IRC log of #schooltool for Monday, 2004-10-25

**** BEGIN LOGGING AT Mon Oct 25 13:26:55 2004
-->You are now talking on #schooltool13:26
---Topic for #schooltool is http://schooltool.org13:26
---Topic for #schooltool set by Aiste at Mon Oct 18 12:47:26 200413:26
-->thisfred ( has joined #schooltool13:27
---pere is now known as pere_poff15:09
<--Aiste has quit (
<--vinsci has quit (
<--thisfred has quit (
<--SteveA has quit (
<--th1a has quit (
<--pere_poff has quit (
<--rabidbt has quit (
<--hazmat has quit (
<--tvon has quit (
-->thisfred ( has joined #schooltool16:05
-->hazmat ( has joined #schooltool16:05
-->Aiste ( has joined #schooltool16:05
-->vinsci ( has joined #schooltool16:05
-->SteveA ( has joined #schooltool16:05
-->pere_poff ( has joined #schooltool16:05
-->th1a ( has joined #schooltool16:05
-->rabidbt (~rabidbt@ has joined #schooltool16:05
-->Kreisquadratur ( has joined #schooltool18:12
<--Kreisquadratur ( has left #schooltool18:15
-->tvon ( has joined #schooltool18:37
---Disconnected (Connection timed out).19:20
**** ENDING LOGGING AT Mon Oct 25 19:20:20 2004
**** BEGIN LOGGING AT Mon Oct 25 20:26:43 2004
-->You are now talking on #schooltool20:26
---Topic for #schooltool is http://schooltool.org20:26
---Topic for #schooltool set by Aiste at Mon Oct 18 12:47:26 200420:26
<--tvon has quit ("Leaving")20:40
-->bskahan ( has joined #schooltool21:09
-->jinty (~jinty@ has joined #schooltool21:09
-->gregweb ( has joined #schooltool21:40
---gregweb is now known as gregaway21:45
-->meimei ( has joined #schooltool21:48
-->pere ( has joined #schooltool21:48
<--meimei has quit (Client Quit)21:50
-->meimei ( has joined #schooltool21:53
-->stockholm ( has joined #schooltool22:29
th1ajinty:  just got your email22:35
jintyah, and? Does it fit with your idea of a community process?22:39
jintya little bit of a social experiment on my part;)22:40
th1aWas it a joke?22:40
th1aI'm a little confused.22:41
jintyno, serious, i think the writer of the spec shouldn't have any own ideas. only listen22:41
th1aOh.  I think I'm refering to a different email.22:41
th1aI meant the earlier one.22:42
th1aI just saw the "Bringing stories..." email.22:42
th1aOK.  That looks good.22:43
jintyyes, that is the one i mean. I want to run every story by the community before it becomes a specification22:43
th1aThat's a good idea.22:43
th1aThe fact of the matter is that the community is largely mute, for various reasons.22:44
th1aBut we have to try anyhow.22:44
jintyI know, but I don't mind sending a lot of mails like that to the list.22:44
th1aI'm going to try to be more open in the process of drafting the next round of specs for POV/Etria as well.22:45
th1aThe primary challenge is getting started early enough.22:45
jintyGood, I think the list is much better because it creates a public record.22:46
jintyYes, to discuss every point can take months22:47
jintyim not expecting anything fast from the timetabling spec22:47
th1aGood.  Because what I'm thinking is that I should spend the next few weeks working out how curriculum will work and have that done in December, which will then make it much easier to do both timetabling and scheduling.22:48
jintyyes, I wanted to talk to you about that.22:49
th1aI've been thinking about curriculum, grade levels, etc.22:49
th1aI think we'll need to use state machines.22:49
jintydon't know anything about curriculum or grade levels.22:50
jintywhat is the use of a state macine?22:50
jintyfor this situation.22:51
th1aSo in the standard US high school, you have four grade levels (9 -12).22:51
th1aEach grade level will be represented by a different state.22:51
th1aThe transition from on state to the next is triggered by specific events.22:52
jintysuch as passing exams?22:52
th1aNot in US high schools, actually, but probably in the UK.22:52
th1aI think the basic pattern of a state machine should cover the various international differences.22:53
th1aIn the US you pass if you get a passing grade for the semester overall, which isn't tied to a specific exam, btw.22:53
jintyOk, so the state macine is the basic infrastructure and the transitions are the customizations.22:54
th1aSo your scheduling app wants to be able to get the state of each student to determine which classes they require or are eligible for.22:55
th1aYeah.  The state machine is the basic model.22:55
th1aThe states and transitions would be customized.  Mostly by country.22:55
jintybut this type of specification is a long time away for the scheduling app.22:56
jintyI mean that the inital work will be in creating something that can solve ANY timetabling problem22:56
jintywith some way of adding customizations for a specific type of problem22:57
th1aYeah.  My perspective on this is skewed, because I'm only worried about defining the constraints.22:57
th1aI'm sure we can find someone to solve the combinatorial problem, once we've properly defined it.22:58
jintyyou mean automatically solving the timetable?22:59
th1aCheck this out:
jintythat is the point of the TTML. That all problems can be converted to TTML.23:00
jintytherefore, in theory a automatic solver for TTML can solve ANY timetabling problem.23:01
jintyIt makes for an ideal interface because we can let the automatic solver people do what they want23:02
th1aSo are there such automatic solvers?23:03
th1aOr is that what we're paying for?23:03
jintyDr Ozcan has one23:04
jintyAt least he solves a sample TTML problem with one23:04
jintybut i have no idea about it23:05
th1aHm.  So really, we could go ahead an pay for work on a TTML solver right now.23:05
th1aWhat do you think about the idea of having a contest?23:06
jintyWe just have to be carefull, they may write a solver for a basic TTML problem23:06
jintyi.e. not taking into account all of the constraints we need.23:07
th1aWell, but we could write a problem that included all the mathematically significant types of constraints, right?23:08
jintyprobably first we need to make a TTML problem that can be a reference problem for them to solve23:08
th1aI guess I need to dig into TTML a bit more.23:08
th1aPast a certain point, though, this is way more math than I can grok.23:08
jintyusing all the functionality we need. Then base the competition on that.23:08
jintyIt took me a while to get my head around it, but I still havn't fully comprehended it.23:09
th1aHm.  So really, we could give the timetabling algorithm geeks something to do.23:10
th1aWhich is my primary short term objective...23:10
jintyUnleash the wild horses, yes:)23:10
SysTray Integration Plugin unloaded23:10
Python interface unloaded23:10
Tcl interface unloaded23:10
**** ENDING LOGGING AT Mon Oct 25 23:10:59 2004

Generated by 2.15.1 by Marius Gedminas - find it at!