 
                           idea sharing: Where insight comes
together...
- http://www.OgOg.org, RSS
feed articles rating, a meritocracy...
- ComLinToo, a computational linguistics toolset written in
Perl

::Being:
- In the world, wavy hair, go figure (http://nl.log
ilogi.org/HomE/WyboWiersma)

------------------------------------------------------------
-------------
This SF.net email is sponsored by: Microsoft
Defy all challenges. Microsoft(R) Visual Studio 2008.
http://clk.atdmt.com/MRT/go/vse0120000070mrt/direct/01/
_______________________________________________
LogiLogi-list mailing list
LogiLogi-listlists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/logilogi-
list

Re: The Voting Power Algoritm
user name
2008-02-22 07:23:40
In answer to Feng's question about the voting-power
algoritm:

Users get (more) voting power when the logi's they wrote are
rated well.

- All anonymous users start off with 0.1. 
- Users with an account have 1.0.
- The admin (founder of a peergroup) has 3.0.

This info can be found in config/config.rb (is loaded into
GlobalCongfig's db when the db is initialized...)

Anonymous users only have voting-power in the general
peergroup... For
all other peergroups one has to be a member before one can
vote in
them at all.

One becomes a member of a peergroup when someone already a
member of
it votes positive (> 0) for a logi you wrote (real-time,
see
app/models/rating.rb, def vote). 

In that case one gets 1.0 voting-power there plus what the
well-
rated logi gives you.

Which is it's score ** 2 * 0.05 at this moment:

For example (range 0..8):

---(Ruby typed in irb)
[0, 1, 2, 3, 4, 5, 6, 7, 8].collect {|i| i ** 2 * 0.05}

[0.0, 0.05, 0.2, 0.45, 0.8, 1.25, 1.8, 2.45, 3.2]
--

Voting-power determines the weight of your vote when you
rate a logi
of somebody else, just like how this works with
high-school-grades in
the Netherlands... 

For example a 9 with weight 2 and a 7 with weight 1, gives
the weighted 
average of: (9.0 * 2 + 7.0) / 3 = 8 1/3.

In addition; in order to prevent some Logi's to become
impossible to
re-judge, the weight behind a score is reduced over time.
This is done
via a half-life. Currently it is set to one week (7 days)...
(also in
config/config.rb). The reduction is done in daily amounts.

This all can still be tweaked and is open for discussion...

Hope my description is clear enough...

Wybo

------------------------------------------------------------
-------------
This SF.net email is sponsored by: Microsoft
Defy all challenges. Microsoft(R) Visual Studio 2008.
http://clk.atdmt.com/MRT/go/vse0120000070mrt/direct/01/
_______________________________________________
LogiLogi-list mailing list
LogiLogi-listlists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/logilogi-
list

[1-3]

about | contact  Other archives ( Real Estate discussion Medical topics )