IRC log of #schooltool for Monday, 2015-11-23

*** khildin has quit IRC00:45
*** khildin has joined #schooltool09:24
*** yvl has joined #schooltool10:34
*** khildin has quit IRC14:41
*** replaceafill has joined #schooltool17:14
*** khildin has joined #schooltool17:25
*** th1a has joined #schooltool17:33
th1ahi replaceafill.17:35
replaceafillhey th1a17:35
th1aSorry my shower ran long...17:35
th1aOK, catching up with the Integration Update thread.17:37
th1aFor #1 -- this is not done by section, right, so we don't really have any idea of who is doing the attendance for a given student?17:38
replaceafillthe user of the app17:38
replaceafillis recorded in the registry17:39
replaceafillsorry, register17:39
th1aOK, sorry.17:39
th1aSo this isn't really a problem per se, just a change?17:39
replaceafilli was just expecting the teacher17:39
th1aOK, so are there any outstanding *problems*?  Just a few changes we understand?17:40
replaceafillwell, we just need to decide what to do about the registers with empty "completed at" attributes17:41
replaceafilldo you understand that problem?17:41
th1aIs this if the teacher hasn't done all the students in the group?17:42
replaceafillif the teacher hasn't finished the register17:43
th1aWhich *shouldn't* happen according to Steph?17:43
replaceafillthe integration looks for X days in the past17:43
replaceafillthe blueprint specifies 3 iirc17:43
replaceafillbut that's just a setting17:44
th1aSo what are the choices?17:44
replaceafillright, according to Steph all the registers should be complete17:45
replaceafillbut i found a few coming from the production data that had this problem17:45
replaceafillso, choices17:45
replaceafilli propose just ignore incomplete registers17:46
th1aIs a "register" the group or a single student mark?17:46
replaceafilla group17:46
replaceafillsingle student mark is called student instance in journey land17:47
replaceafilli'm not sure what's best for them17:47
replaceafillto see just a few students with attendance for a day17:48
replaceafillor none at all17:48
replaceafilli mean, what would make them react17:48
th1aYou don't want to lose what data you do have.17:48
replaceafillit's in journey, right?17:49
th1aWell, I would certainly expect to sync what data there is.17:49
th1aAnd they can still tell the difference in ST between no score and present, right?17:50
replaceafillthe integration doesn't overwrite ST data btw17:50
replaceafillif there's already a score in ST, journey scores won't overwrite that17:51
th1aSo at this point we always check the past three days?17:51
th1aSo if we bring across partial data, they have three days to get the rest in there and then we ignore it?17:51
th1aThat seems ok.17:52
th1aI think that's fine.17:53
replaceafillok, so we'll just ignore incomplete "student instances"17:53
th1aI would not essentially "time out" a whole group of students because one instance was incomplete.17:53
replaceafilland use the register to get the date and evaluator17:54
replaceafilli was even thinking of recording registers + their completed at attribute in ST17:55
replaceafillto know what had been retrieved already17:55
th1aYes, I was too.17:55
replaceafilland skip that in following syncs17:55
th1aThat certainly makes sense.17:55
replaceafilli think that wouldn't be hard to add17:56
replaceafilland would improve transfer times17:56
replaceafilljust a container with objects (id and completed_at)17:56
replaceafillcatalog probably for looks up17:57
replaceafilli'll work on that17:57
replaceafillsince i have learned not to promise demos anymore...17:57
replaceafilli'm aiming for a wednesday demo17:57
replaceafilland i made other changes too17:57
replaceafill(i explained them in my mail)17:57
replaceafilli guess the most important one is the "attendance group" attribute for streams17:58
replaceafilli've put that in the rising side17:58
replaceafillleaving core streams untouched17:58
replaceafillbut it seems like Fran now wants that across the other projects :S17:59
replaceafillat least that's what i understand from this: "WE SHOULD MAP THIS THE SAME WAY WHEN WE TRANSFER ACROSS TO OTHER VERSION TOO"17:59
th1aOK, well, we can charge for it.17:59
th1aShe's pre-paid, so there is little stopping us from saying, "OK, we're counting 4 hours for that," or whatever.18:00
replaceafillah ok18:00
replaceafilli guess that's it18:00
replaceafilli have enough to continue18:01
replaceafilli'll add the container for saving sync'ed registers18:01
th1aI don't know if we should just leave something like this more open ended next time.18:02
th1aSince we *really* don't know how many surprises there will be.18:02
replaceafillthis project has had many more surprises than the canvas one18:02
th1aAnyhow, we can cross that bridge when we get to it.18:02
th1aI think the warning sign is something thrown together from a generic toolkit compared to a mature educational application.18:03
th1aThere will be more idiosyncracies in how it is put together.  Corners cut in the model.18:03
th1aOK.  Thanks replaceafill.18:04
replaceafilland probably just a few users18:04
replaceafillvs (i pressume) a lot of canvas users18:04
th1aI think they're the only users.18:04
th1aSingle purpose databases are going to have no more complete or systematic a model than is absolutely necessary.18:05
*** The_Dylan has joined #schooltool20:16
*** The_Dylan has quit IRC21:40
*** replaceafill has quit IRC23:46
*** khildin has quit IRC23:48

Generated by 2.15.1 by Marius Gedminas - find it at!