IRC log of #schooltool for Monday, 2012-03-26

*** alga has quit IRC02:24
*** alga has joined #schooltool02:28
*** aks has joined #schooltool05:48
*** alga has quit IRC06:00
*** aks has quit IRC06:08
*** aks has joined #schooltool06:08
*** yvl has joined #schooltool09:13
*** alga has joined #schooltool14:24
*** th1a has joined #schooltool15:18
*** aks has quit IRC15:35
*** replaceafill has joined #schooltool16:21
*** th1a_ has joined #schooltool16:26
th1a_Hi aelkner, replaceafill, menesis yvl.16:31
replaceafillgood morning/afternoon16:31
yvlhey guys16:31
th1a_Things are a little crazy here today... my baby sitter is out sick.16:33
th1a_Also I have someone coming to clean the gutters at some point here...16:33
th1a_So... yvl, how's the model looking?16:34
yvloh sorry16:36
yvlwas out to make some coffee :)16:37
yvlit is not finished yet16:37
yvlcurrent status16:37
yvlall of the data structures are in place, as well as supporting code16:37
yvlI'm finishing validation at the moment16:38
yvlit's finicky16:38
yvlthings on my TODO list...16:38
th1a_When are you validating?16:38
yvlon node relationship changes16:38
yvland on node layer changes16:38
yvlso that the change does not break the model16:39
yvlI also implemented skills as normal requirements16:39
yvland I'll have to add catalogs for fast grade access16:40
yvlstoring grades in one root place would be worth it when writing from scratch16:40
th1a_I would think it is finicky since it is flexible.16:40
yvlone could think that16:41
yvlreplaceafill could actually start writing "projects" part of the cando gradebook quite soon16:41
yvlah, I need to check skill retirement16:42
yvlbut other then that, it should be pretty much workable16:42
yvloh, and no security declarations in place yet16:42
yvleasier to do when you actually have some views16:42
yvlso the TODO thing16:43
yvlI need some minor helpers, like availableParentNodes and availableNodeLayers16:43
yvlfor node linking views16:43
yvlI want to do a simple importer16:43
yvlI need simple views to display current document model16:44
yvl(both layers and nodes)16:44
yvland add stupid relationship views for now, for:16:44
yvlskillset -> course linking16:44
yvlnode-node linking16:45
yvlnode-layer linking16:45
yvlah, layer vocabulary16:45
yvland that's pretty much that16:45
yvlthe nearest plan could be:16:46
yvlI push changes (after I look at project's skill retirement), replaceafill starts working on project gradebook16:46
yvlsee what's missing, add that16:46
yvlI finish validation and do a basic importer16:46
yvladd the course->skillset linking16:47
yvlthat should be enough to start working on skill gradebook16:47
yvladd the rest of the stuff16:47
th1a_This is just the skill side of the model right?  Not the scores?16:47
yvlalso the scores16:48
yvlskills are requirements, we'll be using the default evaluation storage mechanism16:48
th1a_OK, so I should consider this both your model tasks.16:48
th1a_That has a big impact on whether you're ahead or behind.  ;-)16:48
yvltrue :)16:49
th1a_OK.  It sounds good.16:49
yvlI think we'll have to communicate closely with replaceafill16:49
* replaceafill too16:49
th1a_Are there any potential minefields in sight?16:49
replaceafilli'll definitely have questions :)16:49
yvlbecause how exactly what should be added where may not be clear from first glance16:49
yvlthe little things16:50
yvllike - skillsets always go in per-year container16:50
yvlone project worksheet == one skillset copy16:50
yvlthere's also a minor thing on direct and indirect skill equivalency16:50
yvla === b16:51
yvlb === c16:51
yvla == c16:51
yvlwell, I'll email replaceafill when I push16:52
yvlhe should take a look, and we should do a hangout after that16:52
* yvl done16:53
th1a_OK.  Thanks yvl.16:53
menesisdon't think I have done anything for schooltool last few days16:54
menesisother than finishing the release16:54
menesisand splitting phillipinnes stories into detailed tasks16:54
th1a_How's that looking?16:55
menesisnot finished yet16:59
*** ignas has joined #schooltool16:59
menesiswas doing that with Laurynas but both were busy with other tasks so little time17:00
menesiswill have to start doing instead of planning asap17:00
th1a_Well, if there is anything I can do to help push things forward, don't hesitate to get in touch.17:01
*** jelkner has joined #schooltool17:02
menesisnot at the moment17:02
th1a_Thanks menesis.17:02
replaceafilllast week i helped Carlos Ramires with his demo importing problems17:03
replaceafillhe finally was able to import his data17:03
replaceafilli also made the journal to use the new gradebook design/functionality17:03
replaceafilli still have to fix some popup options, but i think it's done17:04
replaceafilli experimented a little with qunit tests for the gradebook js17:04
replaceafilli like qunit :)17:04
replaceafilli added a check to the date data converter17:05
replaceafillfor years lower than 190017:05
replaceafillwhich fail using the strftime method17:05
replaceafillthanks to yvl for reviewing that change17:05
replaceafillon saturday a bug was reported on non-ascii characters17:06
replaceafillin the delete view of a section17:06
replaceafillso i fixed that one too17:06
replaceafillwhich made me wonder a few things17:06
replaceafillcould you take a look at that one please ^?17:07
th1a_Is there a certain pattern of unicode tests we should be applying everywhere?17:07
replaceafillSEND_KEYS_TO_ELEMENT in webdriver doesn't allow non-ascii strings. Should we encode in the browser extensions or in WebElement.type method or in the tests?17:07
th1a_That could be pointed out to all the developers...17:08
replaceafillth1a_ actually this is the first non-ascii selenium test we have i think17:08
replaceafillanother thing i don't like:17:08
replaceafillI can't use non-ascii strings like Año in print statement in a ftest/stest, instead of A...o17:09
th1a_Or should we just make a point of using a non-ASCII character in all our test strings.17:09
th1a_These are super-common bugs, so we have to be able to test for them.17:09
replaceafilli agreee17:09
replaceafillyvl, i'm not sure if my last complain can be fixed using encoding of the stest setup17:10
replaceafilli remember i tried that some time ago, but didnt work17:10
yvlselenium should be able to process this17:10
yvlI'll check that17:11
replaceafillthose two things would be a great help for this kind of test17:11
replaceafillthis bug reminded me of this other one:17:11
replaceafillCrash trying to book a resource with unicode characters17:12
replaceafilli checked and it's a similar problem with url query parameters17:12
replaceafilli didnt fix it, but i know how to do it now17:12
th1a_Feel free to fix it.17:13
replaceafillin other things, i also check a fix for the "Are you sure you want to leave this page?" dialog for the gradebooks17:13
replaceafillbig mess with firefox and chrome17:13
replaceafilldifferent functionality :(17:13
replaceafillbut thanks to th1a_ we have a way to solve it )17:13
replaceafillso i have to do that too17:13
replaceafillyvl, quick question17:14
replaceafilli was writing a test for the gradebook17:14
replaceafilland i wanted to print the "current" state17:14
replaceafillwithout submitting the changes17:14
replaceafillfor example i take one cell and i change its value from 100 to 9017:14
replaceafilli wanted to print right there17:14
replaceafilland get the 9017:14
replaceafillthe problem is that my print helper searches for the @value attribute17:15
replaceafillwhich at that point is not set for the input17:15
replaceafilli'm using queryHTML from the analyze module for the helper17:15
replaceafillthen i thought i should use selenium for the helper17:15
replaceafill(as we discussed some time ago)17:16
replaceafillbut then i found out that selenium has a similar problem17:16
replaceafillif the attribute is not set in the element, selenium will return undefined17:16
replaceafillor None17:16
replaceafillso i have two options:17:16
replaceafill1. not print the current state of the gradebook, which sucks17:16
replaceafillbecause i wanted to test that17:16
replaceafillfor the validation of the cell17:17
replaceafill(valid, error, extracredit)17:17
replaceafillor 2.17:17
aelkneri have a similar problem to report, btw17:17
replaceafilluse execute_script somehow to get the value of the input through jquery17:17
replaceafilljquery does return the value i want17:17
replaceafillbut it feels messy17:18
replaceafill(i didn't try it though)17:18
yvlcan you access the value you want...17:18
yvlwithout looking at html from selenium17:18
yvli.e. looking up the element and printing element.value or whatever17:19
replaceafillah ok17:19
replaceafilli'll move the helper to use selenium then17:19
replaceafilland report back17:19
* yvl is not 100% sure that will help17:19
replaceafilli think getValue as they call it in the mailing lists searches for the @value17:20
replaceafillbut i didnt test it17:20
replaceafillbut i will :)17:20
replaceafillok, last thing17:20
replaceafillth1a, do you want this fixed in trunk, like this week?17:20
replaceafillexport data as excel or csv17:20
th1a_If it fits in reasonably for you.17:21
replaceafilli guess i'm done17:22
th1a_It sounds simple and very little risk so why not?17:22
th1a_Thanks replaceafill.17:22
aelknerok, i finished the custom login task17:23
aelknerthat involved a tricky change to the app package to split selenium tests between old and flourish layers17:23
aelkneryvl, thanks for the help with the __name__ problem17:23
aelkneri created tests for the server tabs view17:24
aelknerth1a, and in adding the missing breadcrumbs that you caught17:25
*** th1a has quit IRC17:25
aelkneri realized that we should have a tests for the breadcrumbs in our selenium tests17:25
aelknerthe first time a view is brought up, we do that everywhere17:26
aelknerwe could, i mean17:26
aelkneri found a problem similar to what replaceafill was talking about with testing values17:26
aelkneryvl, if you run my server_tabs.txt test, you notice that although the checkboxes changed17:27
yvlreplaceafill, problem with sending unicode keys confirmed, found a fix, looking how to apply it17:27
replaceafillyvl, excellent!17:27
aelknerthe radio buttons didn't have the values changed17:27
aelkneryvl, so i didn't bother to put tests for the change to the radio button17:29
aelknerso as not to have false tests in there17:29
aelknerthe fact is, the expected changes happen upon submit17:29
yvloh, so if I run it, I can't see what's wrong?17:29
aelknerno, you'd have to add a tests for whatever you think should happen17:29
aelknerlook at the first time i click the radio button17:30
th1a_aelkner:  You're referring to testing the dynamic changes in the form, not it initial state, right?17:30
aelknerbtw, it's 'checked' not 'value' that fails to change17:31
aelkneryet disabled changes ok in the checkbox controls17:31
yvlsorry aelkner, I won't bother to figure out what tests to put in to reproduce whatever is wrong17:31
aelkneras my test shows17:31
aelkneroh geez, fine, i'll put one in, merge to trunk, etc.17:32
aelkneri can do that after i'm done chatting here17:32
yvldid I say you can merge that to trunk?17:32
aelknerok, you can merge from my banch17:33
yvllook, if it works for you, it works for me :)17:33
th1a_If only we had some way of distributing changes in a non-centralized way.17:33
aelknerif only17:33
yvland I have your latest branch checked out already, aelkner17:33
th1a_Perhaps I could suggest that Mark fund the development of such a tool.17:33
th1a_Or maybe Linus Torvalds would like to take on the problem.17:33
aelknerall sarcasm aside, it's not the tool, but the time spent communicating what braches we're talking about, etc.17:34
yvlyou could, th1a_, but you should ask Matt Mackall first17:34
aelknerassuming trunk is our central commuicating point saves time in my opinion17:34
aelknerbut, if you want to merge with y branch, then i'll show you that later17:35
yvlit's like 8 seconds17:35
aelkneryeah, np17:35
* yvl has a bound branch, because you asked to look at you changes few days before17:35
aelknerok, so i fixed the student detail report to have the same period sort as the absences by day report17:37
aelkneri notived that there hasn't been a commit by menesis to trunk since i last added the other period fix17:37
aelknerso i fell it's safe to assume that they will be released together17:37
aelkneranyway, the importer now has automatic teachers and students being added to their groups17:38
aelkneras we discussed on friday17:39
aelkneri added a special xls test file  to go along with the txt test17:40
aelknerit's a smaller version of test_data.xls for reproducing the test17:40
replaceafillaelkner, i wonder what will happen if i try to import only the persons sheet17:40
replaceafillwith no school year17:40
aelknertry it :)17:40
aelknerit shouldn't cause a problem17:41
aelknerhere's the thing17:41
aelkneras we discussed on friday, the import does not check for limit groups17:41
aelknerit can't anyway, so that was the right decision17:41
aelknerjust joking actually, but that omission is logical considering the year independence of the person object17:42
aelknerreally we can't control whether a person's membership to a group like teachers17:42
aelknervaries from year to year17:42
aelknerand its effect n whether a person can have a demo value for that field17:43
aelknerand the person edit form17:43
th1a_Overall people having attributes they shouldn't is not really a problem.17:43
aelknerthat uses the current term17:43
aelknerso that's how that works17:43
replaceafillwhat should be the status of this one?17:44
th1a_The purpose for this is hiding irrelevant things, but if values are assigned, then they aren't necessarily irrelevant for that person anyhow.17:44
aelknerreplaceafill, we're going to have to work together after the meeting to continue to define17:47
aelknerthe problems you are having with importing as we build a list of tasks for me in that area17:47
replaceafillproblems i'm having with importing?17:48
aelknerthat bug report has a long thread of back and forth that involves many issues17:48
*** th1a_ is now known as th1a17:48
aelknersome are ust user error, some perhaps not17:48
replaceafillaelkner, it was just a communication thread i think17:49
th1aThe bug is just the limited demographic fields one.17:49
aelknerwe discussed some of the language dependency issues that are not always thus17:49
aelkneroh, so that bug doesn't have the unicode issues?17:49
replaceafillaelkner, that would be my only problem with the importer17:49
replaceafillaelkner, no17:49
replaceafillthat's the other one17:49
aelknersorry, because the other one does17:49
replaceafillthe one at the end17:49
aelkneractually, i'm not sure yet if its unicode or user error yet17:50
yvloh, btw, replaceafill - unicode typing fixed committed17:50
replaceafillyvl, thanks17:50
replaceafillyvl, will i be able to do:17:50
replaceafill>>> print foo17:51
replaceafillor is it a fix for the type(...) method_17:51
replaceafillactually i should just look :D17:51
th1aaelkner:  I don't know that replaceafill actually spent any time on that but.17:52
aelkneri thought that was the one that caused him to bring p the language issue17:52
replaceafilli just realized that the timetable importer uses english days of the week17:52
replaceafillaelkner, it is17:52
aelknerthose labels?17:52
replaceafillbut again, it's a broader issue i think17:53
aelknerthe ones that we are changing to 1-717:53
replaceafillit's like17:53
replaceafill"make the importer to match the language in the app forms"17:53
replaceafillif the person add form shows a label17:54
replaceafillthat should match the importer column17:54
replaceafill(or something like that)17:54
aelknersample_data.xls is in english, yes17:54
replaceafillbut never mind, these are crazy ideas from mine17:54
aelknerwe could create one for each language17:54
replaceafillaelkner, to me it should be generated17:55
th1aNot now.17:55
aelknerbut did you try exporting in portugese17:55
th1aHere's what we are doing.17:55
* replaceafill shuts up17:55
aelknerbecause there, if it gives english, that would be a problem17:55
th1aI want the importer code to be not dependent on English.17:55
aelknerreplaceafill, don't shut up17:55
replaceafillyvl, did you push to trunk?17:55
aelknerhow come no one listens to me ;(17:56
replaceafillaelkner, i'm listening to you17:56
th1aI'm less concerned with automatically generating a translated template.17:56
aelknerthe importer does not check labels17:56
th1aThat is a good idea, but not so important.17:56
th1aThe import code cannot be dependent on any specific cells being in english.17:56
aelknerok, and when did it ever do this thing that you are talking about?17:57
aelknerok, perhaps it never did and you are just saying, don't have bugs or something17:57
aelkneranyway, replaceafill, it's no problem if you find a bug17:58
th1aWhat is the nature of ?17:58
aelkneri just haven't seen one yet pointing to language dependence17:58
replaceafillaelkner, export an xls with a timetable in a language that is not english17:58
th1aWell, I am asking you to check, which you should be able to do while you're reviewing the importers.17:58
replaceafilland try to import it back17:58
th1aDoes it break in demographics?17:59
replaceafillth1a, the nature?17:59
replaceafillno, it breaks the timetable importer17:59
aelknerreplaceafill, ok, it's an export bug17:59
th1aWe are exporting i18n-ed days and we don't know how to import them?17:59
replaceafilla user would expect the exported xls in their language imo17:59
replaceafillth1a, correct18:00
replaceafillmaybe i'm not going to import it :)18:00
replaceafillth1a, i don't understand your question about the nature of the bug?18:01
th1aSo we're talking about days of the week only in this case.18:01
th1aI'm just trying to figure out what kind of bug this is.18:01
th1aWhich we are discussing.18:01
aelknerwe may have a bug in the exporter, or we may have a unicode interpretation problem upon reimporting18:01
th1aThe bug is we're using DIFFERENT WORDS.18:02
th1aIt would still fail in ASCII.18:02
th1aIf in my country "Monday" is translated as "ElknerDay" it will break our import/export cycle.18:02
replaceafillbecause the importer expects Monday18:03
th1aHow are days stored internally?  Numbers? 0-6?18:03
aelknerreplaceafill, have you found the code where it tests for 'Monday'?18:03
replaceafill['Monday', 'Tuesday'...]18:03
replaceafillaelkner, timetable importer18:03
aelknerah, there it is, titmetables!18:04
replaceafillday_title = self.dows[int(day_id)]18:04
th1aAnyhow we're now back to the solution, which has been clear all along anyhow.18:04
th1aThere needs to be a row in the importer with numbers identifying the days, which is what we'll actually look at in the importer.18:05
th1aRow in the sheet.18:05
th1aWe'll let the Lithuanians go home...18:05
th1aThanks guys!18:05
th1aSee you Wednesday.18:05
* th1a drops the bag of gravel.18:05
replaceafillthanks everybody18:05
aelknerthanks guys, until we meet on wednesday...18:06
aelkneralthough, yvl, i could use your help a second here with timetables18:07
yvlthanks guys18:07
aelknerthe importer has to create a DayTemplate object and pass it a title18:07
* replaceafill hates when bzr merge hides some revisions :(18:07
aelknerit's about process, too, not just tools :)18:08
yvlsorry for that guys18:08
yvlaelkner, please go ahead18:09
* yvl has to run soon18:09
th1aI don't think we need yvl to sort this import question out.18:09
aelknerwell, i'm creating DayTemplate objects, and the question is18:09
aelknershould i be using _('Monday'), _('Tuesday'), etc. for that18:10
aelkneror 1-7 as we discussed as our change for the import labels18:10
th1aSo in the data model they are actually in English?18:10
aelknerthat's why i;m asking yvl, because he is more fluent in timetables18:11
th1aI guess that is more explicit than having to remember which day Day 0 is.18:11
yvlare you adding them to WeekDayTemplates?18:11
aelknerdoes our UI present day names to the user for approval?18:12
aelknertimetable.periods.templates[name] = day18:12
aelknerday being a DayTemplate18:12
yvlthen again it don't matter, AFAIK18:12
yvlname matters a lot18:13
aelknerthat's a nameschooser name18:13
aelknerit would be easier if we looked at the same lines of code, 480 in importer.py18:13
aelknerin the School Timetables sheet, there are two tables18:14
th1aI think there is some confusion about whether we're talking about "Monday" as the day of the week, the real, calendar defined, objective, universal day.18:14
th1aOr the school's day in the timetable.18:15
yvlday title can be anything18:15
yvlif you make it re-importable, I'm fine with it18:15
aelknerwell, i have to choose it18:15
aelknerand when i do so, i fix it as something that the user neds to know to match to18:15
yvlwell, then choose it18:16
aelkneri'm not sure how we could make that process bullet-proof18:16
yvlsorry, gotta go18:16
aelknerthe user can change the language on an instance between export/re-import18:17
yvlthe answer is - you can use any18:17
aelkneryvl, ok, thanks for you input18:17
yvlgood luck!18:17
yvlsee you Wed18:17
*** yvl has quit IRC18:17
aelknerreplaceafill, what do you think?18:18
aelknerdo i need to change the class attribute, dows, to have the translatable strings18:19
aelknerand just hope that the user doesn't change the language between export and reimport?18:19
replaceafillagain, you'd be relying on translations18:19
replaceafillto me we should use codes18:20
aelknerah yes, we could just use 1-718:20
aelknerok, i assumed because we had the days of the week that the users expected that18:21
replaceafillthat's what the importer expects :)18:21
aelknerwell, the importer expects the user wants the days of the week18:21
aelknerbut i guess if they are happy seeing 1-7 in the ui, then all is well18:22
replaceafillwhy do we have to match the ui to what the importer expects?18:22
th1aIs this error triggered in the term sheet?18:22
th1aIs it where you mark the weekends?18:22
aelknerthe importer creates objects that are displayed in the ui18:22
replaceafilli can see Monday in the ui and use code 1 in the importer ,can't i?18:22
replaceafillth1a, no, the timetable sheet18:23
th1aIs it only for certain timetable models?18:24
replaceafillfor week of days18:24
replaceafilldays of week :D18:24
th1aaelkner:  Do you understand the significance of that?18:26
th1aSo basically, we should have the translated day of the week above a number corresponding to that day, and just ignore the text.18:27
aelknerso, are you looking at the School Timetables sheet, at the top18:27
aelknerthere's a period days18:28
th1aOr perhaps they need to be side by side.18:28
aelknerwhich in the example i'm looking at, is set to week_days18:28
aelknerthat's a model factory18:28
th1aI don't have a days of week timetable example handy.18:29
aelkner    day_templates = (18:29
aelkner        ('calendar_days', CalendarDayTemplates),18:29
aelkner        ('week_days', WeekDayTemplates),18:29
aelkner        ('school_days', SchoolDayTemplates),18:29
aelkner        )18:29
aelknerthat seems to be the list of available factories18:29
aelknerif week_days is chosen, as in our case18:31
aelknerthen it needs to look up the weekday in a list which i have hard-coded in english18:32
aelknerbecause i don't know where else to get the language independent names18:32
th1aSee, this is where you need to add a column that gets a number representing the day.18:33
th1aBeside it can be the translated value for that day.18:33
th1aOr nothing or nonsense.18:33
aelknerthis will take some time looking into as it involves grokking the timetable data model again18:35
th1aIt is pretty superficial as long as you remember that it is only for timetables that correspond directly to the days of the week.18:35
th1aIt is a direct mapping.18:36
th1aSunday:0, Monday: 1, etc.18:36
th1a(or whatever)18:36
aelknerif the chosen model is week_days, then yes, if it's school_days, then not so much18:37
aelknerour samples have always been one fits all18:38
th1aIt is completely irrelevant to the other models.18:38
aelknerok, but i can't discuss it at the same time as grokking the data model, so...18:39
th1aUnless this bug has nothing to do with the names of the days of the week.18:39
aelknerit does, i think, but i can't say for sure yet18:39
aelkneri think that maybe we could have a new column as you suggest18:39
aelknerin our one fits all sheet18:39
aelknerand just expect the user to know what our example sheet suggests when the model changes18:40
aelknerwe don't have too many examples of users using various models and the importer18:40
aelknerso we don't have a lot of experience with this situation18:40
th1aTo be clear, we don't recommend users create or edit timetables in the importer/exporter.18:41
th1aThat is officially unsupported.18:41
th1aAs far as I'm concerned.18:41
aelknerperhaps that's why we don't have as many user reports about it :)18:41
aelknerbut i can look into supporting it18:42
th1aWe will never support that.18:42
th1aThe fix is, "Use the web UI."18:42
th1aWe just have to have import/export cycles that work for School Timetables.18:43
aelknerwell, if we make the change to include the new column, then we are supporting it18:43
aelkneri'll just have to look at all this stuff and see what the data model is expecting18:43
aelknerthen perhaps i can find a way to add a column that makes sense for all three period day models18:45
th1aWe don't need to add a column for all three, do we?18:45
aelknerthe sheets are one fits all18:46
aelknerthe cells are searched by row, column numbers18:46
aelknerso if we insert a column, we have to change the code where the cells after that column are bumped up one18:47
aelknerand that needs to fit for those period day models that don't have both a day of week title and a day number18:47
th1aOK, so give them a blank column then.18:47
aelknerbut the first column can't be blank, so i guess the day number goes there18:48
aelknerthen the second column, day name, is filled in for week_days models only18:48
aelkneri'm just spit-balling here, because i still haven't looked at the data model for period day models18:49
th1aOne problem is you keep making these statements about exploring the data model which just sound very time consuming.18:49
aelknerthey are, believe me18:49
th1aOK, here's the fix then.18:50
th1aWe are done with this.18:50
replaceafillth1a, one comment18:50
th1aChange the exporter to be in English.18:50
replaceafillif it will take too much time from aelkner18:51
replaceafillwe still have the Importe i18n task for later18:51
replaceafillit's scheduled for june i think18:51
* replaceafill goes to see the planner18:51
replaceafillyes, two weeks for that on june18:52
th1aOK, you're off the bug aelkner.18:52
th1aJust focus on the cando import issues.18:52
aelknercan we do something about these bug titles18:53
aelknerwe're talking about timetables this whole time, but the title doesn't mention it18:53
aelknerit's this huge thread is what i'm saying18:53
replaceafill*that* bug was not about timetable issues18:53
replaceafillit was about limited demos not appearing after importing18:53
aelknercomment #13 is18:53
replaceafill1 of 13 :/18:54
aelknerthat's what started this whole discussion18:54
replaceafillyou should focus on bug 901267 i think18:54
aelkneri'm just saying, we could use a separate bug for just the timetable issue18:54
replaceafillwe have it aelkner18:54
replaceafillit's 90126718:54
replaceafillcomment 14 on the bug you mentiond18:54
aelknerright, so th1a, when you say i'm off those, could you update the bugs to reflect that?18:56
aelknerchange of status, assignment, so i know what to think when i come back to this18:57
aelknerin the meantime, we can also talk about cando import18:57
th1ayes aelkner18:57
aelknerwe started talking about that friday, and then it got shifted to the discussion we just finished18:57
aelknerwe did talk about the Students sheet18:58
aelknerand i looked that discussion over last night to prepare myself to continue with you18:59
aelkneri'm wondering why we need it again, if we just put the State ID as a demo18:59
aelknerin the Persons sheet18:59
th1aThis is not only for Virginia.19:00
aelknerand the automatic inclusion of students in the students group during section iport19:00
aelknerthat i just merged to trunk, shouldn't that cover that issue?19:00
th1aBasically it is exactly like the persons import except it puts each person in the students group.19:00
th1aIt should be pretty easy to implement.19:01
aelknerwe discussed having the school year at the top19:01
aelknerand otherwise, you're saying the same columns as the Persons sheets19:01
aelknerbut just make sure the user is added to the students group for that year19:02
aelkneri wonder if i can code the Students sheet importer to call the Persons sheet importer19:02
aelknerso that i woudn't need to clone the whole class19:02
aelkneri think i can19:03
th1aor subclass19:03
aelkneranyway, shall we have a Teachers sheet as well19:03
aelkneryes, a subclass could do the trick19:03
aelknerok, that task is well defined now19:04
aelkneranything else for cando that you have in mind for the importer?19:04
th1aYou have to go through the CanDo instructions and make them work for SchoolTool.19:04
aelknerfor instance, they have their own section enrollment importer19:04
th1aI don't have anything in mind personally.19:05
aelknerbut they should be using the general one19:05
th1aWell, what's the difference?19:05
aelknercheck this url out:19:06
aelknerso we can look at the same thing19:06
th1aIn general retaining their exact format is not the requirement.19:06
th1aWhere are the students listed?19:09
th1aSo there is kind of a header row and then a row for each student?19:10
th1aI actually kind of like this version.19:10
aelknerit's a flat table for all terms19:11
aelknersome times these are convenient19:11
aelknerthey are ideal for sql import, for instance19:11
aelknerour import is based on the export unfortunately19:11
th1aWe can keep both around a while.19:12
aelkneror fortnately, depending on how you look at it19:12
aelknerwe could add a new one19:12
th1aStop the old export but keep the import code around.19:12
aelknerare you making an important decision about removing the export feature?19:13
aelkneris that what you mean about stopping the export19:13
aelknerremoving the view class and menu item?19:13
aelkneror are you saing, don't export the new sheets19:14
th1aI mean in the future exporting in the new format.19:14
aelknercould you restate it more precisely please?19:14
th1aThe official format for importing and exporting in the October release will be a new one based on the CanDo version (but not necessarily exactly).19:15
th1aBut we will keep the old importer so we can still import older spreadsheets.19:15
aelknerwell, i'm sure you're not suggesting creating two versions of the importer view class19:16
aelkneri mean, we have the one view, and it has to be ale to import older spreadhetts19:16
aelknerbut also handle the new ones19:17
aelknerStudents, Teachers, All Sections19:17
aelknerAll Sections is my suggestion for the one that works like cando's example19:17
aelknerour sections sheets are for fixed year/term19:18
aelkneractally, the cando version has no year in it19:18
aelknerso i guess that's because they assume the current year19:18
th1aWell, yes, we have to improve them.19:19
th1aMake it more SchoolTool compliant.19:19
th1aBut I think the format might be better.19:19
aelknerwe could have the year at the top, as with Students and Teachers19:19
th1aI've never like the current one.19:19
th1aThis one seems clearer.19:19
th1aOr per row.19:19
aelkneryeah, it looks messy19:19
th1aThat is probably better -- per row.19:19
aelkneryear per row?19:20
th1aSection 124 | Spring 2004 | 2004-2005 | etc19:20
aelknerso include a year column19:21
aelknerthen it would mean all sections for all years19:21
aelknertoo bad we couldn't do that for Students and Teachers19:21
aelknerwhere we need to ask them to put a year at the top19:21
th1aThose are strictly convenience views.19:21
th1aWe probably won't even export the student and teacher sheets -- with values in them.19:22
aelkneri'm just wondering, if we auto-assign in the section import, why we even need the Students and Teachers sheets19:22
aelknersorry if i ask that again, but i'm just trying to make sure19:23
th1aWe don't strictly speaking.19:23
th1aActually, we do because people may manually set up the actual sections.19:23
aelknerah, ok19:24
aelknerand it's just for convenience anyway, as you said19:24
aelknerand we won't export it19:24
aelknerbut for the sections, we can export that one19:24
th1aWe must.19:24
aelknerok, i got what i need now, thanks19:25
th1aThanks aelkner.19:26
th1aJust to be clear -- this should be using the attributes we've got for SchoolTool sections, not the CanDo meta-data.19:28
th1a(if there is a difference)19:29
aelkneryes, they include course title when they don't need to19:31
aelknersince we have the Courses sheet19:32
aelknerand the __name__, called ID, has to match the All Sections sheet19:32
aelkneryou know, there are various degrees of how far you want to go with flattening19:33
aelknerin theory, we could have a REALLY flat table which defines years, terms, courses, and sections19:33
aelknerand the first row for a given year has the ID and title, and all rows after that19:34
aelknerare blank in those cells until the year changes19:34
aelknerand then, within the year, the Term and its title can be defined19:34
aelknerand then with that, the sections19:34
aelknerthough the courses would be problematic in that case19:35
aelknerbecause they do not fall within a term19:35
aelknerthe basic ting is, we assume the user knows to use the IDs19:35
aelknerthat we must have to match up with our __name__ attributes19:36
aelkneractually, it can't be completely flat and completely defining at the same time19:37
aelknerit's a tricky set of data19:37
aelknerbut so you think i'm onto the right idea when i talk about blank cells19:38
aelknerand the almost tree-like nature of flat data that has a clear heirarchy?19:38
* th1a looks back at xchat...19:38
aelknerthe more i type, the more i think about this, we can't have courses defined in the same sheet19:39
th1aI think the flatness they have is about right.19:39
aelknerthey have course, description as first two cols19:40
aelknerthe course id for them is a state code, i think19:40
th1aLet's assume we're still going to have a separate course sheet.19:40
aelknerweird, i'm looking at the import code in cando.virginia, and it doesn't match their example xls19:42
aelknerit's slightly off19:42
aelknersorry, maybe not19:43
aelkneruh, maybe so19:44
th1aJust remember that we're trying to preserve their overall process not necessarily the specific fields, etc.19:44
aelkneri'll need to talk to dwelsh about their process i think19:45
th1aUh... proceed with caution because he's going to have to change a bit more than he wants to.19:47
replaceafillth1a, quick question19:47
th1aTry to focus on getting him to explain the current process more than what changes we might need to make.19:47
replaceafillis that a valid issue?19:48
replaceafillif it is, we have it in the gradebook too19:48
replaceafillstudent names open popups, not their home pages19:48
th1aWe could add it as a menu item.19:51
replaceafillScore, Report, Visit?19:51
replaceafillin the old days we used "View" for that iirc19:52
th1aI think View is ok.19:53
replaceafilli'll add it then19:53
th1aIdeally maybe it should be in a dialog.19:53
th1aOr not.19:53
th1aLet's assume not for now.19:53
th1aCan we open it in a different window/tab?19:53
th1aLet's do that.19:53
th1aI don't want it to take you away from the gradebook.19:56
*** paulproteus has quit IRC20:12
*** paulproteus has joined #schooltool20:14
*** menesis has quit IRC20:27
replaceafillgreat, we can use non-ascii in stests now :)21:18
*** jelkner has quit IRC21:22
*** th1a_ has joined #schooltool21:55
*** menesis has joined #schooltool21:58
*** menesis has quit IRC23:33

Generated by 2.15.1 by Marius Gedminas - find it at!