IRC log of openacs on 2002-03-17

Timestamps are in UTC.

00:01:21 [Spork]
Moo
00:04:06 [markd2]
Mü
00:06:15 [Spork]
mu?
00:06:18 [Spork]
why not pi?
00:06:26 [Spork]
Frankly, I prefer cake :)
00:06:31 [markd2]
Pi day is past
00:06:36 [markd2]
In in post-pi deperesison
00:06:40 [Spork]
lol
00:06:45 [Spork]
I got all my channels
00:06:49 [Spork]
How exciting
00:07:20 [Spork]
Guess what I had for dinner
00:07:33 [Spork]
Frog legs! :-)
00:09:00 [markd2]
ribbit
00:09:11 [markd2]
so what you been up to lately?
00:09:18 [markd2]
I don't think I've typed at you in awhile
00:09:31 [Spork]
No you have not :)
00:09:37 [Spork]
I miss my desk job :(
00:09:46 [Spork]
now it's just going on-site all the time
00:10:18 [Spork]
and I'm still an intern
00:10:26 [markd2]
:-(
00:10:35 [markd2]
and you probably don't get the great perks like white-house interns
00:10:55 [Spork]
lol]
00:11:05 [Spork]
I got all the cable channels for free though :)
00:11:23 [markd2]
even the pay and porn ones?
00:11:24 [markd2]
cool!
00:11:38 [Spork]
not pay per view
00:11:43 [Spork]
but supposedly playboy
00:12:25 [Spork]
i think I have to turn it on though
00:12:32 [Spork]
Through parental controls
00:12:49 [markd2]
prenatal controls
00:13:19 [Spork]
yeah cable box comes with 'em
00:25:35 [talli]
talli (~talli@188.muka.lasv.snfccafj.dsl.att.net) has joined #openacs
00:28:39 [davb]
davb (dave@alb-24-58-162-46.nycap.rr.com) has joined #openacs
00:29:06 [talli]
hey guys
00:29:12 [davb]
hi talli
00:29:44 [Spork]
hey guys ;)
00:33:20 [talli]
talli has left #openacs
01:16:20 [markd2]
markd2 has quit ("wheeeee")
01:28:01 [davb]
wow, lars is good.
01:37:15 [talilee]
talilee (~talli@188.muka.lasv.snfccafj.dsl.att.net) has joined #openacs
01:42:11 [davb]
hmmm. acs-api-browser is visible to all registered users...
02:07:08 [GEM]
GEM (~mcmullan@MULTICS.MIT.EDU) has joined #openacs
02:07:21 [GEM]
evening, all
02:23:53 [davb]
hi GEM
02:24:53 [talilee]
talilee has quit (Read error: 110 (Connection timed out))
02:31:23 [davb]
http://homepage.mac.com/iblog/nn4_3col_header.html
02:31:23 [chump]
A: http://homepage.mac.com/iblog/nn4_3col_header.html from davb
02:31:29 [GEM]
hey - is AOLserver 3.4 a good choice for installing OACS4.5?
02:31:36 [davb]
A:|Netscape 4 compatible 3 column CSS layout!
02:31:36 [chump]
titled item A
02:31:43 [davb]
GEM, I think it is ok.
02:32:10 [GEM]
I am installing on my Debian system, and hate to sully it with non-aptable stuff if I can help it :-)
02:32:14 [davb]
but you need nssha1, nsxml, nscache, etc.. are they all in there?
02:32:16 [davb]
right.
02:32:31 [davb]
rbm envisions a day when openacs 4 will itself be a debian package.
02:32:56 [GEM]
I expect that this will happen. If I had more time, I'd learn how to do it and make it available myself
02:33:09 [GEM]
but at the moemnt, there is no way that I can manage that
02:33:20 [davb]
np :)
02:35:28 [GEM]
not sure what is in there, and not sure how best to find out - any ideas? :-)
02:37:51 [GEM]
I'm tempted to just run with what I have and see what breaks, then apt-get remove whatever I need to hand-install
02:39:31 [GEM]
not like it's hard to undo things
02:41:01 [GEM]
hmm. rats - should I have gone with debian standard "www-data" and /var/www or nsadmin and /web? decisions, decisions...
02:41:19 [andyn]
andyn (~andy@12-254-190-230.client.attbi.com) has joined #openacs
02:42:18 [davb]
that is a hard debate. I usually try to go with debian standards for most stuff.
02:42:53 [GEM]
davb: I forgot that you were a debian fan, as well...
02:46:01 [davb]
it is very popular in the channel :)
02:46:23 [GEM]
I wonder why...just cause it is the best distribution available
02:55:23 [talilee]
talilee (~talli@dsl092-013-011.sfo1.dsl.speakeasy.net) has joined #openacs
02:55:46 [talilee]
talilee is now known as talli
03:04:14 [GEM]
'scuse me - time to go to sleep here
03:13:01 [davb]
davb has quit ("Client Exiting")
03:15:22 [jim]
jim (~jim@12-233-187-5.client.attbi.com) has joined #openacs
03:15:31 [jim]
re
03:28:49 [davb]
davb (dave@alb-24-58-162-46.nycap.rr.com) has joined #openacs
03:46:41 [talli]
talli has quit (Read error: 110 (Connection timed out))
03:50:34 [jim]
re davb
03:50:52 [davb]
hi jim
03:54:23 [andyn]
andyn has quit (devlin.openprojects.net irc.openprojects.net)
03:54:23 [tontsa]
tontsa has quit (devlin.openprojects.net irc.openprojects.net)
03:55:14 [andyn]
andyn (~andy@12-254-190-230.client.attbi.com) has joined #OpenACS
03:55:14 [tontsa]
tontsa (tontsa@livingfor.net) has joined #OpenACS
03:55:48 [Zaki]
Zaki (~ddd@212.46.36.39) has joined #openacs
03:55:54 [Zaki]
hi
03:56:21 [davb]
hi Zaki
03:56:57 [Zaki]
jim: thanks but ... i actually asked about a place to chat about issues in starting business on the net.
03:57:08 [Zaki]
davb: hi :)
03:58:02 [jim]
oh :) this is about a web toolkit you can use to do business, or have communities and let it collaborate
04:01:13 [Zaki]
i'm now reading the first page of the website
04:07:45 [jim]
Zaki: let it be one resource among many
04:09:14 [Zaki]
jim: ok, btw, since it seems very serious app, since it's built on aolserver + postgresql :)
04:09:38 [jim]
Zaki: ported from aolserver/oracle
04:10:20 [Zaki]
which server side web programming language used in aolserver?
04:10:26 [jim]
davb: what about getting back into those drop scripts?
04:10:28 [jim]
TCL
04:10:31 [davb]
now it runs on aolserver + oracle OR postgresql, your choice.
04:10:43 [davb]
jim: not a bad idea. won't happen before release I suspect :)
04:11:03 [jim]
not a big deal, I suppose...
04:11:04 [davb]
I don't think don wants to get into changing pratically every pacakge :)
04:11:29 [davb]
no, but one of the standard tests simon setup is unmount and drop package :)
04:11:48 [davb]
which won't work unless you remove all package content first as we well know.
04:12:11 [jim]
and most packages would fail that test if not?
04:12:35 [davb]
yes.
04:13:23 [jim]
then as I read the situation, we have a standoff atm
04:13:49 [jim]
but we can still provide infrastructure...
04:15:56 [davb]
right. Would it be better to fix all the places where on delete cascade can be used?
04:16:12 [davb]
and put it in the developer docs to use it?
04:17:22 [jim]
well, I dunno. first we determine where to do it (it's still my opinion that the glue language (tcl) is the wrong place to do it, for short and long-term reasons)
04:21:16 [jim]
I think it should be in sql or pl/sql because it involves purely database actions, and no tcl shoudl be used because the drop scripts, being in .sql files themselves, couldn't call the tcl
04:22:22 [davb]
jim: we have two different issues. dropping a package instance after it has been mounted and data added to it. and dropping a package entirely from the system including all mounted instances and data, right?
04:22:58 [jim]
davb: I see the possibility, of doing this:
04:23:40 [jim]
dropping a package instance involves removing the data associated with the instance
04:24:07 [jim]
dropping a package iterates thru all instances, calling drop-instance for each one
04:24:43 [jim]
refactoring in this way has the advantage that the code only need be written once
04:24:50 [davb]
right.
04:25:33 [davb]
so here is my compromise that I think will work: pl/psql proc that handles the actual deletion. tcl proc that calls the pl/sql proc from a service-contract so the apm can ask a pcakge to delete an instance of itself.
04:26:13 [jim]
I guess another possibility is to have the drop-something script be able to drop the whole package or restrict the same query to a single instance
04:26:59 [jim]
so the service contract would be a small amount of -glue- code that calls plsql
04:27:04 [davb]
right.
04:27:08 [jim]
that's a fair compromise
04:27:31 [davb]
so that the apm doesn't have to bother figuring out which pl/sql proc is the right one to call.
04:28:25 [jim]
we should alter a "notes" package to do that, given it's the hello of the acs
04:28:59 [davb]
sure. there is already a notes with a service contract for search.
04:29:26 [jim]
and given that's the example, we document to the hilt
04:30:29 [davb]
sure. and the package_instance delete code should check if any objects are children of a pcakge before deleting and give a message back instead of failing spectacularly.
04:31:11 [jim]
hmm, that's a new one...
04:31:14 [jim]
example?
04:32:05 [davb]
of a failure?
04:32:10 [jim]
the package instance delete of a given package should only delete what uses of the package created...
04:32:22 [jim]
of a situation where there are children
04:32:29 [davb]
right.
04:33:01 [davb]
any package that uses acs_objects. the context_id goes back to the package_instance.
04:33:26 [jim]
it shouldnt' need to check, it should be customized to the package's requirements... or am I missing something?
04:34:30 [davb]
the apm code that attempts to call the delete pl/sql proc should check to make sure it worked. or if a service-contract doesn't exist, it should let you know there is data still in there.
04:34:32 [jim]
ok, let's go thru a hypothetical...
04:35:28 [jim]
a package creates an object (its create sql script already having created the object type)...
04:36:04 [jim]
now, the package delete-instance script is about to delete that object...
04:36:22 [jim]
first calls the deletes for the aux tables
04:36:49 [jim]
then calls acs_object__delete(say on 1234)
04:37:45 [jim]
we've already seen that it's a mistake to set the context_id to null, because that defeats the index and causes problems
04:38:10 [davb]
also it is totally unecessary :)
04:38:30 [jim]
so we don't do that, and the object is deleted, hopefully with the rows in the index table going with it
04:40:58 [jim]
my gut feeling tells me that each drop-instance script -must- be customized for each package... drop-instance contract missing for a package? can't delete instance if there is any data. sorry, here's veto.
04:41:12 [davb]
right.
04:41:30 [jim]
(which is the situation we have now)
04:41:43 [jim]
(except hard to fix :)
04:41:57 [davb]
the script will be unique. ie. ETP creates cr-folders. so the script has to delete all the content items first, then the folders.
04:42:13 [jim]
yeah
04:43:45 [davb]
so the only way to fix it is to go through each package and figure out what has to go when an instance is deleted.
04:45:30 [jim]
right...
04:46:45 [jim]
if we get the infrastructure available to people first and teach them how to use it, the fix is a snap and (potentially anyway) the entire set of packages could get their instance droppers in a month of everyone getting it going
04:47:21 [davb]
sounds good.
04:47:40 [davb]
Have you read up on acs-service-contract?
04:48:45 [jim]
to some degree
04:49:29 [jim]
I don't know the precise details, but I gather it stores a representation of something like a C prototype
04:49:59 [jim]
and each package signals its willingness to implement it
04:50:32 [jim]
is one contract -one- such function signature? or many?
04:50:47 [davb]
a contract can consist of for than one operation.
04:50:59 [davb]
so I think many is the answer.
04:51:03 [jim]
ok
04:51:21 [jim]
but this contract only needs one, true?
04:51:28 [davb]
i think so.
04:51:43 [jim]
unless we decide later that we also need instance-new
04:52:02 [jim]
but that part works
04:52:10 [davb]
right.
04:52:12 [jim]
so I wouldn't wanna screw with that
04:52:25 [davb]
I wonder how it works. :)
04:52:47 [jim]
the package mgr would load the data model
04:53:10 [jim]
hmm, then what...
04:53:16 [davb]
but how is a new instance handled....
04:53:28 [jim]
can operations be optional in sc?
04:53:48 [jim]
the site map creates the instance
04:53:51 [davb]
does a implementation have to cover every operation? I don't think so.
04:54:06 [jim]
or is the ui for creating the instance
04:54:29 [davb]
I think all site map actually does is create a row in apm_packages and the corresponsing one in site_nodes. it doesn't actually cause the package to do anything itself.
04:55:12 [davb]
it must do more, like create attributes, etc...
04:55:27 [jim]
so the idea is that initially, there are no additional rows needed in aux tables for new apps
04:55:48 [jim]
and all packages must presently follow that maxim
04:56:00 [davb]
actually I think the apm stores all the info to create a package instance when the package is installed.
04:56:22 [davb]
there is a proc to make some magic happen. I have to find it again...
04:56:51 [jim]
magic is bad for understanding
04:56:55 [jim]
and vise versa :)
04:57:11 [davb]
would it be easier to jus tmake it a package pareameter, and put the pl/sql proc name in the .info file?
04:57:36 [jim]
to create a new one or delete the old one?
04:58:23 [davb]
delete. instaed of the service contract stuff. the apm already stores metadata about packages in the package_attrtributes table.
04:59:10 [jim]
I'd say let the magic lie as it is for now, get a instance_remove(object_id_of_site_node) in a service contract...
04:59:29 [davb]
ok.
04:59:44 [jim]
the service contract would be helpful in the case where a package cannot delete instances that way yet
05:00:27 [jim]
a package should say: "Yes, you -can- use ______ to delete single instances"
05:02:02 [jim]
and what it really does is remove all data in the instance, such that it's like it was when the site node and app to mount there was first created, and before any data got in
05:02:17 [davb]
right.
05:03:16 [jim]
unfortunately, I can't see anything to be done about the non-implementing packages other than let them error out as they do now
05:03:35 [davb]
well, maybe a nicer error message.
05:03:49 [jim]
the only way to do that is actually provide a function that checks for it
05:04:14 [davb]
there is a way to check is there is a binding for a contract.
05:04:27 [jim]
(whether data still exists, where "data" and "exists" are defined differently for each package)
05:04:53 [jim]
hmm
05:05:19 [jim]
but still no way to check for existing data... and that's not the problem we should solve presently
05:05:55 [davb]
we can check if anything is tied to the acs_object of the package_instance.
05:06:14 [jim]
so we see if the delete-instance binding exists... and if it does not, then we have no idea if there is data in that instance or not
05:06:22 [davb]
but we know many packages do not use the object system quite correctly.
05:06:29 [davb]
jim: right.
05:06:36 [jim]
yes, but again, that's a different problem...
05:06:47 [davb]
yes, there are enough of them.
05:07:55 [davb]
right, sorry. I need to focus. I just read an good article by Lars about that. I want to fix everything. :)
05:08:42 [davb]
I think mainly I am unsure if I am going to leave something out and make it harder to do something I haven't thought of later.
05:09:41 [jim]
ok, let's look at the contract binding itself...
05:09:47 [jim]
there are two cases:
05:09:56 [jim]
a package implements the contract
05:09:57 [jim]
or
05:09:59 [jim]
not
05:10:48 [jim]
if it does,
05:10:58 [jim]
call that binding
05:11:03 [jim]
drop that binding
05:11:11 [jim]
if not,
05:11:26 [jim]
# it's already gone cause it was never there
05:11:44 [davb]
jim: the binding can stay, it belongs to the package install, not the package_instance.
05:11:50 [davb]
I think.
05:11:52 [jim]
ok
05:12:54 [jim]
you being right makes sense here...
05:18:20 [jim]
I guess now we'd need a concrete example
05:18:36 [jim]
I don't have an acs atm; drive blew up
05:18:53 [davb]
ouch.
05:19:14 [davb]
I have plenty of acs's . I think I have 4 different installs here on this box.
05:19:56 [davb]
I just am not too sure how to write the contracts.
05:20:31 [jim]
http://etp.museatech.net/doc/ doesn't have sc docs
05:20:31 [chump]
B: http://etp.museatech.net/doc/ from jim
05:20:48 [jim]
can we delete B: :)
05:21:12 [davb]
I really need to fix chump to allow that.
05:21:56 [jim]
chump, forget B:
05:22:05 [jim]
nope :/
05:22:43 [davb]
openacs.org/doc/ doesn't have package docs yet either.
05:23:53 [davb]
if you register at www.thedesignexperience.org you can read my copy. I turned on read permission.
05:24:56 [davb]
try http://www.thedesignexperience.org/register/ there is no link from the home page , oops
05:32:45 [jim]
let's regroup in a coupla days, by then I'll have my openacs at 4.5
05:32:53 [davb]
ok, sounds good.
05:33:00 [jim]
plus, I'm playing with daemontools on debian
05:33:23 [davb]
daemontools is very handy,
05:34:04 [jim]
plus seems easy to set up
05:34:16 [davb]
yes. is it the latest version?
05:34:32 [jim]
well, 0.7.6?
05:34:40 [davb]
that sounds like it.
05:43:12 [davb]
wow, its late for me. see ya!
05:43:44 [jim]
nite davb
05:49:52 [davb]
davb has quit ("Client Exiting")
07:16:52 [Zaki]
Zaki has quit (Connection timed out)
09:06:05 [andyn]
andyn has quit ("leaving")
10:10:29 [andyn]
andyn (~andy@12-254-190-230.client.attbi.com) has joined #openacs
10:23:17 [andyn]
andyn has quit (Remote closed the connection)
11:32:26 [talli]
talli (~talli@talli.dialup.access.net) has joined #openacs
13:21:46 [davb]
davb (dave@alb-24-58-162-46.nycap.rr.com) has joined #openacs
13:50:34 [davb]
davb has quit ("Client Exiting")
16:27:09 [davb]
davb (dave@alb-24-58-162-46.nycap.rr.com) has joined #openacs
17:41:31 [Spork]
Dave
17:41:39 [Spork]
Do you know aye-ess-pee?
18:07:52 [davb]
aigh, noooooo
18:08:06 [davb]
you mean VBscript?
18:23:44 [talli]
talli has left #openacs
18:33:28 [bartt]
bartt (~chatzilla@dsl-64-34-176-81.telocity.com) has joined #openacs
18:36:23 [bartt]
bartt has quit (Client Quit)
18:38:26 [talli]
talli (~talli@lti-4.dialup.access.net) has joined #openacs
18:39:00 [jim]
jim has quit (Remote closed the connection)
18:58:58 [Spork]
I guess VBScript
19:11:49 [davb]
it is the programming language most commonly used with ASP.
19:11:53 [davb]
Another project?
21:11:40 [Spork]
sorta
21:11:49 [Spork]
I have to make a database update form
21:16:44 [Spork]
with ASP and IIS
21:40:45 [davb]
I have done that before. Download the vbscript reference from microsoft.com.
21:41:15 [davb]
check this out, really good info for ASP and databases
21:42:45 [Spork]
i'm using DreamWeaver ultradev
21:43:06 [Spork]
it has a premade template for a db update
21:43:15 [Spork]
I got it to work once...and now I can't get it to work
21:44:46 [davb]
ah. that is a spiffy app.
21:45:02 [davb]
I really don't know how it works, I just fooled around with it.
21:46:07 [rbm]
eeeek
21:46:16 [Spork]
oooom!
21:46:22 [Spork]
Ack, I mean moooo!
21:46:52 [rbm]
* rbm has an epyleptic convulsion at the thought of having to use ASP+IIS
21:47:20 [Spork]
I didn't choose to do it like this
21:47:27 [rbm]
Spork: I believe you.
21:47:38 [Spork]
heh
21:47:58 [Spork]
Roberto, have you seen anything about Digital cable and stuff on demand via coax yet?
21:48:32 [rbm]
Here at USU, they recently spent a boatload of money to "redesign" the website and load it with a gazillion graphics. The "consulting" firm wanted ASP. Computer services didn't want IIS. So they came up with a compromise: Apache+Chilisoft ASP+Win2k
21:49:53 [Spork]
heh
21:50:12 [rbm]
Now they are buying into this "Portals" thing and now they are looking at spending 500,000 just on hardware for this expensive proprietary portal software (that itself costs $500,000)
21:50:46 [Spork]
We (my group) chose IIS, ASP, VB because they only had 4 months development time to come up with a solution
21:51:50 [rbm]
Speaking of portals, has anybody looked into the OACS 4 portals package?
21:52:05 [davb]
nope. I was waiting for the new dotlrn stuff.
21:52:19 [rbm]
dotlrn will have a portals package?
21:52:27 [davb]
Spork: we have video on demand on our cable now. only one movie though so far.
21:52:35 [rbm]
Isn't it going to be pre-released this month?
21:52:42 [davb]
yes and yes.
21:52:53 [Spork]
Dave, that's cool...what provider?
21:53:01 [davb]
Time Warner
21:53:08 [Spork]
they only have one movie?
21:53:11 [Spork]
LOL
21:53:19 [rbm]
That'd be awesome. I'll take a look at it and then try to convince folks to not commit this atrocity of wasting all this money on a proprietary solution.
21:53:24 [Spork]
i thought they were further along than that
21:53:35 [davb]
well we have about 80 channels of movies starting every half hour.
21:53:44 [Spork]
heh
21:53:53 [davb]
good luck rbm.
21:54:35 [davb]
I am hoping I can stop my work from using the blackboard web based course service.
21:54:55 [Spork]
Blackboard sucks
21:54:57 [Spork]
I hated it
21:55:11 [talli]
rbm: dotlrn rewrote portals completely
21:55:29 [talli]
ben posted that they were planning on releasing it "in a couple of days"
21:55:37 [talli]
you may want to write to him and find out what is going on
21:58:34 [talli]
what license is postfix under?
21:58:57 [hazmat]
ibm pl.
21:58:58 [talli]
"
21:58:58 [talli]
21:58:58 [talli]
This document is a FAQ for the Postfix Mail Transfer Agent (MTA) on Red Hat Linux. Postfix is being developed by Wietse Venema and Licensed under the IBM version of the GPL.
21:58:59 [talli]
"
21:59:29 [hazmat]
thats a strange concept.
22:00:40 [rbm]
talli: Wow! Cool!
22:01:10 [talli]
rbm: that may be cool, but it was supposed to be released in january
22:01:37 [davb]
yeah! a new error!
22:01:39 [rbm]
software schedules :)
22:01:51 [talli]
i guess.
22:02:16 [rbm]
I wish they'd update the website
22:02:27 [talli]
i wish they'd release the code
22:02:46 [talli]
or at least the specs
22:03:51 [rbm]
yeah
22:04:11 [rbm]
I can understand why they haven't, but I still wish they did.
22:04:20 [talli]
i can't
22:04:36 [rbm]
(I wish they released it anyways)
22:04:45 [talli]
nothing stopping them from releasing the specifications and design docs
22:05:11 [rbm]
Didn't Ben or Arjun post a message with the reasons why?
22:05:25 [talli]
they weren't good reasons
22:06:27 [talli]
and they were pretty defensive
22:11:05 [rbm]
having the specs so we at least knew what to expect would be nice.
22:11:20 [talli]
yes, it would.
22:16:58 [rbm]
* rbm goes back to studying for his monday test
22:21:54 [rbm]
So I installed the latest OpenOffice stable build. It's a lot faster than last time I tried it.
23:33:19 [davb]
cool.