IRC log of #schooltool for Friday, 2004-10-01

**** BEGIN LOGGING AT Fri Oct 1 12:37:54 2004
-->You are now talking on #schooltool12:37
---gregweb is now known as gregaway13:02
---Disconnected ().13:59
**** ENDING LOGGING AT Fri Oct 1 13:59:17 2004
**** BEGIN LOGGING AT Fri Oct 1 15:30:11 2004
-->You are now talking on #schooltool15:30
-->alga ( has joined #SchoolTool15:56
-->pips ( has joined #schooltool16:09
<--gregaway ( has left #schooltool16:19
-->gregweb ( has joined #schooltool16:20
<--pips has quit ("Chatzilla 0.9.64f [Mozilla rv:1.7/20040707]")16:21
gregwebHi Marius!16:36
gregwebI'm Gregoire. You probaly do not remember me. But we met at the Europython sprint more than a year ago.16:37
gregwebI'm currently looking at schoolbell. I have some questions regarding the timetable schema.16:39
gregwebDo have time at the moment?16:39
algayes! you were "the responsible one" in Louvain-La-Neuve!16:42
algawe pair-programmed on mail system a bit :-)16:42
gregwebOne of them, yes!16:42
gregwebYes, we worked on a global smtp service as I remember.16:42
gregwebwork == I looked and asked, you typed :-)16:43
gregwebhow are you!16:43
algafine, and you?16:43
gregwebI know you're currently finishing a project.16:44
gregwebfine, also!16:44
algain fact we're getting ready for the second chunk of work on schoolbell16:44
gregwebI'll have some time during october to work also on school tool.16:45
gregwebI'm currently like to understand timetabling of m716:46
algawhat are you going to work on?16:46
gregwebw. schooltool?16:46
gregwebI'd like to be able to make a demo mid november at a school here.16:47
gregwebWe probably have different timetable requirements than others.16:47
algahave you seen this post by steve:
gregwebAnd now I like to first understand hwo it works in m7 and what can be done to fit it to our needs (or at least partly).16:48
algahe outlines our assumtions there16:48
*gregweb reading16:48
gregwebConcretely: What is a timetable schema? What is it meant for?16:49
mgedmina timetable schema is a definition of how a timetable looks like16:51
mgedminit defines what days there are in a timetable16:51
mgedminand what periods there are in each day16:51
mgedmin(different days may have a different set of periods)16:51
mgedmin(or the same periods but in a different order)16:51
mgedmindays in a timetable schema may match week days, or they may be completely unrelated16:52
mgedminwe had a bunch of example timetables from various schools and then tried to abstract everything16:52
mgedminthe result might be a bit too abstract and therefore hard to understand16:52
gregwebok, Q: Is 08:15-09:00 a period?16:52
mgedminsort of16:53
mgedmina timetable schema also has a mapping of period to times16:53
gregwebI understand16:53
mgedminso "Period 1" may be "8:15-9:00" on Mondays, but "8:15-8:45" on Wednesdays16:53
gregwebOk. I understand why I didn't understand :-)16:54
gregwebWe don't need/have this mapping.16:54
gregwebSo for us teh best would be to name the period e.g. 16:45-18:15 and then fill in 16:45-18:15 for every day (same times every day).16:57
gregwebThere are no names given to this periods in our case.16:57
mgedminI think Period 1, Period 2, Period 3 is better16:57
mgedminbut it's up to you16:58
mgedminwhen a concrete timetable is defined, you only have days and period names16:58
mgedminand then you fill in lectures in that matrix16:58
mgedmin(actually you cannot do that through the web interface, but only with the wxWidgets clients)16:59
mgedminin Lithuanian schools we usually have numbered periods, so 1, 2, 3 etc seems very natural for me16:59
gregwebYes, I understand16:59
mgedminI've seen timetable examples where periods were named A, B, C and then mixed up in different days (A, B, C on Monday, B, C, A on Tuesday, etc.)16:59
gregwebWe always talk about times themselves: One pupil to the other: "They moved our math class to tomorrow at 15:15, did you realize"17:00
gregwebQ: what to do if a period that exists monday to friday does not exist e.g. at saturday?17:04
gregwebCan we leave empty the fileds?17:05
mgedminif you do not assign a time to a period on a particular week day, that period is skipped17:06
mgedminif timetable days correspond to week days, then it is even better to exclude that period from the timetable schema17:07
mgedminhave you seen the schoolbell web interface?17:07
mgedminspecifically the timetable schema definition view?17:07
gregwebmay we send you a screenshot of what we want to achieve?17:07
gregwebWe're currently on teh web interface17:07
mgedmindo you know my email?17:07
gregwebYes we're filling out the timetable schema of the web interface17:08
gregwebe-mail: yes17:12
gregwebyou should just have gotten the e-mail from my pair17:16
gregwebmgedmin: did you get teh screenshot?17:17
mgedminone moment17:18
mgedminyes, I hot it17:18
gregwebWhen we press update the saturday times move up to the first and second row.17:20
gregwebDo you understand how we like to define the timetable schema?17:21
gregwebby teh way: if you're short on time, mention it.17:22
mgedminwhy do you want to keep the periods at the bottom?17:23
mgedminthat's not possible in schoolbell -- the day just has an ordered list of periods17:24
gregwebyou mean the 08:15-10:00 one?17:24
mgedmingaps are not allowed in that list17:24
mgedminwhen you assign times to periods, then you can have gaps17:24
mgedmin(gaps in time, that is)17:24
gregwebIn our case tehre are no classes in the vening at saturday.17:25
gregwebNevertheless we have to define the evening periods.17:25
mgedminwhy do you have to define them?17:26
gregwebIs it then later possible to not allocate classes in teh vening of saturday?17:26
gregwebIs it possible to have more than one timetable schema? I think yes.17:27
gregwebShall we define one for saturday and one for the week days?17:27
gregwebwhy do you have to define them?: A: Define Saturday or what?17:28
mgedmincould you send me a mockup of an actual timetable17:30
mgedminit would be easier for me to understand what exactly you want to get17:30
gregwebWe currently realize that we have to give the periods names.17:30
gregwebwe send you a timetable (one minute)17:33
gregwebjust sent17:36
gregwebgot it?17:37
---alga is now known as alga|away17:43
*mgedmin is mildly busy17:50
mgedminok I got it17:52
mgedminI suggest you take a set of all times that appear on Mondays and use that set as period names17:53
mgedminditto for all other days17:54
-->gregweb_ ( has joined #schooltool17:54
mgedminlost connection?17:54
gregweb_Sorry, my battery went down...17:54
mgedmin<mgedmin> I suggest you take a set of all times that appear on Mondays and use that set as period names17:54
mgedmin<mgedmin> ditto for all other days17:54
gregweb_I currently try another approch (perhaps the one you mention):17:55
gregweb_I define one day and within this day I define all possible periods.17:57
-->th1a ( has joined #schooltool18:02
gregweb_Is that the only way currently?18:05
gregweb_for our case?18:05
gregweb_I made a pdf out of the above pages. You may delete them if you like.18:05
mgedminI think I understand your question now18:06
mgedmindo you assume that because 16.45-18.15 on Freitag and 8.15-10.00 on Samstag are displayed in the same row on the first table that they necessarily have to occur at the same time?18:07
mgedminit is not so18:07
mgedminon the bottom table you will get different rows for different period names18:07
mgedminand you may assign different times for them18:07
gregweb_In the weekdays there are no classes during the day and saturday none at the evening.18:08
<--gregweb has quit (Read error: 238 (Connection timed out))18:08
mgedminmy mockup matches the PDF you sent me18:09
mgedmin(at least I tried to, I did not study the PDF extensively)18:09
mgedminI think that you're trying to do something different now18:09
mgedminanyway the gifs you sent me look fine18:10
gregweb_Sorry, there are some classes in the morning. I didn't realize. This has to do with the fact that two schools share rooms.18:11
gregweb_One school mainly has classes during the day and teh other in the evenings (mainly :-).18:12
gregweb_The mockup is only for one department. They seem not to have classes saturday.18:14
gregweb_I have a knot somewhere. I'll play more... how long will you be on the chat today?18:14
gregweb_Thanks a lot!18:15
gregweb_Actually there are three schools sharing resources ...18:18
mgedminI'll be here for at least 2 more hours18:22
mgedminprobably 3 at least18:22
gregweb_Is it ok to eventually ask some more questions?18:23
<--tvon|x31 has quit ("Leaving")18:23
---alga|away is now known as alga18:26
algahi Tom!18:26
th1aHi Alga.18:26
th1aWe had a stressful night last night here in the USA.18:27
th1aThe first presidential debate.18:27
algaah :-)18:27
th1aThe fate of the free world was hanging on John Kerry's shoulders.18:27
th1aHe did ok.  Bush did pretty badly.18:28
th1aAnyway, what have you got on the proposal?18:29
-->gregweb ( has joined #schooltool18:32
<--gregweb_ ( has left #schooltool18:32
algaI added a new story, Recurring events18:32
th1aOK.  I think that's necessary.18:33
algathis is a tough one, a lot of internal calendar infrastructure rework, and the spec is messy18:33
th1aI know.18:33
algaso it will be something like 2 weeks of work or so18:33
th1aEek.  OK.18:34
th1aThere's no getting around it.18:34
algamy blurb on it is:18:34
algaHandling recurring events in the calendar.  Each event will support18:34
algathe full iCalendar specs recurring event model.  iCal views will18:34
algasupport retrieving and creating recurring events, so will web views.18:34
th1aSounds good.18:35
algathen, there's a URI refactoring story18:35
th1aOK.  What's the time estimate on that?18:35
algaRefactor URIs to be normal objects, not interfaces.  They should be18:35
algaavailable to the client over a restive inteface.18:35
algaEstimate: 4 developer days18:35
th1aIs this on the wiki page?18:35
algaI'm editing in on our internal wiki18:36
algaI'll send you the result once I'm done18:36
algaOr maybe I should put it on
th1aYeah, put it on the website.18:36
algathen, there are unclear things re. the basic functionality18:36
algayou expressed some concerns regarding Group, person creation, rename, deletion18:37
algaand also enforcing uniqueness of names18:37
algacould you phrase what exactly you want done?18:38
th1aWhen a person, resource or group is created SchoolBell should warn the user if another object with that title or first and last name already exist, either through the web interface or an error code over the REST interface.18:41
mgedminhow do you deal with collisions?18:42
mgedmin(there are cases where people have the first and last names in real life)18:42
algain a sample of 1000 people there's usually a name clash18:43
mgedminour university faculty stopped giving students email aliases of the form "name.username@domain" for this reason18:43
th1aWell, you need to warn the user when it comes up. 18:43
mgedmina confirmation dialog then?18:43
th1aI guess it is trickier on the REST interface to give a warning.18:43
algawhat about changing usernames?18:43
th1aYeah.  Confirmation.18:43
mgedmin"A user called John Ivanauskas already exists, do you want to add another user with the same name?"18:43
mgedminyes, that would catch people accidentally entering the same persion into the system twice18:44
th1aHaving people entered twice tends to be a big problem in school systems.18:44
th1aAs does confusing two people with the same name :-)18:44
th1alet's see...18:46
th1aOh.  when groups or people are created through the web interface they should be assigned a parent group.18:47
mgedminI suggest a drop-down in the adding form18:47
th1aProbably there should be a default group, like 'people' and a drop down.18:48
mgedminAdd this person to group [People v]18:48
th1aAnd also it would be good if showed links to all the persons.18:49
th1aSame for resources and groups.18:49
algamgedmin: told you!!!18:49
th1aI'd think that should work for the REST interface as well as the web interface.18:50
th1aI guess the last big thing is HTML calendar views for resources.18:51
mgedminI'm pretty sure that works for the REST interface18:51
th1aIt does.18:52
algalists of groups and persons will be 1/2 day18:52
mgedminalga: I was always for showing all persons, wasn't I?18:52
th1aBut you can't set the ACL yet.18:52
mgedminand resources18:52
mgedminth1a: what do you mean about the ACL?18:52
mgedminallowing only certain users to see the list of all persons/groups/resources?18:53
mgedminallowing certain users to add new persons/groups/resources?18:53
mgedmindo I understand you correctly?18:53
th1aOh, I was referring to setting ACL for resources.18:54
th1aWe probably need to start paying a little more attention to access control throughout.18:54
th1aOn those "index" pages of persons/groups/resources, the user should only see listed the ones that he has the 'view' permission for.18:55
th1aPerhaps we can ignore that for now.18:56
algaok, suppose we're deleting a group.18:57
algawhat if some persons are unconnected after that?  put them somewhere?18:57
th1aWell, that's one reason to have the "persons" index page.19:00
th1aI don't know if we should also have a "all persons" group.19:02
th1aIt is a little redundant, but I think the users would appreciate it.19:02
algaif so, should membership in it be enforced?19:03
th1aWell, perhaps only the system should be able to add people to those groups and do it automatically and transparently.19:04
algaso, they should not be like normal groups19:04
th1aRight now it is really hard to navigate to people in ST.19:05
mgedminmaybe we need a search form19:05
th1aWe need to enforce a little more structure behind the scenes.19:05
algaI think a list of all persons will do it in the short term19:05
th1aIn the long run we need to work on a few issues related to searching, but I think that can wait.19:06
mgedminanother thing that we've talked about before: making the timetable schema definition easier to use for simple cases19:07
th1aPeople will have more attributes soon.19:07
th1aYeah.  How many days have we piled onto this milestone at this point?19:07
algaIs it important?19:08
th1aDon't we keep the milestones around a month?19:08
algawell, we found out that our velocity is about 0.6619:09
algathat means that we do a month of project work in 1.5 months19:10
th1amgedmin:  yeah, we should simplify the schema definition for simple cases.19:11
algath1a: I suggest the following action plan: I write down all we came up with, send it to you, you choose what you want done in the next step19:12
algaResponsible for a resource19:16
algaA resource's event has to have a property stipulating who is responsible for19:16
algathe resource during the scheduled block of time.19:16
algaCan a group be responsible for a resource?19:16
th1aI guess so. 19:17
-->hazmat ( has joined #schooltool19:21
algath1a: do you have any concrete stories on ACLs that aren't in the proposal yet?19:22
hazmatalga, have you guys considered davacl as a possible implementation path?19:23
th1aNo, not other than allowing ACL's on resources to be set via the web interface.19:24
algawhat would an acl on a resource allow/limit?19:26
th1aview/add/modify -- just like people19:27
algahazmat: no, not yet19:27
algawhat exactly is it?19:28
algath1a: add a new resource?19:29
algamodify -- add to groups?19:29
th1aalga:  no, I mean the resource's calendar.19:29
th1aadd/modify/view events on the calendar.19:30
algaI thought we have done that already...19:30
hazmatalga, its a dav spec for permissions on resources, taking into account hierarchical permissions, hierarchical principles, and dav operations to manipulate security on resources, including the standard ones being considered here at the moment (view/modify/delete)19:31
th1aThere just isn't an HTML calendar view for a resource.19:31
th1aThat may be the only problem.19:31
algaah, ok19:32
th1aPeople will just want to browse the resource's calendar and add events.19:32
mgedminI think we don't allow that currently19:33
mgedminresource calendars are read-only and only change when events are booked or unbooked19:33
mgedminor am I mistaken?19:33
mgedminah, that only applies for the iCal RESTive views19:34
th1aEvents for resources are fundamentally the same as events for people, right?19:34
mgedminwe have some special magic for resource booking19:35
mgedminthe same event appears in two calendars (the person's who booked the resource and the resource's)19:35
th1aWhich is handy!19:35
algaand should disappear if it is modified/deleted in either place19:36
mgedminpreviously we did not have globally unique IDs for events19:36
mgedminso updating the resource calendar via iCal would break this linking19:36
th1aNo I don't want to do it via iCal.19:36
mgedmintherefore we disabled updates through iCal19:36
th1aYeah, but I want a more graphical method via the web.19:36
th1aThan we have now.19:37
algaso, essentially we have to let the manager specify who can book the resource and who can cancel smb else's booking?19:39
algaand who can view the bookings19:40
algath1a: I counted all estimates, it's ~36 dev days19:42
algawe can commit to do it in 6 weeks or so19:43
th1aOK. 19:43
th1aCan you do the link refactoring first?19:45
-dmwaters_-{global notice} Hi all! You guys can blame this netsplit on me.:) I restarted the firewall on that server, and didn't realize it wasn't running to start with, so it dropped all active connections.... sorry about that.... Thank you for your patience, and thank you for using freenode!19:45
th1aEtria will be getting started next week, too.19:45
algawhat will they be doing?19:46
th1aOh, that's right, you were away, which reminds me, best wishes on your marriage!19:47
th1aThey're doing 'person info' -- a system for handling demographic data, addresses, family relationships, etc.19:48
algacool, the practically-needed missing part!19:48
th1aWe're planning it to be very customizable.19:48
algaoh oh19:52
algaWe had another idea19:52
algathe user should be able to select which groups' calendars should appear on his19:53
th1aYes.  Put that in.19:53
algaso a member of a group could schedule an event for a group, and all members would see it19:53
th1aYou mean in the web interface.19:54
algaiCal does that for free19:54
th1aA composite view.19:54
algathe current view is already a composite19:54
algait merges the r/o timetable calendar onto the editable calendar19:54
algawe thought of firing up the 'dogfood' schoolbell server for ourselves19:59
algaand this was the feature request that immediately popped up for us20:00
algatom, have you figured out the way the timetable manual entry should look like?20:01
algaprobably there should be an individual view for timetables20:03
th1aI have a better idea of what it should look like, I think.  Actually the way the wxPython interface handles it is pretty close.20:04
algabut the web should be able to do it better, right?20:05
th1aWell, I'd really like to use more drag and drop.20:08
th1aBut maybe I'm on crack.20:09
th1aDrag the students onto the classes.20:09
algathat's one thing web deprives you of20:19
algadrag and drop means javascript20:20
algaand usually you don't want it20:20
th1aI know.20:20
th1aIt would be worth using Javascript in this case.20:24
algawith all the added stories the total sum is 60.5 dev days20:24
th1aOh.  Yeah, I thought we were getting into that range...20:24
th1aWell, send me the list.20:24
algaTom, what about splitting the work into 2 milestones?20:32
algathis way we could ask the payment to be delivered in more parts20:33
algaand you would have some kind of checkpoint20:33
algacause this piece of work is getting quite large20:33
th1aSure.  But we may need to do some SchoolTool work in between.20:35
algahm, do you suggest we should work on the schoolbell branch now20:35
th1aI'm talking to HTH on Monday and I may be committing us to have a certain amount of SchoolTool functionality ready by April.20:36
th1aI think that both the work you're doing and the work Etria is doing in October is equally applicable.20:36
th1aSo I don't think we'll need to truly fork yet.20:36
th1aCould we actually use some of the translation machinery to manage the different interfaces for SchoolTool and SchoolBell?20:37
algaI think it is a good idea.20:39
algaok, I have sent the proposal out.20:44
th1agot it20:45
th1aI'm going to go eat lunch now.20:45
algaone other thing we need to discuss is the story acceptance protocol21:15
algawe would like to know the narrative acceprance test before we start work21:15
algaand it would be good to get feedback once we think the story is completed21:16
algaprobably we should mail you once each story is completed so you could chase it up and make sure it is what you wanted.21:16
th1aYeah.  We need to work that out.  I'm feeling more prepared to handle that.21:55
<--SteveA has quit (Remote closed the connection)23:05
algahave a nice weekend everyone!23:46
<--alga has quit ("leaving")23:46
---Disconnected ().23:49
**** ENDING LOGGING AT Fri Oct 1 23:49:59 2004

Generated by 2.15.1 by Marius Gedminas - find it at!