Hi, Here is the log of the chatroom at http://farsides.com/chat/ (or irc://irc.freenode.net/#farsides) for the 2011-07-11.
2011-07-11 06:54:45+0000 ShadowBlackfire (ShadowBlackfire) is now online 2011-07-11 06:55:14+0000 <ShadowBlackfire> Hi 2011-07-11 07:24:42+0000 tempuramerino (tempuramerino) is now online 2011-07-11 08:04:59+0000 tempuramerino (tempuramerino) is now online 2011-07-11 08:30:26+0000 arbrandes (arbrandes) is now online 2011-07-11 08:45:40+0000 dachary (dachary) is now online 2011-07-11 10:44:24+0000 ShadowBlackfire (ShadowBlackfire) is now online 2011-07-11 10:44:47+0000 <ShadowBlackfire> Hello world 2011-07-11 10:46:28+0000 <dachary> hi 2011-07-11 10:47:45+0000 <ShadowBlackfire> What's up ? 2011-07-11 10:49:20+0000 <dachary> going to rmll.info 2011-07-11 13:03:27+0000 tempuramerino (tempuramerino) is now online 2011-07-11 16:17:09+0000 dachary (dachary) is now online 2011-07-11 16:24:33+0000 arbrandes (arbrandes) is now online 2011-07-11 17:27:27+0000 tempuramerino (tempuramerino) is now online 2011-07-11 17:42:15+0000 dachary (dachary) is now online 2011-07-11 17:50:57+0000 divya (divya) is now online 2011-07-11 18:01:46+0000 dachary (dachary) is now online 2011-07-11 18:02:21+0000 <dachary> arbrandes: hey 2011-07-11 18:02:31+0000 <dachary> on the train toward RMLL. 2011-07-11 18:02:52+0000 <arbrandes> hi Loic!~ 2011-07-11 18:03:05+0000 <arbrandes> I envy you Europeans and your trains 2011-07-11 18:03:10+0000 <arbrandes> Best means of transportation :) 2011-07-11 18:03:12+0000 <dachary> :_D 2011-07-11 18:03:15+0000 <dachary> indeed 2011-07-11 18:03:23+0000 <dachary> the railway station is 10 minutes from home 2011-07-11 18:03:29+0000 <dachary> very convenient 2011-07-11 18:04:25+0000 <arbrandes> Very cool indeed! I love hopping around different cities by train when I visit my sister (in Cologne) 2011-07-11 18:08:32+0000 <dachary> On the phone Xavier told me that the game primarily had success with children. 2011-07-11 18:08:44+0000 <dachary> I just created the http://tickets.farsides.com/issues/218 task for Matjaz 2011-07-11 18:08:55+0000 <arbrandes> Ahh, yes, I saw the post on buddypress, awesome 2011-07-11 18:09:07+0000 <dachary> There is a post ? 2011-07-11 18:09:10+0000 * dachary checking 2011-07-11 18:09:19+0000 <arbrandes> "Misc" group 2011-07-11 18:09:39+0000 <dachary> thanks 2011-07-11 18:10:31+0000 <arbrandes> Good idea, to have a jquery plugin for chat 2011-07-11 18:11:30+0000 <dachary> There is not much space on the screen but we'll see how it goes 2011-07-11 18:12:31+0000 <dachary> Are you going to use an existing code base for facebook authentication ? I kind of remember you explored the existing alternatives a few weeks ago. But I don't remember what your conclusion was. 2011-07-11 18:12:34+0000 <dachary> arbrandes: ^ 2011-07-11 18:16:15+0000 <arbrandes> I'm exploring them now, as a matter of fact 2011-07-11 18:16:20+0000 <arbrandes> This one looks nice: https://github.com/aidanlister/django-facebook/tree/master/django_facebook 2011-07-11 18:16:42+0000 <arbrandes> I'm not sure if I'll actually use it all 2011-07-11 18:16:59+0000 <arbrandes> But it'll be good for an idea of where to start, in any case 2011-07-11 18:18:33+0000 arbrandes_ (arbrandes_) is now online 2011-07-11 18:19:58+0000 <arbrandes_> <arbrandes> I'm exploring them now, as a matter of fact 2011-07-11 18:19:59+0000 <arbrandes_> <arbrandes> This one looks nice: https://github.com/aidanlister/django-facebook/tree/master/django_facebook 2011-07-11 18:19:59+0000 <arbrandes_> <arbrandes> But it'll be good for an idea of where to start, in any case 2011-07-11 18:19:59+0000 <arbrandes_> <arbrandes> I'm not sure if I'll actually use it all 2011-07-11 18:20:07+0000 <arbrandes_> dachary, not sure you got that, I got disconnected ^ 2011-07-11 18:21:53+0000 <dachary> I got it ;-) 2011-07-11 18:22:35+0000 * dachary reading http://aidanlister.com/2010/11/better-integration-with-facebook-for-django/ 2011-07-11 18:23:52+0000 <dachary> arbrandes_: Django Facebook version 2.0 2011-07-11 18:24:02+0000 <dachary> Thierry Schellenbach refers to this in a comment 2011-07-11 18:25:14+0000 <dachary> it looks like they are competing with each other ;-) 2011-07-11 18:25:15+0000 <dachary> https://github.com/tschellenbach/Django-facebook 2011-07-11 18:26:48+0000 * dachary running gource on both for fun 2011-07-11 18:28:52+0000 <arbrandes_> Hahaha, yes, but django-facebook (with small "d"), seems to be a bit lighter weight 2011-07-11 18:30:47+0000 <dachary> https://github.com/tschellenbach/Django-facebook has gained a lot of momentum recently. The other one is a one man operation mostly. That's the kind of stuff gource shows and it's much more difficult to figure out otherwise. 2011-07-11 18:31:10+0000 <dachary> well 2011-07-11 18:31:14+0000 <dachary> it's not exactly true 2011-07-11 18:31:19+0000 <dachary> git shortlog -nes 2011-07-11 18:31:24+0000 <dachary> says 2011-07-11 18:31:25+0000 <dachary> 26 Aidan Lister <[email protected]> 2011-07-11 18:31:25+0000 <dachary> 4 Andrew Ash <[email protected]> 2011-07-11 18:31:25+0000 <dachary> 2 Aidan Lister <[email protected]> 2011-07-11 18:31:25+0000 <dachary> 1 Chris <[email protected]> 2011-07-11 18:31:57+0000 <dachary> on the otherone it says: 2011-07-11 18:31:58+0000 <dachary> http://pastebin.com/3fPNAGxA 2011-07-11 18:32:22+0000 <dachary> but it does not show how it goes overtime 2011-07-11 18:32:52+0000 <dachary> where gource -s 0.1 tells me within a few seconds that https://github.com/tschellenbach/Django-facebook has seen a lot of activity in the past three months 2011-07-11 18:34:05+0000 <arbrandes_> Interesting, this gource 2011-07-11 18:34:51+0000 <arbrandes_> installing it now 2011-07-11 18:35:12+0000 <dachary> you need 3D but it works with even the most basic 3D driver 2011-07-11 18:35:49+0000 <dachary> arbrandes_: do you think it would be better to re-use code that is simple and copy it ? 2011-07-11 18:36:00+0000 <arbrandes_> Ahhhhhhhh, I've seen it before, it really is awesome 2011-07-11 18:36:31+0000 <dachary> arbrandes_: it is awesome yes. And the best part of it is that it is simple to run. 2011-07-11 18:37:12+0000 <arbrandes_> dachary, usually if I'm not going to use a whole bunch of stuff that comes with these libraries, I tend to just "get the drift" and implement a simpler sub-set. This is easier to maintain. 2011-07-11 18:37:42+0000 <dachary> facebook is scary because they keep changing the API 2011-07-11 18:37:57+0000 <arbrandes_> Also, sometimes the library simply doesn't do what you want, and you end up having to hack it all to pieces anyway. 2011-07-11 18:37:59+0000 <arbrandes_> But you're right 2011-07-11 18:38:02+0000 <arbrandes_> on the Facebook thing 2011-07-11 18:38:14+0000 <arbrandes_> It is a good argument to choose a well-maintained, active library 2011-07-11 18:38:23+0000 <dachary> but there does not seem to be any 2011-07-11 18:38:46+0000 <dachary> very active during a few months does not really qualify as well-maintained, mature and reliable ;-) 2011-07-11 18:38:55+0000 <dachary> my train arrives, ttyl ! 2011-07-11 18:38:56+0000 <arbrandes_> dachary, agreed on all counts :) 2011-07-11 18:39:00+0000 <arbrandes_> dachary, later! 2011-07-11 18:39:05+0000 <dachary> really tough choice 2011-07-11 18:39:23+0000 <arbrandes_> I'll try to come to a conclusion by tonight, or early tomorrow, and we can discuss it further 2011-07-11 20:31:45+0000 arbrandes (arbrandes) is now online 2011-07-11 22:10:22+0000 dachary (dachary) is now online 2011-07-11 22:10:26+0000 <dachary> back 2011-07-11 22:13:17+0000 <dachary> arbrandes: xavier & david & tartarugafeliz say hi 2011-07-11 22:13:52+0000 <dachary> we had a chat about facebook interface 2011-07-11 22:16:32+0000 <dachary> Xavier opinion is that only a *very* active django plugin is worth depending on 2011-07-11 22:16:49+0000 <dachary> he base his experience on the work he had to do with wordpress 2011-07-11 22:16:59+0000 <dachary> he tried dozens of facebook connect plugins 2011-07-11 22:17:11+0000 <dachary> and ended up chosing the one that was updated daily 2011-07-11 22:17:21+0000 <dachary> because it was the only one that actually worked 2011-07-11 22:17:32+0000 <arbrandes> Hi dachary! Say hi to them for me, too! :) 2011-07-11 22:17:34+0000 <arbrandes> Ok, so... 2011-07-11 22:17:44+0000 <arbrandes> If it's NOT updated very frequently 2011-07-11 22:17:52+0000 <arbrandes> Better to write our own code? 2011-07-11 22:18:10+0000 <dachary> That's what I think actually, yes. 2011-07-11 22:18:15+0000 <dachary> The rationale is as follows: 2011-07-11 22:19:10+0000 <dachary> facebook changes its API so often that *all* efforts to maintain a generic library (drupal, joomla, wp, django, synfony ... ) eventually give up after a few months, maybe a year 2011-07-11 22:19:13+0000 <dachary> that's a fact 2011-07-11 22:19:36+0000 <dachary> you won't find anywhere a library that has been successfully maintained to keep up with facebook during more than a year 2011-07-11 22:20:02+0000 <dachary> it means that if you depend on it, you will end up being the maintainer of said library within a year 2011-07-11 22:20:20+0000 <arbrandes> Agreed on all points 2011-07-11 22:20:30+0000 <dachary> and that will most certainly be *more* work than writing your own code from the start, for the subset of the facebook functionality you can afford 2011-07-11 22:21:16+0000 <dachary> given the size of farsides and your own workforce, it means a subset of facebook behavior that is compatible with X days of your time every W weeks 2011-07-11 22:22:15+0000 <arbrandes> Excellent, that makes complete sense, and I will actually be much more comfortable maintaining our own code. 2011-07-11 22:22:40+0000 <dachary> Xavier is very aware that facebook will require a *regular* effort. He does not think you can come up with a solutio nthat won't require regular maintainance. It's sad but it's facebook fault. 2011-07-11 22:22:44+0000 <arbrandes> (like I did with registration, instead of using django-registration) 2011-07-11 22:23:04+0000 <dachary> s/regular/periodic/ 2011-07-11 22:23:25+0000 <dachary> in french regulier (sounds like regular) means periodic ;-) 2011-07-11 22:23:37+0000 <dachary> sorry for the confusion 2011-07-11 22:23:52+0000 <arbrandes> Yes, I agree, from what I've already seen of the API, it's kind of a mess. 2011-07-11 22:23:54+0000 <arbrandes> np :) 2011-07-11 22:23:56+0000 <dachary> and yes, since it matches your own approach with the registration plugin, it is all good 2011-07-11 22:24:11+0000 <dachary> ;-) 2011-07-11 22:25:06+0000 <dachary> from what I understand (please clear this with Xavier to be 100% sure) what is desirable is to have a facebook connect button that asks permission to copy in the django registration base as much information as possible from the player facebook account 2011-07-11 22:25:59+0000 <arbrandes> The "as much information as possible" is what needs to be decided on. Apparently, if you ask for too much, people just don't use it (because they can see what you're asking for). This is according to FB's own docs. 2011-07-11 22:26:06+0000 <dachary> when we discussed the subject originaly I think the idea was to use the oauth facebook API for which there are a few examples outthere. 2011-07-11 22:26:30+0000 <dachary> Yes, "as much information as possible" is very tricky. 2011-07-11 22:26:56+0000 <arbrandes> I was thinking to start with just email and name, to match current registration. 2011-07-11 22:27:09+0000 <dachary> A first step is to require as little as possible. But the goal will eventually to ask more such as mail confirmation, etc. 2011-07-11 22:27:34+0000 <arbrandes> Agreed 2011-07-11 22:27:36+0000 <dachary> maybe the email is already too much for facebook 2011-07-11 22:27:41+0000 <dachary> for a first step 2011-07-11 22:27:48+0000 <dachary> maybe it's an advanced topic ;-) 2011-07-11 22:28:51+0000 <dachary> I think facebook connect just give you a login name + a unique identifier to start with 2011-07-11 22:28:59+0000 <dachary> that would be ok for a first step 2011-07-11 22:29:14+0000 <dachary> I think it is required to ask the user permission to get more 2011-07-11 22:29:29+0000 <arbrandes> Well, from what I understood of previous chats, we're not going to be using Facebook for login, we're actually going to register the user *using* FB information . To do so, currently we'd need at the very least email, which if I'm not mistaken, is precisely the identifier in question. 2011-07-11 22:29:35+0000 <dachary> and maybe the good time to ask for that is when inviting people or doing something of the sort 2011-07-11 22:30:04+0000 * dachary thinking 2011-07-11 22:30:06+0000 <arbrandes> (Facebook has no usernames, just emails) 2011-07-11 22:30:19+0000 <arbrandes> Aside from real names, I mean 2011-07-11 22:30:55+0000 <dachary> do you know if facebook will provide you the email without confirmation from the user ? 2011-07-11 22:31:19+0000 <arbrandes> Let me check 2011-07-11 22:32:49+0000 <dachary> I understand the distinction now, thanks for clarifying this. Using facebook for login would mean being integrated in facebook and not importing anything in the current registration database. What you aim at doing is to use facebook connect to collect just enough information to create an entry in the registration database and then connect the user as if it used the regular registration form. Is my understanding correct ? 2011-07-11 22:34:28+0000 <arbrandes> Apparently, getting the user's email with oauth DOES require an addition permission. Minimal information is just name, gender, and list of friends. 2011-07-11 22:34:53+0000 <arbrandes> Yup on you last comment 2011-07-11 22:34:56+0000 <arbrandes> *your 2011-07-11 22:35:25+0000 <arbrandes> Otherwise, we would depend on Facebook for all subsequent logins of that user. 2011-07-11 22:35:57+0000 <arbrandes> I'm not even sure how that would work for the current notification system, which depends on emails. 2011-07-11 22:36:37+0000 <arbrandes> I foresee a Facebook cardstories plugin, too, by the way. :) 2011-07-11 22:36:44+0000 <arbrandes> (for notification) 2011-07-11 22:36:55+0000 <dachary> I think the first step should be to create users with just the name 2011-07-11 22:37:16+0000 <dachary> notifications are disabled if the login is not an email 2011-07-11 22:37:24+0000 <dachary> no harm but less functionality 2011-07-11 22:37:48+0000 <dachary> we can solve this notification problem later 2011-07-11 22:38:25+0000 <dachary> either by using facebook advertisement (like publishing a new game on your wall for your friends to see) 2011-07-11 22:38:46+0000 <dachary> in any case you cannot give cardstories permissions to disclose the mails of your friends 2011-07-11 22:38:57+0000 <dachary> therefore you won't be able to notify *them* by email 2011-07-11 22:39:19+0000 <arbrandes> agreed 2011-07-11 22:39:56+0000 * arbrandes thinking about the implications of just having a name 2011-07-11 22:40:23+0000 <dachary> actually I think there will be a django plugin that will grow. At the moment it's django auth. But I think more and more will be done by django, including notifications. But that's just a hunch. 2011-07-11 22:40:46+0000 <dachary> arbrandes: from the cardstoriespoint of view having just a name is not a problem at all 2011-07-11 22:41:07+0000 <dachary> http://cardstories.org/?player_id=arbrandes 2011-07-11 22:41:11+0000 <dachary> works like a charm 2011-07-11 22:41:13+0000 <arbrandes> dachary, yup 2011-07-11 22:41:25+0000 <arbrandes> I'm thinking about the User model in django 2011-07-11 22:41:36+0000 <arbrandes> (which we're currently using) 2011-07-11 22:41:45+0000 <dachary> ah 2011-07-11 22:41:47+0000 <dachary> right 2011-07-11 22:41:52+0000 <arbrandes> It requires an unique "username", which at the moment is being hijacked for email. 2011-07-11 22:43:00+0000 <arbrandes> There must be a common solution to this that I'm not seeing 2011-07-11 22:43:28+0000 <arbrandes> I have to dig in a bit deeper into the API, and look at what the available plugins do about it. 2011-07-11 22:44:29+0000 <arbrandes> Maybe there's an unique Facebook identifier to be used as "username", that would solve it. 2011-07-11 22:44:42+0000 <arbrandes> (in addition to email, I mean) 2011-07-11 22:45:12+0000 <dachary> hum 2011-07-11 22:45:28+0000 <arbrandes> well yeah, there it is: "user id", heheheh 2011-07-11 22:45:34+0000 <dachary> you have a numerical identifier in addition to the email in your current registration ? 2011-07-11 22:45:45+0000 * dachary checking 2011-07-11 22:46:04+0000 <arbrandes> dachary, only the autoincrement for id 2011-07-11 22:48:24+0000 <dachary> and facebook will provide you with a numerical id too, right ? 2011-07-11 22:52:37+0000 <arbrandes> https://developers.facebook.com/docs/reference/api/user/ 2011-07-11 22:52:50+0000 <arbrandes> It says "string", but its possible it's actually an integer. 2011-07-11 22:52:59+0000 * arbrandes trying to find out 2011-07-11 22:54:01+0000 <arbrandes> yup, numerical 2011-07-11 22:54:21+0000 <arbrandes> Pretty neat simulator, too: https://developers.facebook.com/tools/explorer/?method=GET&path=me 2011-07-11 22:56:35+0000 <arbrandes> I see there's a "username", too. If all user's have usernames, then we can use that, and problem solved. 2011-07-11 22:58:29+0000 <dachary> :-) 2011-07-11 22:59:11+0000 <dachary> well, as long as you have a unique numerical id you don't have worry too much about duplicate user names, do yoU ? 2011-07-11 22:59:38+0000 <arbrandes> Oops, not everybody has usernames. The numerical ID will have to be the key. 2011-07-11 22:59:48+0000 <arbrandes> Yup 2011-07-11 23:06:00+0000 <dachary> it's going to be a ride ;-) 2011-07-11 23:06:24+0000 <dachary> I'll get some sleep now, better be fresh and ready for tomorrow. GNU GPLv3 translation most of the day. 2011-07-11 23:06:27+0000 <arbrandes> Oh, definitely! ;) 2011-07-11 23:06:39+0000 <arbrandes> Ok dachary, thanks once again 2011-07-11 23:06:46+0000 <arbrandes> Have a good night! 2011-07-11 23:06:47+0000 <dachary> thank you 2011-07-11 23:06:51+0000 <dachary> good night :-) _______________________________________________ Farsides mailing list - [email protected] Wiki: http://farsides.com/ List: http://farsides.com/ml/ Forum: http://farsides.com/forum/ Ideas: http://farsides.com/ideas/ Chat: http://farsides.com/chat/
