| *** jinty has joined #schooltool | 00:22 | |
| *** jinty has quit IRC | 00:32 | |
| *** jinty has joined #schooltool | 00:32 | |
| *** didymo has joined #schooltool | 01:09 | |
| *** jinty has quit IRC | 01:39 | |
| *** th1a has joined #schooltool | 07:49 | |
| *** th1a has left #schooltool | 07:50 | |
| *** hoffman has joined #schooltool | 07:51 | |
| *** Aiste has quit IRC | 08:57 | |
| *** Aiste has joined #schooltool | 09:51 | |
| *** kitblake has joined #schooltool | 10:29 | |
| *** thisfred has joined #schooltool | 10:58 | |
| *** kitblake has quit IRC | 12:08 | |
| *** alga has joined #SchoolTool | 13:10 | |
| *** ignas has joined #schooltool | 13:26 | |
| *** mgedmin has joined #schooltool | 13:37 | |
| *** didymo has quit IRC | 13:49 | |
| *** mgedmin has quit IRC | 13:53 | |
| *** mgedmin has joined #schooltool | 13:55 | |
| *** mgedmin has quit IRC | 15:21 | |
| hoffman | Hello all. | 16:30 |
|---|---|---|
| hoffman | Good morning/afternoon. | 16:30 |
| *** SteveA has joined #schooltool | 16:30 | |
| srichter | good morning | 16:31 |
| *** mgedmin has joined #schooltool | 16:32 | |
| mgedmin | hi | 16:32 |
| hoffman | Let's get started. | 16:33 |
| hoffman | I'll go first. | 16:33 |
| hoffman | I didn't do much actual *programming* last week, but I did work some things out with how requirements will work with the gradebook. | 16:34 |
| hoffman | That is, it seems to make sense to use the requirement package Stephan and Paul started | 16:34 |
| hoffman | to assign grades to a student for a section. | 16:35 |
| hoffman | And also to evaluate individual assignments. | 16:35 |
| srichter | ok | 16:35 |
| hoffman | I think the requirements package is going to be a great foundation. | 16:35 |
| hoffman | I'm going to work on creating an "activity" or assignment package for the gradebook this week. | 16:36 |
| srichter | cool | 16:37 |
| srichter | note that I would call it schooltool.gradebook | 16:37 |
| hoffman | Well, I'm wondering if I should make schooltool.activity separate from schooltool.gradebook | 16:38 |
| srichter | I personally think that it would be overkill, but then I do not know what you intend to do with activities later | 16:39 |
| hoffman | activity is a pretty foundational object... | 16:39 |
| hoffman | Perhaps I'll keep them together for now. | 16:40 |
| srichter | yeah, we can separate later if we need to | 16:40 |
| hoffman | srichter: What's the status of the tutorial? | 16:40 |
| srichter | I have been stuck :-) | 16:40 |
| srichter | or :-( | 16:40 |
| * mgedmin remarks that we already use the term 'activity' for timetabling, and wonders if the two terms are the same | 16:41 | |
| srichter | I am thinking about the easiest way of doing this and staying inside the patterns | 16:41 |
| hoffman | mgedmin: What's an "activity" in timetabling? | 16:41 |
| srichter | I will finish this the week after Christmas | 16:41 |
| mgedmin | the thing that sits in a timetable slot | 16:41 |
| hoffman | Inside the pattern of doctests? | 16:41 |
| mgedmin | ITimetableActivity | 16:41 |
| srichter | this week I really have to do some research, so do not expect much availability from me | 16:41 |
| srichter | hoffman: no common Zope 3 / SchoolTool programming patterns | 16:42 |
| *** SteveA has quit IRC | 16:42 | |
| srichter | also, last week was reading period at the school, so I was massively tied up grading | 16:43 |
| srichter | (but all this is over now) | 16:43 |
| hoffman | srichter: What are the options in terms of patterns? | 16:43 |
| srichter | starting after new year's I will be available 50/50 for this and research | 16:43 |
| srichter | hoffman: as discussed we are not using a namespace, which means that we have to do some other trickery, similar/equal to the notes feature | 16:44 |
| srichter | the other option is to write a pluggable HTTP traverser | 16:44 |
| srichter | either way, I think that a lot of the simplicity of the first two writings will fade away | 16:45 |
| hoffman | But you'd prefer to use a namespace? | 16:45 |
| srichter | (now I wish I would have gone wioth namespaces) | 16:45 |
| hoffman | Are you part way into it now? | 16:45 |
| srichter | yeah, I now consider it easier, because I can use more established patterns for editing, contents and adding | 16:45 |
| srichter | no, grading stopped me from making much progress | 16:46 |
| hoffman | Right. | 16:46 |
| alga | srichter: will schooltool automate grading in the future? :-) | 16:46 |
| *** pingswept has quit IRC | 16:47 | |
| hoffman | We need to integrate some latent semantic analysis packages. | 16:47 |
| srichter | alga: I actually worked on such tools before, and you can automate a lot in science grading actually | 16:47 |
| mgedmin | and OCR | 16:47 |
| srichter | actually, you can ask students to take tests and HW online | 16:47 |
| srichter | my undergrad institution did this successfully using a very simple basic program | 16:48 |
| srichter | it can actually be more instructive, since you can see animations, models, etc. | 16:48 |
| hoffman | Anyhow, when we discussed this step in the tutorial before did mgedmin and I think namespaces seemed like overkill? | 16:48 |
| srichter | hoffman: mgedmin mainly hates the +'s | 16:48 |
| srichter | hoffman: I spoke too soon back then thinking without namespaces it would be easier | 16:49 |
| hoffman | Where do we use namespaces now? | 16:49 |
| srichter | nowhere | 16:50 |
| srichter | but Paul is using them in Cando | 16:50 |
| srichter | (after I showed him how to use them) | 16:50 |
| hoffman | Well, I suppose we shouldn't use them in the tutorial if we aren't going to use them widely in SchoolTool. | 16:51 |
| hoffman | It is a tricky question, though. | 16:51 |
| srichter | right, that was the counter argument in my head as well :-) | 16:51 |
| hoffman | srichter. You should shoot for having this done for inclusion in the snapshot on the 3rd. | 16:51 |
| srichter | ok, I can manage that | 16:52 |
| hoffman | Good. | 16:52 |
| hoffman | OK. How have things been going at chez POV? | 16:52 |
| alga | ahem | 16:53 |
| alga | We've been working on the realtime attendance form and all the backend infrastructure it needs | 16:54 |
| alga | Basically the backend is done, and the form is in the works | 16:54 |
| alga | Unfortunately, we have to suspend work on SchoolTool for the forthcoming week | 16:55 |
| alga | but we'll be | 16:55 |
| alga | erm | 16:55 |
| alga | but we'll definitely have realtime attendance working for the new year release | 16:55 |
| alga | over | 16:55 |
| hoffman | I see. | 16:56 |
| hoffman | Is there anything I can see in the web UI in the trunk yet? | 16:57 |
| mgedmin | section meeting calendar events have a link to the real-time attendance form | 16:57 |
| alga | but the form itself is not yet there | 16:57 |
| hoffman | OK. | 16:58 |
| mgedmin | well, you'll see something if you follow that link | 16:58 |
| mgedmin | but it will not be very formy | 16:58 |
| alga | Vidas has been in offline-land working on the sparklines | 16:59 |
| hoffman | Yes. Is that going to be a separate package? | 17:00 |
| alga | we cried YAGNI | 17:00 |
| hoffman | OK. | 17:00 |
| hoffman | Let's move on to discussing package structure & mgedmin's proposal. | 17:01 |
| hoffman | Does everyone have his email? | 17:01 |
| srichter | yep | 17:01 |
| hoffman | I think his proposal sounds reasonable. | 17:02 |
| srichter | I think the main evil is that most packages depend on getSchoolToolApplication | 17:02 |
| srichter | even if we make this a utility (I am +1), we still need to import ISchoolToolApplication | 17:02 |
| hoffman | Why is that evil? | 17:02 |
| mgedmin | in my proposal most of the packages depend on ISchoolToolApplication | 17:02 |
| ignas | i think we need a plugable SchoolTool.Preferences | 17:02 |
| mgedmin | in my proposal most of the packages depend on schooltool.app | 17:02 |
| mgedmin | not the other way around | 17:03 |
| srichter | that would be great | 17:03 |
| ignas | that would decrease the need for getSchoolToolApp, and allow adding items to the main preferences view | 17:03 |
| srichter | how feasible is that? | 17:03 |
| srichter | mgedmin: how feasible is that? | 17:03 |
| srichter | I think one of the biggest issues there are the tests | 17:04 |
| mgedmin | I'm not 100% confident it is possible to make everything work without having a single package that depends on everything else and I'm not confident it is possible to make everything work without having a single package that depends on everything else and brings everything together | 17:04 |
| srichter | ok | 17:04 |
| mgedmin | especially functional tests | 17:04 |
| srichter | right, I personally think that functional tests should be outside the package structure | 17:04 |
| mgedmin | which is the reason why I decided that having everything depend on schooltool.app is inevitablr | 17:05 |
| mgedmin | hmm ouside? | 17:05 |
| srichter | because they really test the application SchoolTool and not the functionality of a few packages | 17:05 |
| mgedmin | this was the original idea | 17:05 |
| hoffman | srichter: I was thinking the same thing, | 17:05 |
| mgedmin | and schooltool.app was the application | 17:05 |
| mgedmin | and all ftests lived inside schooltool.app | 17:05 |
| srichter | right, I knew you would be saying this ;-) | 17:06 |
| srichter | so, ST has the same issue as Zope 3 itself | 17:06 |
| mgedmin | perhaps we need to split st.app into st.core and st.app | 17:06 |
| srichter | app is too big | 17:06 |
| mgedmin | the first one would be generic things like ISchooltoolApplication | 17:06 |
| mgedmin | everything could depend on st.core (persons, groups, timetabling, etc) | 17:06 |
| mgedmin | nothing would depend on st.app | 17:06 |
| mgedmin | st.app would depend on everything | 17:06 |
| ignas | maybe packages SchoolTool.schoolToolSkin / SchoolTool.SchoolBellSkin (i know it is dead) would be better than adding tests to the core ? | 17:07 |
| mgedmin | I cannot tell beforehand how big would st.core/st.app be | 17:07 |
| srichter | that sounds cleanest to me | 17:07 |
| ignas | like having SchoolTool.core separate from views/integration with other packages | 17:07 |
| mgedmin | my email proposal is equivalent to st.app being empty and everything being st.core (without the name change) | 17:07 |
| mgedmin | but it might turn out to be unfeasible | 17:07 |
| * mgedmin likes the st.core more and more | 17:07 | |
| mgedmin | we could refactor it slowly | 17:08 |
| mgedmin | move one thing to st.core at a time | 17:08 |
| mgedmin | slowly remove dependencies on st.app | 17:08 |
| srichter | right | 17:08 |
| mgedmin | while continuing to work on other stuff | 17:08 |
| hoffman | Yes, no more grand reorganizations. | 17:08 |
| srichter | though I would not move st.person to st.core for example, since it would bloat the package, but simply consider it coer | 17:08 |
| mgedmin | yep, st.person would depend on st.core | 17:09 |
| mgedmin | st.app would depend on st.person | 17:09 |
| srichter | ok, we are ont eh same page | 17:09 |
| * mgedmin 's goal is an acyclic dependency graph | 17:09 | |
| srichter | right | 17:09 |
| srichter | ok, I think the key to success is getting the UI and tests separated | 17:10 |
| srichter | we have to do a better job using stubs | 17:10 |
| mgedmin | +1 | 17:10 |
| ignas | +1 | 17:10 |
| alga | this would place a lot more responsibility on integration tests | 17:10 |
| srichter | the second step is finding an answer to the package dependency problem through ZCML; I will tackle that next week | 17:10 |
| mgedmin | what is the problem? | 17:11 |
| srichter | that right now the st.requirement pacakge dependes on Section (fore example) just because of the ZCML declaration | 17:11 |
| srichter | or a better example is the notes package | 17:12 |
| mgedmin | ah, ok | 17:12 |
| srichter | somewhere we have to say this object has IHaveNotes | 17:12 |
| mgedmin | the st.app zcml could be responsible for all this | 17:12 |
| srichter | and I think this should be outside st.note and st.person (for example) | 17:12 |
| mgedmin | or we could use zcml:condition | 17:12 |
| srichter | mgedmin: yep, that would be ideal | 17:12 |
| srichter | I think zcml:condition would not serve us well there | 17:13 |
| srichter | it was not designed for those type of tasks | 17:13 |
| mgedmin | you're probably right | 17:13 |
| srichter | we could also have another ZCML glue directory, like integration/ | 17:14 |
| srichter | here is a sci-fi run: | 17:14 |
| srichter | SchoolTool is distributed without st.note | 17:14 |
| srichter | But I decide that I would like st.note to be installed | 17:15 |
| srichter | so I install the package | 17:15 |
| srichter | and then I run a little script where it declares several components as IHaveNotes | 17:15 |
| srichter | this generates a ZCML file that is placed into integration/ | 17:16 |
| srichter | <over> | 17:16 |
| mgedmin | mhm... | 17:16 |
| mgedmin | I don't like this additional step | 17:16 |
| srichter | but how do we know ahead of time which components the user wants to associate notes with? | 17:17 |
| srichter | right now we make the decision for them | 17:17 |
| * mgedmin is a GNOME user | 17:18 | |
| mgedmin | if there's a sensible default, use it | 17:18 |
| srichter | but I think it should be a more obvious way to change it | 17:18 |
| mgedmin | yes | 17:18 |
| srichter | except that GNOME looses gradually their user base because it is too limiting | 17:18 |
| mgedmin | what would you think about putting these extra declarations into the package include file? | 17:18 |
| hoffman | OK, let's not go there. | 17:18 |
| mgedmin | the user could then customize it | 17:18 |
| srichter | that's a good idea | 17:19 |
| hoffman | We don't need a GNOME/KDE flame war. | 17:19 |
| mgedmin | of course you would say that, you use Mac OS X! | 17:19 |
| mgedmin | :) | 17:19 |
| * mgedmin promises to stop joking now | 17:19 | |
| hoffman | And GNOME. | 17:19 |
| mgedmin | I think we have a general agreement, except for the placement of glue zcml | 17:20 |
| mgedmin | but I do not care about that much | 17:20 |
| srichter | I think putting the glue into the package include file, is a low tech approach right now | 17:21 |
| srichter | since we do not know what we want, it is as good as any to try | 17:21 |
| hoffman | Can someone email a summary of the consensus on package structure? | 17:21 |
| mgedmin | I can do that | 17:21 |
| hoffman | OK. Thanks mgedmin. | 17:22 |
| mgedmin | when I get back to the office | 17:22 |
| hoffman | In other news... | 17:22 |
| hoffman | I got an email from Kit today, and I'm going to go ahead and finally propose to Mark that we hire Infrae to work on SchoolTool starting in February. | 17:23 |
| srichter | ok, cool | 17:23 |
| hoffman | So I'll be writing that up today. | 17:23 |
| hoffman | I think we're done then. | 17:24 |
| srichter | then we have a third strongly-opinionated ST developer :-) | 17:25 |
| hoffman | Should we skip next Monday's meeting? | 17:25 |
| srichter | yeah, I will not be here for sure | 17:25 |
| mgedmin | +1 | 17:25 |
| srichter | thanks for the reminder, I meant to mention that | 17:25 |
| hoffman | OK. No meeting next Monday. | 17:26 |
| hoffman | Have a Merry Christmas and a Happy New Year, everyone! | 17:26 |
| mgedmin | you too! | 17:26 |
| srichter | you too! | 17:27 |
| alga | Merry Christmas | 17:27 |
| ignas | +1 | 17:28 |
| * hoffman bangs the virtual yuletide gavel. | 17:28 | |
| tiredbones | I like a Merry Christmas too! | 17:28 |
| mgedmin | tiredbones, then by all means have some | 17:29 |
| hoffman | Merry Christmas tiredbones! | 17:29 |
| tiredbones | I wonder if I'll ever understand Zope 3??? | 17:30 |
| mgedmin | nobody can fully understand Zope 3 | 17:30 |
| mgedmin | :) | 17:30 |
| *** mgedmin has quit IRC | 17:32 | |
| tiredbones | I just read the Adapter.txt doc and I'll probably will have to read many before I understand it. | 17:32 |
| hoffman | I'm finally starting to understand adapters. | 17:33 |
| hoffman | At least enough to understand what they DO, if not how they work. | 17:34 |
| srichter | once you understand adapters, it's all downhill from there | 17:34 |
| hoffman | I think that's the big hurdle. | 17:34 |
| srichter | I think it is a big hurdle, because a lot of developers are not good OO developers either | 17:35 |
| alga | developers developers developers developers :-) | 17:36 |
| srichter | LOL | 17:36 |
| hoffman | I also think the metaphor of the adapter is too simple to describe what's actually happening. | 17:36 |
| *** alga has quit IRC | 17:36 | |
| srichter | hoffman: really? that's interesting to hear... (it never occurred to me that way) | 17:37 |
| tiredbones | Our interfcaes join by adapters? If that is the proper way of asking the question. | 17:38 |
| hoffman | I'm not sure I can fully articulate why. | 17:38 |
| hoffman | Hm... are interfaces joined by adapters... | 17:43 |
| hoffman | An adapter allows one interface to act like another. | 17:43 |
| hoffman | So the simple metaphorical case is a power adapter. | 17:44 |
| hoffman | Lets you use a US plug in a European outlet. | 17:44 |
| hoffman | Zope 3 adapters are a bit more ambitious. | 17:44 |
| hoffman | They'll adapt your electric razor to act as a hairdryer. | 17:45 |
| hoffman | If you want. | 17:45 |
| tiredbones | Thanks for the reply. I think my think is in the right direction. I just have to reread and play with some more. | 17:47 |
| hoffman | Yes. | 17:47 |
| srichter | I'll note that all programming adapters are that ambitous | 17:52 |
| srichter | (at least the ones I saw) | 17:52 |
| hoffman | Yeah, it is probably not the case that Zope 3 adapters are different, but that there is a leap between the basic metaphor and examples and the many ways the technique can be implemented. | 17:54 |
| srichter | right | 18:01 |
| *** mgedmin has joined #schooltool | 18:02 | |
| tiredbones | srichter, have you read philKON blog on the over use of zcml and do you agree with it? | 18:08 |
| * mgedmin read that, and /me agrees | 18:11 | |
| srichter | tiredbones: do you have a link? | 18:20 |
| * srichter is not reading or writing blogs; I have no time for that | 18:20 | |
| tiredbones | srichter, http://www.z3lab.org/sections/blogs/philipp-weitershausen/2005_12_14_zcml-needs-to-do-less | 18:43 |
| tiredbones | srichter, I put the pointer here because I wasn't sure your where on zope3-dev. | 18:47 |
| srichter | note that his thought have been expressed first by Jim Fulton about 1-1.5 years ago | 18:48 |
| srichter | in fact, a lot of recent changes to the very core have been pushing in this direction | 18:48 |
| srichter | also, see the new forms package | 18:49 |
| tiredbones | Glad to know this. Thanks for your reply. | 18:50 |
| tiredbones | srichter,Glad to know this. Thanks for your reply. | 18:50 |
| srichter | I'll also note that this is only a blog entry, so he makes a lot of suggestions in a fairly non=public place for Zope 3 developers | 18:53 |
| srichter | I'll wait till he writes a proposal | 18:53 |
| *** erchache has joined #schooltool | 19:15 | |
| erchache | hi | 19:15 |
| *** thisfred has quit IRC | 19:21 | |
| *** mgedmin has quit IRC | 19:55 | |
| *** erchache has quit IRC | 20:32 | |
| *** mgedmin has joined #schooltool | 20:33 | |
| *** alga has joined #SchoolTool | 21:23 | |
| *** hoffman has quit IRC | 21:23 | |
| *** alga_ has joined #SchoolTool | 21:48 | |
| *** alga has quit IRC | 21:49 | |
| *** alga_ has quit IRC | 22:19 | |
| *** ignas has quit IRC | 22:19 | |
| *** ignas has joined #schooltool | 22:20 | |
| *** ignas has quit IRC | 22:25 | |
| *** Aiste has quit IRC | 22:31 | |
| *** mgedmin has quit IRC | 22:36 | |
Generated by irclog2html.py 2.15.1 by Marius Gedminas - find it at mg.pov.lt!