| *** tpb has joined #melange | 00:00 | |
| *** solydzajs has joined #melange | 01:41 | |
| *** ChanServ sets mode: +o solydzajs | 01:41 | |
| *** penyaskito has quit IRC | 01:47 | |
| *** solydzajs has quit IRC | 01:49 | |
| *** jamtoday has joined #melange | 02:43 | |
| *** tansell_laptop has joined #melange | 02:50 | |
| *** rwatson has quit IRC | 03:41 | |
| *** Merio has joined #melange | 05:16 | |
| *** MatthewWilkes has joined #melange | 06:42 | |
| *** dr__house has joined #melange | 07:45 | |
| *** rwatson has joined #melange | 08:10 | |
| *** mithro_ has quit IRC | 08:58 | |
| *** MatthewWilkes has quit IRC | 09:51 | |
| *** MatthewWilkes has joined #melange | 09:51 | |
| *** florinciu has joined #melange | 09:56 | |
| *** madrazr has joined #melange | 10:04 | |
| *** ajaksu has joined #melange | 10:06 | |
| *** MatthewWilkes has quit IRC | 10:33 | |
| *** madrazr has left #melange | 10:34 | |
| ajaksu | jamtoday: ping | 10:34 |
|---|---|---|
| *** MatthewWilkes has joined #melange | 10:35 | |
| *** rwatson has left #melange | 11:18 | |
| *** MatthewWilkes has quit IRC | 12:13 | |
| *** dhans has joined #melange | 12:13 | |
| Merio | hi dhans ^_^ | 12:15 |
| dhans | hello Merio:) | 12:15 |
| Merio | I've just cloned the hg repository from bitbucket and tried the collect statistics | 12:15 |
| dhans | and you got error? | 12:15 |
| Merio | The final json string is {"0": 0, "1": 0, "2": 0, "3": 0, "4": 0, and so on till 99... is it normal? | 12:15 |
| Merio | (I mean in the current status of your code) | 12:16 |
| Merio | It gives me always "records completed => 0" | 12:16 |
| Merio | (I've done the seed_db) | 12:17 |
| dhans | but you collected stats on your local instance? not on dhansmelange? | 12:18 |
| Merio | Yes, in my local instance | 12:18 |
| dhans | ok, and did you add some students? | 12:19 |
| dhans | because I just collected the stats on dhansmelange it works, because it says "35" : "10" | 12:19 |
| Merio | Hmm well the seed_db creates students with birth date as db.DateProperty.now(), | 12:20 |
| Merio | However I think they should be shown as "0".. no? | 12:21 |
| dhans | generally yes | 12:21 |
| dhans | maybe I did not cover this edge condition that someone is born now:P | 12:22 |
| Merio | eheheh :P | 12:22 |
| dhans | it's even impossible to pick such a date from calendar... it ends at 2008:) | 12:23 |
| Merio | To be fair, in the Datastore viewer I can't see any student entity | 12:24 |
| Merio | Oh well, all ok | 12:26 |
| Merio | I'll figure out myself no probs | 12:26 |
| dhans | ok, because I just checked out and it works for someone that's born today:) | 12:27 |
| Merio | Yes, simple seed_db doesn't seed Students | 12:27 |
| Merio | That's ok, need to use seed_many | 12:27 |
| Merio | Well, let's go to point (4) Memcache issue (and statistic logic) | 12:28 |
| Merio | It's ok for you | 12:28 |
| Merio | ? | 12:28 |
| dhans | yeah, because I just wanted to mention that I almost fixed this issue:) | 12:28 |
| Merio | oh well great! | 12:29 |
| dhans | if you pull my new commit it almost works, but I'll make a few more changes to make the code look nicer:) | 12:29 |
| Merio | what's the "almost"? | 12:29 |
| dhans | just as said, it works, but I will make some stylish changes | 12:30 |
| Merio | pulled | 12:30 |
| Merio | Going to open the logic | 12:31 |
| Merio | What was the problem? | 12:31 |
| dhans | after updating an entity I didn't actualize memcache | 12:32 |
| dhans | ok actualize is a wrong word (just checked in dictionary:) I meant update | 12:33 |
| Merio | Yes, understood anyway ^_^ | 12:33 |
| Merio | where is "self.updateEntityProperties()"? | 12:34 |
| Merio | It's in some parent class? | 12:34 |
| Merio | ok got it it's in base :) | 12:34 |
| dhans | yes, it's right there:) | 12:35 |
| Merio | Ok so | 12:36 |
| *** dr__house has quit IRC | 12:38 | |
| Merio | We can go for (5) | 12:38 |
| Merio | JS Layer and JS stuff | 12:38 |
| dhans | ok | 12:38 |
| Merio | Is it ok for you? | 12:38 |
| Merio | Great | 12:38 |
| Merio | So, I've explored jQueryUI today | 12:38 |
| Merio | I've found it very very powerful | 12:39 |
| Merio | We can use it for iGoogle style dashboards | 12:39 |
| *** dr__house has joined #melange | 12:39 | |
| Merio | And it's easy to work with it | 12:39 |
| Merio | At least for simple things | 12:39 |
| dhans | that would be grate to have such a dashboard:) | 12:40 |
| Merio | Also we can do many wonderful things with it, but I'm going to think about it :) | 12:41 |
| Merio | I would like also to implement a sort of Ajax push or something for the "records completed" stuff, but I need to explore some solutions further, beyond the obvious "polling" | 12:42 |
| Merio | So, for the spike solution, I really would like to try using a visualization API and show a sample chart | 12:43 |
| dhans | ok | 12:44 |
| dhans | I'll implement some sort of filter that will convert final_json_string to the format you want | 12:45 |
| Merio | Ok, let me see something about the visualization API | 12:45 |
| Merio | What's your thoughts about leaving the building of the final JSON string as a separate pat | 12:46 |
| Merio | part | 12:46 |
| Merio | So, if we ever need to change it, or to have different generations for different purposes, we're ready | 12:46 |
| dhans | isn't is something that I just suggested? | 12:48 |
| Merio | Yes, I mean to design it with this flexibility in mind | 12:48 |
| Merio | However | 12:48 |
| *** durin42 has quit IRC | 12:48 | |
| Merio | I've just seen a very good thing I didn't notice before | 12:48 |
| Merio | We can use the JSON directly without any iteration | 12:49 |
| Merio | Check it out | 12:49 |
| Merio | http://code.google.com/apis/ajax/playground/?type=visualization#json_data_table | 12:49 |
| tpb | <http://ln-s.net/3EqV> (at code.google.com) | 12:49 |
| dhans | sweet:) | 12:50 |
| Merio | A nc | 12:51 |
| Merio | And | 12:51 |
| Merio | Check this ==> http://code.google.com/apis/visualization/documentation/dev/gviz_api_lib.html | 12:51 |
| tpb | <http://ln-s.net/3EqY> (at code.google.com) | 12:51 |
| dhans | wonderful:) | 12:52 |
| Merio | That surely would help! | 12:52 |
| dhans | it's great you found it! | 12:52 |
| Merio | So probably we don't need any JavaScript to handle the responses for visualizations | 12:53 |
| dhans | but as you can see (if I understand correctly), there are different json format for different visualization, aren't they? | 12:53 |
| Merio | What do you mean? The "JSON string"/"JSON response"/"JavaScript String" thing? | 12:53 |
| dhans | for example: the JSONObject in the first link that you sent is different than the JSONObject that you used in the example on wiki | 12:55 |
| dhans | but that's quite obvious:) | 12:55 |
| Merio | Yes, that's because this is something preformatted to pass the JSON directly to the Visualization constructor | 12:56 |
| dhans | yes | 12:56 |
| Merio | Well... can you give me some time (about an hour or so) to explore those things further? | 12:56 |
| dhans | and if I understand you correctly, yesterday you wanted me to create final_json_string in the format you used on wiki? | 12:57 |
| Merio | It seems very promising on our integration | 12:57 |
| Merio | Yes, but I've found that link only now :D | 12:57 |
| dhans | ok:) | 12:57 |
| Merio | Also there are many different interesting things | 12:57 |
| dhans | because my suggestion is to keep final_json_string as simple as possible and convert it later | 12:57 |
| Merio | So... well... let's just wait for me to read some documents, I'm finding great things that might help our work ^__^ | 12:57 |
| dhans | sure:) | 12:58 |
| *** florinciu has quit IRC | 12:58 | |
| Merio | Are you here in an hour? | 12:58 |
| dhans | yes | 13:02 |
| *** durin42 has joined #melange | 13:03 | |
| *** ChanServ sets mode: +v durin42 | 13:03 | |
| *** dr__house has quit IRC | 13:46 | |
| *** dhaun has joined #melange | 13:47 | |
| Merio | dhans: ping | 13:59 |
| dhans | pong | 14:00 |
| Merio | So, I think there's something interesting out there for JS integration, but some solutions might imply different design, so I think it's better to ask the devs what they think about. I'll send an email to the mailing list as a RFC | 14:01 |
| Merio | dhans: What are you going to do in the meantime? Just for info and syncing | 14:03 |
| dhans | meantime - it depends how long it will be:) | 14:04 |
| Merio | ahah :) | 14:04 |
| Merio | Well, I think I'm going to send it in the next hour or so | 14:04 |
| Merio | But we might be waiting at least a day for our mentors to respond | 14:05 |
| dhans | but now I want to finish this memcache thing | 14:05 |
| Merio | That's ok | 14:05 |
| *** MatthewWilkes has joined #melange | 14:39 | |
| *** madrazr has joined #melange | 14:49 | |
| *** madrazr has quit IRC | 14:51 | |
| *** tlarsen has joined #melange | 15:38 | |
| *** ChanServ sets mode: +o tlarsen | 15:38 | |
| Merio | dhans: ping | 15:38 |
| *** James--Crook has joined #melange | 15:52 | |
| James--Crook | Hi Merio - you wanted to talk with me? | 15:52 |
| James--Crook | Merio: also, can I delete the higher quality version of the F1 I have? Basically the DVD had everything on I hope and so the original is not needed anymore (can free up some space for Bobby). | 15:54 |
| Merio | James--Crook: yep, of course ^__^ I didn't know you were keeping it :) | 15:55 |
| James--Crook | ...just wanted to be sure :-) | 15:55 |
| Merio | James--Crook: save some space for Sunday 7th if you can :P | 15:58 |
| *** tlarsen has left #melange | 15:59 | |
| James--Crook | Will do :-) | 15:59 |
| ajaksu | James--Crook: I've got some cool survey stuff to show off and no appengine verified account, do we have one for hacky temporary demos? :) | 16:00 |
| James--Crook | I believe there is, but sverre is the person to talk to about it. | 16:03 |
| James--Crook | I lurk on wiki/issue tracker/e-mail.... Not actively developing... | 16:03 |
| James--Crook | I am keen to see cool survey stuff :-) | 16:04 |
| James--Crook | What have you got? | 16:04 |
| Merio | James--Crook: so, I was looking in the visAPI page, and found this => http://code.google.com/apis/visualization/documentation/release_notes.html#whatsnew | 16:05 |
| tpb | <http://ln-s.net/3EuQ> (at code.google.com) | 16:05 |
| James--Crook | Merio: ..looking | 16:06 |
| Merio | James--Crook: in that I've seen | 16:06 |
| Merio | James--Crook: details about a visualization API query language => http://code.google.com/apis/visualization/documentation/querylanguage.html | 16:06 |
| tpb | <http://ln-s.net/3Eua> (at code.google.com) | 16:06 |
| Merio | James--Crook: Which you can even use live ==> http://code.google.com/apis/ajax/playground/?type=visualization#using_the_query_language | 16:07 |
| tpb | <http://ln-s.net/3Eub> (at code.google.com) | 16:07 |
| Merio | James--Crook: Unfortunately that query language has only one implementation in Java, and seems one in PHP (even if not perfectly compliant, not released by Google itself) | 16:08 |
| James--Crook | Do you think the query language is good? By that I mean, a good design, one that we would want? | 16:09 |
| Merio | James--Crook: well actually they're queries sent by JavaScript... so they can be customized in a very easy and dynamic way. I mean... a click and we can have different kind of queries... dynamically changing the charts | 16:12 |
| James--Crook | I am thinking that it does not solve one of our most important problems. | 16:12 |
| James--Crook | It is strictly one table at a time, as far as I can see. | 16:13 |
| James--Crook | If we're driving filters (select) and aggregation (group_by, sum) from a GUI, then we are possibly just as well off doing the selections and grouping 'directly' ourselves... | 16:14 |
| Merio | There are many things we could do | 16:15 |
| Merio | As you can see, in this very moment, we can use that feature only by using Google Spreadsheets | 16:16 |
| Merio | So we can have an exporting to it (in some way aggregating tables) and then use the query language over it | 16:17 |
| James--Crook | Hmm. That makes it a bit too limiting. We want to be able to go direct from appengine to chart. | 16:17 |
| Merio | However | 16:17 |
| Merio | I've found also another project that might be useful | 16:17 |
| Merio | Which is jLinq ==> http://www.hugoware.net/ | 16:18 |
| tpb | Title: jLinq - Javascript Query Language (at www.hugoware.net) | 16:18 |
| James--Crook | ..looking... | 16:18 |
| James--Crook | Oh. I Like! | 16:18 |
| Merio | If say, for example, we come with some JSON objects from the appengine | 16:18 |
| James--Crook | It's not artificially introducing a string version of the query. | 16:19 |
| Merio | We can query over it/them in many many ways | 16:19 |
| James--Crook | That makes it nicer for tying up to controls on a dashboard. (I think) | 16:19 |
| ajaksu | James--Crook: sorry for the delay, I've got 'pick_many' via checkboxes, starting to work on a nice 'edit choice field' interface and multiple rendering options for a given field :) | 16:19 |
| James--Crook | ajaksu: wow! I must see this! | 16:19 |
| Merio | There are also other projects like jLinq.. there is JSONPath and JSONQuery. Both working with "strings", and are inspired by XPath | 16:20 |
| James--Crook | ajaksu: Merio has been looking at jQueryUI. Are you using any extra / new libs? | 16:20 |
| Merio | ajaksu: James--Crook: yep, I think it's definitely a must-have | 16:21 |
| Merio | ajaksu: what do you need for the interface? | 16:21 |
| James--Crook | Merio: personal bias showing here... but the stringing together of the filters/aggregators in code needs to be done whatever is done, so I am reluctant to introduce a parser/string-creator that is not needed... What do you think? | 16:21 |
| ajaksu | James--Crook: no, I know very little about JS and jquery, but just reading Merio's messages I see we could do things a lot better :) | 16:21 |
| James--Crook | ajaksu: well.. with more code :-) | 16:22 |
| James--Crook | some of the jQuery UI features are very very nice though. | 16:22 |
| ajaksu | Merio: currently we add and remove form fields and sprinkle some events magic I have a superficial understanding of.... | 16:22 |
| ajaksu | I want to have add/remove/edit/reorder options/fields | 16:23 |
| Merio | James--Crook: well, actually we need something like a "query constructor" anyway if we plan to change dynamically the data (and thus the chart). Then we can output code for jLinq or a string, that's not too different to me, but I might be wrong | 16:23 |
| James--Crook | ajaksu: we'll be wanting to use (gmail like) checkbox lists elsewhere in Melange. I'm assuming that they are currently closely tied in to surveys? | 16:24 |
| Merio | ajaksu: jQuery UI should work well on it (for example on reordering) but I should see in depth what are your needs to help better | 16:24 |
| ajaksu | Merio: our main JS file is at http://github.com/jamslevy/Melange/blob/94efbb25f0eac0486cf05f6a4589bd25556c1b49/app/soc/content/js/edit_survey.js and we're about to refactor that now I'll add real edit fields stuff :) | 16:25 |
| tpb | <http://ln-s.net/3Eux> (at github.com) | 16:25 |
| James--Crook | Merio: mmm... main difference is that code output does not need to be parsed by more code, it's ready to run. Of course I agree, there will be a constructor object and we could generate multiple things in the same way if we so choose. Hmm. And we will need a viewable representation of the list of steps in processing to show to the user too.. | 16:27 |
| ajaksu | James--Crook: right now, yes, but I plan to make the code general by creating two (pick_one and pick_multi) kind of generic fields, so the 'add new option' features can be reused... we could then refactor the 'add new field' logic to make it a generic dynamic form creation tool | 16:28 |
| James--Crook | Merio: This is *way* ahead of what we need at the moment :-) | 16:28 |
| ajaksu | James--Crook: add that base serialization format you mentioned and hey, we've got something way cool :) | 16:28 |
| Merio | James--Crook: so you think jLinq might be better? | 16:30 |
| Merio | ajaksu: have you seen in the jQuery site if there are plugins that already do those things? | 16:31 |
| James--Crook | ajaksu: cool. What do you want to happen now in melange? You're probably the free-est active coder person in Melange, and it is very much up to you where you want to work next, what you wnat to know and do... | 16:32 |
| James--Crook | Merio: Yes, admittedly I've not had a long to look at it yet, but I like the look of jlinq. It looks exactly the approach we would want. | 16:33 |
| James--Crook | The only questions in my mind is how much we might possibly want to do server side, and how we deal with two (or more) table queries. | 16:33 |
| ajaksu | Merio: no, I'm following James Levy's lead and working with the JS he adds... you were calling people to discuss a nice approach to JS, right? | 16:33 |
| James--Crook | I think development will be faster and more flexible if we do complex queries like these client side, not to mention more responsive. | 16:34 |
| ajaksu | James--Crook: I still have to finish my part of surveys, so I'll be the freeest next week, hopefully :) | 16:34 |
| * James--Crook grins | 16:35 | |
| Merio | ajaksu: yep I was trying to :) | 16:35 |
| Merio | James--Crook: jLinq has a join function too | 16:36 |
| James--Crook | ajaksu: In terms of surveys being ready for GSoC use, it is nearly there isn't it? I get the impression that mostly it is that you wnat to tidy the code, that there is not functionality missing. Incidentally, surveys will make a great test case for stats.... | 16:36 |
| Merio | James--Crook: that's for sure ^_^ | 16:36 |
| James--Crook | Can we talk a little about that? | 16:37 |
| ajaksu | Merio: we brought in a plugin to expand textareas and it broke things, so I'd rather get things working in a simple JS way and help pushing for a standardized JS approach :) | 16:37 |
| Merio | James--Crook: about surveys for stats? | 16:37 |
| James--Crook | On the wiki stats page there is a note that there may be some questions (like what years did you participate in in the past) that are helpful to us in gathering *useful stats for improving the program*. | 16:38 |
| Merio | ajaksu: why it broke things? | 16:38 |
| Merio | James--Crook: yes, that's definitely the most interesting part of the stats | 16:38 |
| ajaksu | James--Crook: sort of, I can't run James's latest commits as-is, we've diverged a bit in code, so I have some real work to do before I consider us to be in the polish-things-up period :) | 16:39 |
| James--Crook | Merio: so we are going to have the most difficulty with the free text fields in surveys, which will be little use to us in stats. So we need to think about what we can get that is numerical (and useful) and confirm that survey can actually do it... | 16:39 |
| ajaksu | Merio: I don't know exactly why or how, looks like it added hidden fields and we had buggy form handling code, so it went in then out before I got to know it :) | 16:40 |
| James--Crook | ajaksu: got it. It sounds excellent progress, but I will hold off on syaing that until I see it on an appspot site :-) | 16:40 |
| Merio | James--Crook: I think a thread in the dev list should help :) | 16:40 |
| James--Crook | +1 on that | 16:40 |
| James--Crook | Actually +2. It is something where many points of view will help. | 16:40 |
| Merio | James--Crook: also because I would like LH to join in the discussion (don't know when she'll come back) | 16:41 |
| James--Crook | I believe she is at Google I/O.. but don't quote me. | 16:41 |
| *** madrazr has joined #melange | 16:42 | |
| Merio | ajaksu: eheh :) Well, if you've ever need some JS help just ping me | 16:42 |
| Merio | James--Crook: are you going to start the thread? | 16:42 |
| James--Crook | Me? Nope. You start it :-) | 16:42 |
| *** madrazr has quit IRC | 16:43 | |
| Merio | James--Crook: That's ok, I'll do :) Now dinner is ready here ^__^ | 16:43 |
| Merio | James--Crook: I also want to start a thread about jLinq and all the query stuff | 16:43 |
| James--Crook | OK. Merio, You also wanted to talk about other aspects of more flexible stats or is that covered by the jlinq discussion? | 16:43 |
| Merio | James--Crook: but I'll do tomorrow probably, want to put some sensible thought on them too :) | 16:43 |
| James--Crook | Makes sense. | 16:44 |
| ajaksu | Merio: I almost interrupted your yesterday chat with dhans to ask about how you think we could re-organize our code, but it was such a great conversation I just sat and watched :) | 16:44 |
| Merio | James--Crook: well actually I think we can go for a thread in the dev list, because it might imply some changes on the design | 16:44 |
| James--Crook | Main thing will be the exchanges with dhans... The collection on the surver is interesting to discuss. | 16:44 |
| James--Crook | Yes to that. | 16:44 |
| James--Crook | The Access rights needs more working out in particular. | 16:45 |
| James--Crook | It is possible to do anything for host. | 16:45 |
| Merio | ajaksu: well, let's just plan a meeting on it so ^__^ Perhaps that would be better when they'll be back from the all-hands | 16:45 |
| James--Crook | And that might be the way we do it for now (i.e. for next two weeks or so) | 16:45 |
| James--Crook | and then we start thinking more seriously about what tables we could supply more generally... | 16:46 |
| Merio | James--Crook: completely agree | 16:46 |
| James--Crook | and there will also be some discussion about caching stats derived by host, but for general view... | 16:47 |
| ajaksu | Merio: great, thanks a lot, I'll ping the list next week as we should also have a much better list of wants and wtfs by then :) | 16:47 |
| James--Crook | What I mean by that... | 16:47 |
| James--Crook | is host works on their chart in a totally flexible jlinq way... | 16:47 |
| James--Crook | then publishes a fixed view of their chart. | 16:48 |
| James--Crook | At that stage they might push some aggregated data back out to GAE (via a JSON object) | 16:48 |
| *** ankitg has quit IRC | 16:48 | |
| James--Crook | so that users will load the reduced JSON object, not the raw data. | 16:48 |
| James--Crook | anyway we can discuss later. | 16:49 |
| James--Crook | So, you get two actions... two e-mails to write... one today... one tomorrow... and I get off scot free :-) | 16:49 |
| James--Crook | ajaksu... anything you need at the moment? | 16:49 |
| ajaksu | James--Crook: nope, I'll ping Sverre about the demo instance, thanks a lot for the guidance :) | 16:50 |
| James--Crook | cool. Looking forward. | 16:50 |
| Merio | James--Crook: today you mean the stats/survey one? | 16:50 |
| James--Crook | bye-for-now all. | 16:50 |
| James--Crook | Merio: yep... what stats do we want to do on mid term surveys. What questions would be good to ask - e.g to get years of previous participation which are not easily obtainable any othe rway... | 16:52 |
| Merio | James--Crook: ok, perfect :) | 16:53 |
| James--Crook | and if we see patterns in how happy students are with their mentors vs size of group, how long org has been doing gsoc for etc... | 16:53 |
| ajaksu | bye :) | 16:53 |
| James--Crook | bye | 16:54 |
| *** James--Crook has left #melange | 16:54 | |
| *** dhaun has quit IRC | 17:17 | |
| *** ajaksu has quit IRC | 17:18 | |
| *** dhans has quit IRC | 17:58 | |
| *** Merio has quit IRC | 18:50 | |
| *** ankitg has joined #melange | 20:25 | |
| *** ankitg has quit IRC | 20:26 | |
| *** mithro_ has joined #melange | 20:30 | |
| *** ChanServ sets mode: +v mithro_ | 20:30 | |
| *** MatthewWilkes has quit IRC | 21:00 | |
| *** tansell_laptop has quit IRC | 21:07 | |
| *** tansell_laptop has joined #melange | 21:07 | |
Generated by irclog2html.py 2.17.2 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!