IRC log of openacs on 2002-02-14

Timestamps are in UTC.

00:36:45 [talli]
http://cspar181.uah.edu/RbS/JOB/prog00.html
00:36:46 [chump]
A: http://cspar181.uah.edu/RbS/JOB/prog00.html from talli
00:36:59 [talli]
A: Some Important Theological Questions Would Be Answered If God Were A Computer Programmer
00:36:59 [chump]
commented item A
00:37:21 [talli]
A: Where will I go after I die? Onto a DAT tape.
00:37:21 [chump]
commented item A
01:51:18 [markd2]
markd2 (~Snak@r-41.69.alltel.net) has joined #openacs
01:56:06 [davb]
davb (~chatzilla@alb-24-58-161-172.nycap.rr.com) has joined #openacs
02:08:38 [k2pts]
k2pts (~nkd@adsl-168-174.cytanet.com.cy) has joined #openacs
02:10:08 [vinod]
vinod (~vinod@216-164-248-20.s2496.apx1.sbo.ma.dialup.rcn.com) has joined #openacs
02:10:21 [CSari]
CSari (~CSari@pm202.isomedia.com) has joined #openacs
02:10:34 [CSari]
*waves hello*
02:10:48 [vinod]
hi CSari
02:11:37 [CSari]
today is plunge day.. I'm downloading oacs4...
02:11:44 [vinod]
woohoo!
02:12:10 [CSari]
at least to look at. if it scares me too badly i might stick with 3.2.5 ...
02:12:17 [vinod]
the beta install docs are up at http://kurup.com/acs/openacs4
02:12:34 [CSari]
excellent, thanks :)
02:12:50 [davb]
hi vinod
02:12:52 [vinod]
np - let me know if any of the docs are confusing
02:12:57 [vinod]
hey davb
02:13:10 [davb]
4 is the way to go.
02:13:31 [davb]
You can mostly ignore the parts you don't understand yet, that's what I do :)
02:13:52 [vinod]
i think davb knows more of openacs4 than anybody, though!
02:13:58 [davb]
heh
02:14:02 [davb]
that is _very_ scary
02:14:08 [CSari]
*smiles* hopefully i'll understand the parts I need anyway... I've gotten kinda attached to 3x .. at least I've figured out the parts I needed to change.
02:14:17 [davb]
I suspect DonB and company know a little more
02:14:34 [davb]
4 is worth it just for the template system.
02:14:43 [vinod]
agreed
02:14:49 [vinod]
templating is very cool
02:15:01 [k2pts]
davb: I finished the datamodel. I wanna post them to the forums tomorrow for discussion (4 o' clock here)....sending by email (that's dave at designexperience.org)
02:15:26 [davb]
thedesignexperience.org
02:15:30 [k2pts]
ok
02:15:36 [davb]
cool. I'll look at tomorrow
02:15:37 [k2pts]
I'll include some comments
02:15:53 [CSari]
i didn't do anything with templating with 3x. My one "real" 3x site I did with hacked ad_header and ad_footer. Stupid, but true.
02:15:55 [k2pts]
I have to fix something...spent the night answering emails...
02:16:06 [k2pts]
so have that in mind...
02:16:18 [davb]
np.
02:16:27 [markd2]
yeah. I'd use oacs4 just for the templating system
02:16:41 [vinod]
CSari: yeah, the templating is much more powerful in 4.x -actually makes it worth using. i used to be a ad-header/footer junkie ;-)
02:17:00 [k2pts]
talli: what's luke's email (luke at museatech.net)?
02:17:22 [vinod]
plus the db-abstraction api makes things a lot easier (although that was also there in acs 3.4.x)
02:17:24 [davb]
k2pts: that works
02:17:35 [davb]
(lukes email)
02:17:36 [k2pts]
ok
02:17:38 [k2pts]
thx
02:18:06 [davb]
I need to actually write some code instead of just porting stuff so I can learn more.
02:19:48 [CSari]
hmm... anyone done Jonathan's rpm install? call me lazy, but if i can do it without goofing up my 3x install, i'd rather..
02:20:20 [vinod]
i haven't, but he's really responsive if things don't work
02:20:39 [davb]
I used the first one and it worked well, but most of use use debian :)
02:21:57 [CSari]
i've got a mandrake system, and the 3.2.5 rpm worked like a dream. i'm just looking at his oaks4 directions and getting worried since it discusses a fresh install
02:22:08 [CSari]
...like I've got a spare test box kicking around. *sigh*...
02:23:40 [docwolf]
docwolf has quit ("What happens if i press this bu")
02:24:16 [davb]
vinod: is there a doc somewhere about running 3.x and 4?
02:24:45 [davb]
I think as long as you use a different tcl file and different port number, you are ok.
02:24:53 [vinod]
kinda - my brief openacs install guide talks about it a bit
02:25:07 [davb]
different aolserver config file that is.
02:25:10 [CSari]
i've already got virtual hosting for 2 sites going, so i figured it shouldn't be bad.
02:25:39 [davb]
what version of aolserver are you using?
02:25:54 [vinod]
http://openacs.org/new-file-storage/one-file?file_id=114
02:26:02 [CSari]
3.3.1+ad13
02:26:15 [davb]
ok. all you need extra is nsxml
02:26:28 [markd2]
well, and make that check payable to "Vinod Kurup"....
02:26:38 [vinod]
:-) :-) :-)
02:26:39 [CSari]
hehe
02:26:57 [CSari]
excellent, thank you for the link
02:27:18 [CSari]
you forgot the paypal button, of course.
02:27:28 [vinod]
unfortunately, markd2 says i have to set it all aside for taxes, c/o badgertronics.com
02:29:16 [talilee]
talilee (~talli@xcdfddb76.ip.ggn.net) has joined #openacs
02:29:35 [talilee]
well lookee here. is the circus in town?
02:29:39 [talilee]
everyone's here tonight
02:30:03 [vinod]
i'd just like everyone to notice how civilized things were up until a few moments ago
02:30:30 [talilee]
* talilee picks up his club and opens up his can of Savagery Special
02:30:38 [vinod]
* vinod ducks
02:30:41 [davb]
vinod: we all know you were saving up until talli got here...
02:31:07 [vinod]
my repertation presedes me
02:33:10 [markd2]
anyone know how to find the technical contact for an arbitrary domain?
02:33:34 [vinod]
whois?
02:34:11 [vinod]
whois badgertronics.com -> DM Hostmaster (DH3299-ORG) hostmaster@DIGITAL-MISSION.COM
02:35:54 [davb]
http://www.mnot.net/cgi_buffer/
02:35:54 [chump]
B: http://www.mnot.net/cgi_buffer/ from davb
02:36:01 [davb]
B:|cgi_buffer
02:36:01 [chump]
titled item B
02:36:14 [talilee]
i got a weird question
02:36:27 [talilee]
does vinod really look better in fishnets?
02:36:31 [talilee]
i got another one
02:36:44 [talilee]
anyone know steve wozniak's email addy?
02:36:48 [markd2]
vinod: thanks
02:36:58 [vinod]
can anything make vinod look better than he already does?
02:37:00 [davb]
B: cgi_buffer is a group of libraries used to improve performance of CGI
02:37:00 [chump]
commented item B
02:37:03 [vinod]
markd2: np
02:37:11 [davb]
B: scripts (and other content generation engines) in some circumstances, by
02:37:11 [chump]
commented item B
02:37:21 [talilee]
vinod: perhaps a shave of the back?
02:37:27 [davb]
B: applying performance-enhancing HTTP mechanisms that are typically not supported by them.
02:37:27 [chump]
commented item B
02:37:29 [vinod]
should be on woz.org somewhere
02:38:12 [vinod]
talilee: what were you doing looking at my back?
02:38:43 [talilee]
i was looking to test my new lawn mower
02:44:27 [davb]
http://arstechnica.com/paedia/n/net/net-1.html
02:44:27 [chump]
C: http://arstechnica.com/paedia/n/net/net-1.html from davb
02:44:39 [davb]
C:|Microsoft .Net on ArsTechnica
02:44:39 [chump]
titled item C
02:44:51 [davb]
C: a technical overview
02:44:51 [chump]
commented item C
02:45:25 [talilee]
talilee has left #openacs
02:45:53 [Starets]
Starets (~petej@home.clickvision.com) has joined #openACS
02:47:00 [markd2]
yo
02:47:04 [davb]
C: very interesting reading, the first reall information I have seen
02:47:04 [chump]
commented item C
02:47:06 [Starets]
oi.
02:48:59 [davb]
I have a question. Should keywords be different from categories for describing OpenACS content? Or is it all the same thing?
02:49:43 [vinod]
well, they sound like 2 different things to me
02:49:55 [vinod]
an item should belong to one category, but can have many keywords
02:50:14 [vinod]
but i haven't been following any of the category discussions at all (over my head)
02:50:17 [davb]
vinod, why can't a item be in more than one category?
02:50:20 [davb]
:)
02:50:36 [vinod]
i spose it could
02:50:39 [davb]
I nned to know what the difference is myself, or if there should be one.
02:51:00 [davb]
I think maybe the categories should be arranged in a tree, but keywords should just be a list.
02:51:14 [Starets]
not a tree...a network.
02:51:56 [davb]
ok right. I think that is how k2pts is doing it.
02:52:08 [davb]
now I am reading about a keyword category...
02:52:23 [Starets]
keywords arise from the content of an item, whereas categories are usually assigned, so it's possible to have a category for an item where the name of the category doesn't appear in the item "text"
02:52:38 [Starets]
imo
02:52:47 [davb]
Ok. that is a good start.
02:53:44 [davb]
I think I need a librarian :)
02:53:50 [Starets]
I think of categories as a semantic organization tool, while keywords provide unstructured search assistance
02:54:12 [davb]
yes, that is what I was thinking.
02:55:04 [vinod]
is that how cr_keywords works now?
02:55:20 [davb]
no
02:55:22 [Starets]
"fruit flies" could be about the aerodynamics of organic seed distribution media, or it could be about biological organisms; you'd have to assign the category, but the keywords just happen.
02:55:56 [davb]
well I think keywords can also be assigned to make an item more relevant for certain searches.
02:56:27 [davb]
vinod: in cr_keywords, there is a list of keywords. If a keyword has children, its called a keyword category.
02:56:32 [Starets]
I guess what I'm saying is that I wouldn't infer meaning from the keywords for an item, but I would infer meaning from the categories.
02:57:19 [davb]
ok. I am looking around and alot of library sites have search by keyword, subject, category, so they seem to think thery are different things.
02:57:36 [Starets]
Librarians are smart.
02:58:29 [CSari]
i agree keywords and categories are different.. although I'd think about subjects as being categories with lots of children...
02:59:04 [davb]
Well in a library the subject is a special part of the book metadata I think.
02:59:33 [Starets]
Starets has quit ("Fixx'm 1.5ad MacOS X")
03:00:59 [davb]
ok, I think I have it. categories might be part of a site structure where keywords would just describe the document.
03:01:52 [vinod]
sounds good to me :-)
03:02:07 [til]
til has quit (Read error: 113 (No route to host))
03:03:41 [davb]
Ok, thanks for listening! I'll be back tomorrow :)
03:03:49 [k2pts]
davb: I have emailed the datamodel
03:03:57 [k2pts]
and some ideas/thoughts
03:04:20 [k2pts]
I'll post to the forums tomorrow evening (I hope) a request for comments/suggestions
03:05:02 [davb]
ok.
03:05:14 [davb]
I'll look at it tomorrow.
03:05:53 [k2pts]
thx, later
03:05:56 [k2pts]
k2pts has left #openacs
03:07:12 [davb]
cool, it looks good.
03:07:17 [davb]
well goodnight
03:07:18 [davb]
davb has quit ("ChatZilla 0.8.5 [Mozilla rv:0.9.8/20020204]")
03:10:58 [talli]
uhm. was there actual OACS4 discussion going on tonight?
03:11:13 [talli]
like, people were talking about the merits of one approach over another?
03:11:43 [vinod]
no. davb was mostly just talkin to himself
03:11:53 [vinod]
i added reassuring grunts here and there
03:12:00 [talli]
man, can't you prescribe something for that vinod?
03:12:19 [talli]
for davb, and for your grunting problem
03:12:23 [vinod]
nah, i don't believe in modern medicine
03:12:24 [talli]
mostly for your grunting problem
03:12:42 [vinod]
it just pays the bills
03:12:46 [talli]
oh yeah? how do the MIT students feel about your blood letting and leeches procedures?
03:13:15 [vinod]
as long as it gets them back to doing their problem sets, they're happy
03:13:18 [CSari]
hmm... i got told how wonderful the new templating system is.. that was oacs4 related.
03:13:28 [vinod]
oh yeah!
03:13:29 [CSari]
it didn't last long, of course.
03:13:37 [vinod]
that's 2 on-topic discussions in 1 night
03:13:44 [talli]
dangerous. dangerous.
03:13:55 [vinod]
both when talli wasn't around
03:14:04 [CSari]
don't worry.. vinod will prolly have his pants off in a minute or two, since talli is here.
03:14:13 [talli]
* talli curses the moment CSari came into the channel. a woman will always get men to focus!
03:14:18 [vinod]
did CSari just ask me to take my pants off?
03:14:20 [talli]
whoa.
03:14:23 [talli]
she wins!
03:14:23 [CSari]
no.
03:14:39 [talli]
that was a serious touche CSari
03:14:48 [talli]
we may even need a ruling by markd2 on that one
03:15:09 [CSari]
lol
03:19:33 [talli]
vinod, you know i keep forgetting to mention to you that the docs you emailed to me were the original docbook files from aD, not the ones you were working on
03:19:35 [markd2]
* markd2 wakes up
03:19:44 [markd2]
CSari is toucheing vinod?
03:19:57 [markd2]
and Vinod's webcam isn't on!
03:20:01 [talli]
* talli strokes markd2 head reassuring to get him back to sleep
03:20:41 [markd2]
* markd2 purrs
03:20:45 [markd2]
* markd2 goes back to sleep
03:20:48 [talli]
markd2, we need a ruling on CSari's retort. i think it might be pretty decent.
03:20:59 [vinod]
talli: are you sure? i used their template, but i changed a lot of the text
03:21:01 [markd2]
I give it an A
03:21:06 [vinod]
i can send you a newer copy
03:21:10 [talli]
yeah, do so
03:21:40 [talli]
ding ding. CSari, an A is good, but you'll have to try harder to get on the quotes file. ask vinod what it takes...
03:22:03 [talli]
thx vinod, for the new copy
03:22:19 [CSari]
hey, only my second day here... and i don't even have oacs4 installed yet.
03:22:26 [talli]
true true.
03:22:27 [markd2]
I think she's doing pretty well
03:22:30 [talli]
you done good, girl.
03:22:35 [markd2]
* markd2 is talking in his sleep, of course
03:22:55 [talli]
* talli wonders who is typing what markd2 is speaking
03:23:21 [vinod]
that's markd2's secret project
03:23:33 [markd2]
shhh!
03:23:38 [talli]
ooo!! secrets! i love secrets!
03:23:45 [talli]
i watch the X Files every week
03:24:12 [talli]
and i keep a journal hidden behing my bed's head board where i write down how much i love vin - d'oh!
03:24:20 [vinod]
why do you think he's been programming in those dead languages fortran and objective-c
03:24:29 [vinod]
awwwww
03:24:29 [talli]
black magic?
03:24:35 [vinod]
i love you too talli!
03:24:57 [CSari]
*shudder* fortran..
03:24:58 [markd2]
muahahahaha.
03:24:59 [talli]
* talli throws a roundhouse at vinod's head
03:25:20 [talli]
* talli follows by throwing a square cabin at vinod's groin
03:25:50 [vinod]
* vinod moves out of his closet into his new square cabin
03:26:01 [CSari]
time to take my A and go.. see you guys later!
03:26:14 [markd2]
l8r
03:26:17 [talli]
finally! i told you would be much happier once you came out!
03:26:20 [talli]
later CSari
03:26:26 [CSari]
CSari has quit ()
03:30:41 [talli]
got the docs, thanks vinod
03:31:20 [vinod]
np
03:31:48 [talli]
vinod, do you pronounce your name vih-node or vih-nod?
03:31:51 [talli]
or perhaps hal-jah-like-ah-kik?
03:32:41 [talli]
ugh
03:32:54 [vinod]
works well in medical circles
03:32:57 [talli]
that's a nice association to have with your name
03:33:05 [vinod]
lymph nodes are great!
03:33:14 [vinod]
they fight illness and disease!
03:33:33 [talli]
hmm... yes, well you should always aim high, i guess.
03:37:26 [estanchier]
estanchier (~rolf@above.badgertronics.com) has joined #openacs
03:37:47 [talli]
talli has quit (Read error: 113 (No route to host))
03:37:48 [estanchier]
Staunch Markd2
03:38:20 [estanchier]
estanchier has quit (Client Quit)
03:39:17 [talli]
talli (~talli@lti-4.dialup.access.net) has joined #openacs
03:39:29 [markd2]
staunch estanchier
03:39:50 [talli]
uh oh. is something getting staunched?
03:40:07 [markd2]
* markd2 smells a staunch in the air
03:40:41 [talli]
whoops. sorry. burrito for lunch.
03:40:58 [talli]
extra pico de stauncho
03:47:44 [markd2]
http://www.himonkey.net/holiday/ghd/index.html
03:47:44 [chump]
D: http://www.himonkey.net/holiday/ghd/index.html from markd2
03:47:56 [markd2]
D:|HiMonkey takes on Groundhog Day
03:47:56 [chump]
titled item D
03:48:05 [markd2]
D: Includes cookies!
03:48:05 [chump]
commented item D
03:58:37 [markd2]
markd2 has quit ("wheeeee")
06:32:01 [vinod]
vinod has quit ("changing universes")
07:07:02 [dlk-gone]
dlk-gone is now known as dlk
08:00:09 [andyn]
andyn has quit (Remote closed the connection)
08:00:57 [andyn]
andyn (~andy@12-254-190-230.client.attbi.com) has joined #openacs
08:01:46 [andyn]
andyn has quit (Client Quit)
08:01:47 [andyn]
andyn (~andy@12-254-190-230.client.attbi.com) has joined #openacs
09:00:53 [andyn]
andyn has quit (Remote closed the connection)
09:01:34 [andyn]
andyn (~andy@12-254-190-230.client.attbi.com) has joined #openacs
09:44:52 [dlk]
dlk has quit (Read error: 104 (Connection reset by peer))
09:45:04 [dario]
dario (dlk@walter.ita.chalmers.se) has joined #openacs
09:45:24 [dario]
dario is now known as dlk
09:45:34 [dlk]
wheeha.... nickserv is back
10:44:45 [dlk]
dlk is now known as dead-man-walkning
13:21:13 [dead-man-walkning]
dead-man-walkning is now known as dlk
13:21:38 [dlk]
dlk is now known as dead-man-walking
13:38:58 [dead-man-walking]
dead-man-walking is now known as dlk
13:59:54 [davb]
davb (~dave@rrcs-nys-24-97-22-203.biz.rr.com) has joined #openacs
14:09:31 [Starets]
Starets (~petej@home.clickvision.com) has joined #openACS
14:12:26 [markd2]
markd2 (~Snak@r-41.65.alltel.net) has joined #openacs
14:18:40 [davb]
davb has quit ()
14:24:53 [davb]
davb (~dave@rrcs-nys-24-97-22-203.biz.rr.com) has joined #openacs
14:56:03 [davb]
davb has quit ()
14:58:28 [markd2]
* markd2 wakes up and looks around at the wreckage
14:58:34 [markd2]
must have been some party last night
14:59:19 [Starets]
Starets has quit ("Fixx'm 1.5ad MacOS X")
14:59:32 [Starets]
Starets (~petej@home.clickvision.com) has joined #openACS
15:04:34 [Starets]
Starets has left #openACS
15:05:39 [Starets]
Starets (~Starets@home.clickvision.com) has joined #openacs
15:08:44 [markd2]
http://slashdot.org/articles/02/02/14/143254.shtml?tid=166
15:08:44 [chump]
E: http://slashdot.org/articles/02/02/14/143254.shtml?tid=166 from markd2
15:08:55 [markd2]
E:| CmdrTaco proposes online
15:08:55 [chump]
titled item E
15:09:20 [talli]
i liked the post, "I sent this in three days ago!!!"
15:09:29 [markd2]
LOL
15:10:17 [Starets]
I liked the poll version of the question
16:01:48 [dlk]
dlk is now known as dlk-gon
16:12:07 [markd2]
markd2 has quit ("wheeeee")
16:31:10 [Starets]
Starets has left #openacs
16:48:18 [dlk-gon]
dlk-gon is now known as dlk-phone-home
16:49:58 [Starets]
Starets (~Starets@home.clickvision.com) has joined #openacs
17:04:28 [davb]
davb (~dave@rrcs-nys-24-97-22-203.biz.rr.com) has joined #openacs
17:08:41 [davb]
http://www.paulgraham.com/taste.html
17:08:41 [chump]
F: http://www.paulgraham.com/taste.html from davb
17:08:53 [davb]
F:|Tase for Makers - Paul Graham
17:08:53 [chump]
titled item F
17:09:22 [davb]
F: "We need good taste to make good things. Instead of treating beauty as an airy abstraction, to be either blathered about or avoided depending on how one feels about airy abstractions, let's try considering it as a practical question: how do you make good stuff?
17:09:22 [chump]
commented item F
17:11:44 [k2pts]
k2pts (~nkd@adsl-168-174.cytanet.com.cy) has joined #openacs
17:12:04 [davb]
hello k2pts
17:12:08 [k2pts]
hey dave
17:12:25 [k2pts]
did you receive my email. I have just posted to the forums about categories in openacs-4
17:13:40 [davb]
yes I received it. I have to look at it more. I checked it a little last night.
17:13:44 [davb]
looks pretty good I think.
17:14:10 [k2pts]
just read my posting. there's not much more in the datamodel...
17:14:23 [k2pts]
that will save you some time...
17:14:50 [k2pts]
actually the posting is very similar to what I have send you last night. sorry, I was lazy to write something else
17:15:16 [davb]
np :)
17:27:46 [davb]
k2pts: you figured out the mising piece I didn't think of. Mapping categories to pacakges, so bboard can have its own categories, but they map back to the master list.
17:30:54 [Starets]
Does the tree_sortkey mechanism work like Celko's nested-set model for hierarchies in SQL?
17:32:22 [k2pts]
Starets: don't know (probably not -- explanation follows). do you have a link for that?
17:32:55 [Starets]
http://www.intelligententerprise.com/001020/celko1_1.shtml
17:32:55 [chump]
G: http://www.intelligententerprise.com/001020/celko1_1.shtml from Starets
17:33:11 [Starets]
G titled Trees in SQL
17:33:20 [Starets]
G: titled Trees in SQL
17:33:20 [chump]
commented item G
17:34:15 [k2pts]
my friends Oleg and Teodor built a nice tree contrib module based on gist. It seems though that our solution is better for our purposes (the contrib module has a restriction on the number of nodes, I think 64K)
17:34:25 [k2pts]
thanks for the link...I'll check it out
17:35:21 [Starets]
This model has the advantage of being very clean in terms of SQL, but I have no idea how it will perform under load on a big dataset. Some of the SQL is deceptively simple, but can modify a large number of rows in the table at one time.
17:37:40 [k2pts]
the tree_sortkey mechanism can represent hierarchies as those in the article and it's very efficient. In the article, each node has a right and left children. Using the tree_sortkey mechanism, you only need the parent of each node...The tree_sortkey can be used to collect the children of a node or it's ancestors...
17:38:40 [Starets]
this one has no limits to number of nodes or children.
17:40:06 [Starets]
G: |Trees in SQL
17:40:06 [chump]
commented item G
17:40:10 [k2pts]
yes but how can you collect nodes with just one pass. afaik (I don't know much about connect by) you cannot do it by one pass. the tree_sortkey mechanism can
17:41:17 [k2pts]
dave: yes, ain't that great?
17:41:22 [Starets]
When you say "collect nodes", do you mean something like finding a node and all children?
17:41:31 [k2pts]
yeap
17:41:40 [Starets]
G:|Trees in SQL
17:41:40 [chump]
titled item G
17:41:46 [Starets]
Aha!
17:42:02 [k2pts]
btw, I'm Neophytos
17:42:04 [Starets]
See page 2 -- it's a single SQL statement. Look at the first and second examples.
17:42:07 [Starets]
I'm Pete.
17:42:13 [k2pts]
Demetriou :)
17:43:05 [Starets]
Selecting subtrees is pretty quick with this; my fear is that there are degenerate cases in insertions and deletions.
17:43:42 [k2pts]
I didn't checkout the second page, earlier. yes, it's like the tree_sortkey mechanism
17:44:02 [k2pts]
s/it's like/it seems like
17:44:03 [talli]
hey k2pts
17:44:03 [Starets]
Pardon my laziness, but is there an URL for tree_sortkey?
17:44:05 [k2pts]
I just browsed
17:44:06 [talli]
hey Starets
17:44:06 [k2pts]
hey talli
17:44:15 [Starets]
hey talli.
17:44:56 [k2pts]
Starets: search the bboard or the 4.x central page. I learned about the tree_sortkey mechanism by reading the source code...
17:45:12 [Starets]
G: Starets struggles to learn chump
17:45:12 [chump]
commented item G
17:45:21 [k2pts]
hi chump
17:45:24 [k2pts]
chump: hi
17:45:24 [chump]
Not understood: hi
17:45:28 [k2pts]
chump, hi
17:45:37 [k2pts]
dave: how do you say hi to chump
17:45:55 [Starets]
'sokay. I see there's googlage.
17:46:29 [k2pts]
np. have to write an announcement for the pre-release of the new openfts...bbiaw
17:47:17 [Starets]
openfts?
17:47:29 [talli]
Starets, openfts is the new search package for PG
17:47:40 [talli]
k2pts is one of the openfts wizards
17:47:56 [k2pts]
http://openfts.sourceforge.net
17:47:56 [chump]
H: http://openfts.sourceforge.net from k2pts
17:47:56 [Starets]
Ah. Thanks. BTW, is there a recommended replication solution for PG yet? Googlage seems murky.
17:48:13 [talli]
not really
17:48:23 [Starets]
Dang.
17:48:30 [talli]
you mean synchronous or asynchronous replication/
17:48:31 [talli]
?
17:48:41 [Starets]
Hoping for async.
17:48:45 [davb]
k2pts: I don't know if you can :)
17:48:49 [talli]
that shouldn't be a prob
17:48:57 [k2pts]
I think docwold did it once...
17:49:04 [davb]
techent.postgresql.org for replication info
17:49:14 [k2pts]
Starets: don't listen to talli about technical stuff :)
17:49:16 [davb]
wait thats oracle :)
17:49:31 [Starets]
Was looking for (sorry) MySQL-equivalent functionality.
17:49:39 [Starets]
* Starets hangs head in shame
17:50:15 [davb]
make that techdocs.postgresql.org
17:50:22 [davb]
it has links to whatever is out there.
17:50:23 [talli]
Starets, in here people end up sleeping with the fishes in new concrete boots for talking stuff like that.
17:50:38 [lilo]
[Global Notice] Hi all. Just a reminder that non-critical announcements and OPN administrative information can be found on WALLOPS. One of the following commands should turn on WALLOPS on your client: /umode +w, /mode <yournick> +w, /quote mode <yournick> +w, /rawmode mode <yournick> +w .... Thanks, and thank you for using OPN!
17:51:09 [Starets]
davb: no clear winner appears, and pages like erserver.com seem not updated in a scary long time.
17:52:48 [davb]
ah. I never checked. Hopefully someone has something in the works.
17:53:34 [Starets]
tree_sortkey is not like Celko's Trees in SQL.
17:53:38 [Starets]
FWIW.
17:54:46 [k2pts]
Starets: please explain. I have just browsed the article but it looked a lot like tree_sortkey
17:55:53 [Starets]
Tree_sortkey appears to maintain the parent node id as part of the item's info, right? And then uses a sortkey value to help get things sorted right in aggregate operations. Am I reading it right?
17:55:58 [talli]
Starets, feel free to correct k2pts english. make fun of it, if you feel like it. (this is in response to k2pts flagrant denigration of my technical knowledge)
17:56:16 [talli]
btw, anyone know why my internet is acting funny? it just doesn't wotk good.
17:56:30 [Starets]
Sorry, hon, I'm from Baltimore. We're not allowed to make fun of how others talk.
17:56:53 [talli]
i think it has to do with the thingie that touches the computer and wraps into the wall
17:56:56 [talli]
but i don't know
17:57:06 [Starets]
The cat?
17:57:41 [talli]
no, the cats are plotting their next attacks on one another. they're far from the comptuers
17:57:53 [talli]
btw, k2pts, isn't synch replication supposed to come into PG soon/
17:57:54 [talli]
?
17:58:15 [Starets]
Does synch repl mean that if one is down, they're all down?
17:59:00 [talli]
no, that one db is replicated immediately.
18:00:29 [Starets]
Anyway, in nested sets, the record keeps it's position in a linear walk of the nodes, from left-to-right. So it knows it's left node, and it's right node. Using aggregate operations in SQL, you can use this to get hierarchical lists from self-joins.
18:00:30 [davb]
bbl lunch
18:00:46 [Starets]
Node IDs can be any increasing sequence; no special values required.
18:01:24 [Starets]
The bummer is that if you insert a node in the "middle", you have to update everything to it's "right", which can, I imagine, be expensive. But it's a simple update, done in one SQL statement.
18:01:48 [Starets]
I've used it before, and it works amazingly well, but I haven't tried it on a really big dataset.
18:02:12 [Starets]
You have to really pay attention to the examples in the article, but they do make sense.
18:03:27 [Starets]
Where I've used nested sets, I have one table with rows for the principals, and a separate table for the hierarchy. The second table, since it just has the pk for the principal, and the left and right node id numbers, sorts and indexes well, so retrieval is really fast.
18:03:37 [k2pts]
"you can use this to get hierarchical lists from self-joins."
18:03:58 [Starets]
Sorry, I know that's confusing.
18:04:05 [k2pts]
with the tree_sortkey you don't need self-joins. just a select statement using the tree_sortkey
18:04:26 [Starets]
Yes, but it looks like you have to treat the sortkey specially.
18:04:29 [k2pts]
sorry, I'm working on the announcement and I check in and out
18:04:34 [Starets]
and the values must be carefully planned.
18:04:50 [k2pts]
what do you mean specially.
18:05:15 [Starets]
for one, tree_sortkey is a 4K varchar, not just an int or something.
18:05:30 [Starets]
i'm guessing that if you get the wrong thing in the varchar, it doesn't sort right.
18:05:34 [k2pts]
select tree_sortkey into v_parent_sk from acs_object where object_id=12345;
18:05:42 [k2pts]
Starets: give a min to explain
18:06:00 [k2pts]
"4k varchar": no, the tree_sortkey uses varbits
18:06:19 [k2pts]
ignore my last sql statement
18:06:27 [k2pts]
to get the child of a node
18:06:42 [Starets]
"sql/postgresql/site-nodes-create.sql" line 58 of 378: tree_sortkey varchar(4000)
18:07:37 [k2pts]
have to check that to tell you why it's using varchar(4000). it's probably something else named in the same way. if you want to check the tree_sortkey mechanism, see acs-kernel/sql/postgresql/acs-objects-create.sql
18:07:38 [Starets]
It appears to me that tree_sortkey is an encoded value, so the relationship with other records is not normalized, in a Codd sense, because the relationship is not expressed strictly in the schema.
18:08:48 [Starets]
Before I start sounding like I'm championing something, let me disclaim:
18:08:56 [k2pts]
Starets: we have both the parent_id and the tree_sortkey. this allows to write connect by statements in oracle. postgresql doesn't support tree-like structures for the moment, therefore the tree_sortkey mechanism
18:08:58 [Starets]
I was just asking if ACS used this technique.
18:09:18 [k2pts]
ok
18:09:20 [Starets]
It appears to be different from tree_sortkey to me.
18:09:47 [Starets]
Since this is a technique I've used, I'm familiar with it; likewise, since I haven't used tree_sortkey, I'm not familiar with it.
18:09:48 [k2pts]
I'll write the article later so you can ask me then :)
18:10:35 [Starets]
(but I will say this this technique would not require a "connect-by", so it would use the same SQL on Oracle, PG, or Sybase, which looks like a maintenance plus to me)
18:11:20 [talli]
you guys should have this discussion on the bboards, btw
18:11:32 [k2pts]
I'm never gonna finish that announcement if we keep it like that :) I'm reading the article now and I'll let you know...
18:12:02 [Starets]
I'm going to feed, so go write your announcement and read the article, and we'll catch up later.
18:12:44 [k2pts]
ok, that's sounds better
18:13:39 [Starets]
* Starets cites k2pts for superfluous use of apostrophe and hopes talli is happy now.
18:14:17 [talli]
gonna have to do better than that. you need to repent a lot more for bring that "M" word in here...
18:14:19 [talli]
:)
18:14:50 [Starets]
My neck hurts from all this "hanging in shame."
18:15:06 [Starets]
Starets has left #openacs
18:16:42 [k2pts]
talli: since you hand around chating all day. In case I'm not here when Starets gets back, tell him it's the same thing. only that the tree_sortkey algorithm is faster because it doesn't have to self-join...
18:16:54 [talli]
ok
18:16:56 [k2pts]
s/hand/hang
18:16:59 [k2pts]
thx
18:17:06 [talli]
i'll just direct him to the chatlogs
18:17:16 [k2pts]
tha's sounds goot
18:17:22 [k2pts]
s/goot/good
18:17:26 [k2pts]
:)
18:29:06 [davb]
also: it looks like Starets has old code, all the tree_sortkeys should have been changed to varbit
18:31:56 [k2pts]
yeap
18:35:04 [davb]
davb has changed the topic to: Free Web Toolkit | http://openacs.org | Sense of humor required
19:15:45 [k2pts]
k2pts has quit ("Client Exiting")
19:23:59 [til]
til (~tils@port-212-202-128-195.reverse.qsc.de) has joined #openacs
19:26:04 [talilee]
talilee (~talli@lti-4.dialup.access.net) has joined #openacs
19:26:05 [talli]
talli has quit (Read error: 104 (Connection reset by peer))
19:31:23 [andyn]
andyn has quit (Remote closed the connection)
19:32:09 [andyn]
andyn (~andy@12-254-190-230.client.attbi.com) has joined #openacs
19:47:17 [Starets]
Starets (~Starets@home.clickvision.com) has joined #openacs
19:59:21 [talilee]
talilee is now known as talli
20:00:52 [Starets]
Starets has left #openacs
20:33:24 [denshi]
denshi (~chatzilla@adsl-216-62-223-193.dsl.rcsntx.swbell.net) has joined #openacs
20:33:43 [talli]
denshi, you think you can just come in here and act like it's your domain?
20:33:57 [denshi]
talli, all your base are belong to me. just accept it.
20:35:06 [talli]
very original. i'll have you know that ... hold on, where's the beef?
20:35:26 [denshi]
I am the master of your domain.
20:35:59 [talli]
you're not even the technical contact of my domain
20:37:16 [denshi]
15-Love.
20:38:01 [talli]
what are you doing here today, denshi? finished editing that hardcore, XXX porn movie yet?
20:38:19 [denshi]
we're up to four X's these days.
20:38:20 [talli]
or is it a taco bell commercial this week?
20:38:37 [talli]
hey, did you follow up with that TOra gig? were you interested?
20:39:17 [denshi]
actually, I forgot to nub you with that. I asked the channel if anyone has used TOra before.
20:39:34 [talli]
kapil (hazmat) has, and he said it was awesome
20:39:39 [talli]
but he hasn't been around for a while
20:39:41 [denshi]
It looks like clean code, but the Oracle assumptions go down *all the fricken' way*.
20:39:51 [talli]
ah, i see
20:39:51 [denshi]
So it would be an incredible pain in the ass.
20:40:02 [talli]
you know how much the dude was paying, though?
20:40:13 [denshi]
I am sending some mail to that character, to hash technicalities.
20:40:23 [talli]
cool
20:40:33 [talli]
so you want to take the gig?
20:40:57 [denshi]
if it's enough, sure I'll go for it. But if it's not it might be better to steal the architecture and redo *huge* elements of it in a more accomodating language.
20:41:08 [talli]
oh, cool.
20:41:13 [talli]
like, uhm, ruby?
20:41:51 [talli]
also, did you ever hear back from paul graham?
20:41:51 [denshi]
no, I don't think Ruby is really stable enough yet. But it's a nice idea.
20:42:01 [talli]
java? c#?
20:42:14 [denshi]
I'm saving my profanity for paul graham to when I throw down some example code.
20:42:22 [talli]
ah, i see
20:42:24 [denshi]
but he didn't respond.
20:42:31 [talli]
bummer
20:42:54 [talli]
what are you thinking of redoing tora in?
20:43:00 [talli]
it's written in C++, right?
20:43:02 [denshi]
it's still in pot-dream theory stages, anyway.
20:43:06 [denshi]
back to TOra..
20:43:14 [denshi]
it's in C++ interfacing with Qt.
20:43:39 [denshi]
I think that might be why C++ is used -- Qt is very nice. I wrote a small game in C++/Qt several years ago.
20:44:00 [davb]
denshi: did you see www.paulgraham.com/taste.html
20:44:41 [talli]
there aren't many replacements of C++ available, are there?
20:44:50 [denshi]
I haven't focused on a language for TOra, after all, I'm not committed. If I were, I'd look for something with easy connection to Qt, and reuse as much GUI widgetry as possible.
20:45:00 [denshi]
I just wish TOra was written in more layers.
20:45:15 [denshi]
replacements for C++?? What do you mean?
20:45:25 [talli]
"a more accomodating language"
20:45:44 [talli]
i was just wondering what you meant
20:45:47 [denshi]
davb: yeah, I saw it. I'd rather he showed me some code. taste.html was a rehash.
20:46:08 [davb]
ah :)
20:46:29 [denshi]
talli: something more LISPy. Even just pulling the Oraclisms out of the C++, embedding a scripting language in TOra, and writing the logic in that would be much nicer.
20:46:42 [talli]
that seems cool
20:47:03 [talli]
now this is totally out of left field, but owuld it be possible to adapt it so that it was kinda OACS specific?
20:47:15 [denshi]
yes.
20:48:11 [talli]
neat.
20:48:22 [talli]
what would you think it might look like?
20:48:36 [denshi]
it would be nice to have some roles for OACS ("oaks") in it, where OACS objects are presented in a sane manner to the user, prioritized over other random things.
20:50:26 [denshi]
so, in a simple example, if you're laying out a data model, acs_objects & such are near at hand and given priority over tables whose linkages are too many hops away from the module being worked on.
20:50:40 [talli]
neat
20:50:47 [denshi]
or in administration, focus would be given to views we know to be troublesome to performance.
20:50:51 [talli]
is that a hard algorthm or behavior to build?
20:51:38 [davb]
http://diveintomark.org/archives/00000137.html
20:51:38 [chump]
I: http://diveintomark.org/archives/00000137.html from davb
20:51:40 [denshi]
that behavior is getting pretty common in IDEs, but I haven't seen it used w/ SQL.
20:51:54 [davb]
I:|Why you should use CSS
20:51:54 [chump]
titled item I
20:51:57 [denshi]
the network algorithms are something I'm trying to focus on.
20:52:13 [davb]
I:"OK, in the spirit of Richard Fish, here is a non-moralistic argument for CSS: because it will save you money and make you money.
20:52:13 [chump]
commented item I
20:52:29 [talli]
what do you mean "network algorithms"?
20:52:45 [talli]
what kind of networking are you referring to? within the DB?
20:53:41 [denshi]
networks as in objects in a network aka non-directed graph. you've heard of trees referred to as 'directed asymmetric graphs'?
20:54:16 [talli]
no, i haven't
20:54:21 [denshi]
now you have.
20:54:25 [talli]
touche
20:55:05 [k2pts]
k2pts (~nkd@adsl-168-174.cytanet.com.cy) has joined #openacs
20:55:11 [denshi]
so, in this case, following a map of sql tables connected by referential constraints could be considered a network.
20:55:14 [talli]
hey k2pts
20:55:17 [k2pts]
hi talli
20:55:20 [k2pts]
hi guys
20:55:26 [talli]
denshi, ah, ok
20:55:28 [talli]
understood
20:55:32 [davb]
hi k2pts
20:55:35 [denshi]
hey k2pts. sorry I got cut off yesterday. my mac melted into goo.
20:55:36 [k2pts]
hi dave
20:55:49 [k2pts]
np, I was busy myself (still am)
20:55:52 [k2pts]
:)
20:58:00 [talli]
denshi, how are you proposing addressing this issue? and how do you think a scripting language can be embedded?
20:59:22 [denshi]
embedding scripting languages are easy. it's deciding on clean flow-of-control between the layers that is tricky. but still kind of easy, just time consuming.
20:59:37 [denshi]
i mean, it's a single-user desktop application.
20:59:48 [talli]
right, cool
21:02:28 [denshi]
but first I'll hear back from nicolas (the proposed sponsor) to see what he wants. I mean, does he think we can fold this into the main tree? I doubt such a thing is possible.
21:02:58 [talli]
becuase it would require such a fork?
21:03:42 [denshi]
indeed. merging the two forks would be equivalent to merging ACS3 & OACS3. Both would need to redesign to generalize.
21:03:50 [talli]
ah, i see
21:04:13 [talli]
i know that many people have been trying to add PG support via patches, but it seems that those patches must be pretty serious hacks
21:05:34 [denshi]
can you send me some links to that? I've just seen patches to the foreign client stuff.
21:05:41 [talli]
is converting the TOra code to PG heavy work or mostly just converting the heavy oraclisms?
21:05:55 [talli]
i just see the mailing list
21:06:10 [talli]
everything should have been folded into the code, i guess
21:06:38 [denshi]
a little of column A, a little of column B....
21:07:04 [talli]
i see
21:07:11 [denshi]
TOra is *really* complete for an Oracle product. But think about how many features in Oracle aren't present in Postgres.
21:07:30 [talli]
that's true.
21:07:51 [denshi]
or how many features shared by them that behave differently, even more that just syntactical differences. so it's a nightmare on that front.
21:08:06 [talli]
yeah, true enough
21:08:47 [talli]
i imagine it would be a nightmare to try and add OACS support so that it would work with oracle and PG
21:08:49 [denshi]
OTOH, maybe that's a blessing in disguise. One could simulate tablespaces in Postgres if the admin operated solely through this tool.
21:09:20 [denshi]
actually, that sounds like a pretty awful idea.
21:09:35 [denshi]
almost as bad as Don promised a Tcl/Java merge. :)
21:09:42 [talli]
:)
21:11:22 [denshi]
ah, I should shut up. this is taking up too much time.
21:11:37 [talli]
story of your life, huh?
21:12:13 [talli]
bbl
21:12:15 [denshi]
that and not enough money.
21:12:18 [denshi]
later.
21:14:12 [denshi]
hey, davb, who wrote the 'ACS Permissions Tediously Explained' doc you linked yesterday?
21:15:29 [davb]
I am not sure. doesn't it say at the top?
21:15:31 [davb]
:)
21:18:56 [denshi]
nope, my only clue is that it's in Vadim's file-storage space.
21:19:05 [denshi]
but it RAWKS!
21:20:28 [davb]
I think that might be the author.. not sure. rbm knows
21:23:52 [davb]
gotta go
21:23:54 [davb]
davb has quit ()
21:38:08 [markd2]
markd2 (~Snak@r-41.166.alltel.net) has joined #openacs
21:38:19 [denshi]
denshi has quit (Read error: 104 (Connection reset by peer))
21:44:16 [k2pts]
k2pts has left #openacs
22:08:12 [markd2]
markd2 has quit ("wheeeee")
22:44:24 [markd2]
markd2 (~Snak@r-41.166.alltel.net) has joined #openacs
22:46:15 [beattiek]
beattiek has quit (Read error: 110 (Connection timed out))
22:48:08 [beattiek]
beattiek (beattiek@bmhd25b1y14ha.bc.hsia.telus.net) has joined #openacs
23:42:20 [anlater]
anlater (~anlater@212.34.222.62) has joined #openacs
23:47:57 [docwolf]
docwolf (~docwolf@adsl-34-183-205.bct.bellsouth.net) has joined #openacs