Friday, 12 April 2013< ^ >
arjan has set the subject to: Zotonic - the Erlang Content Management Framework
Room Configuration

[06:49:02] Jarimatti joins the room
[06:51:03] <Jarimatti> Good morning everyone.
[06:52:18] <Jarimatti> I've sent (or tried to send) two replies to the Zotonic users Google group but they don't show up there yet. Are they still in the moderation queue or have I missed something?
[06:56:53] Jarimatti leaves the room
[06:56:56] Jarimatti joins the room
[07:10:02] Jarimatti leaves the room
[07:53:05] arcusfelis joins the room
[08:14:16] simon.smithies joins the room
[08:32:52] simon.smithies leaves the room
[08:43:52] Jarimatti joins the room
[08:53:57] Arjan joins the room
[08:54:27] Jarimatti leaves the room
[09:00:38] <Marc Worrell> Hi Jarimatti, we have to ok all new e-mail senders. I;ve done it, they should have come through now.
[09:25:32] Maas joins the room
[09:40:32] Arjan leaves the room
[11:12:30] simon.smithies joins the room
[11:12:30] <Marc Worrell> I will meet with Francesco about their non-use of Zotonic :p
[11:17:38] <simon.smithies> Andreas: quite a bit in the log for that server of mine for yesterday and very little today
[11:18:25] <simon.smithies> Have you changed anything on it?
[11:18:43] <simon.smithies> (I haven't)
[12:15:40] simon.smithies leaves the room
[12:30:19] simon.smithies joins the room
[12:30:23] simon.smithies leaves the room
[12:30:25] simon.smithies joins the room
[13:00:05] <Marc Worrell> new tinymce in the works
[13:00:05] <Marc Worrell> http://www.tinymce.com/tryit/basic.php
[13:00:32] <Marc Worrell> though I am still very cautious about giving editors too much freedom with their layout.
[13:00:49] <Marc Worrell> that should be the job of the template - content/presentation separation!
[13:03:11] <Andreas Stenius> yeah, but it would be nice if we could work with the presentation from the web GUI too, though.... ;)
[13:14:41] <simon.smithies> I think the most important thing for the editor is to do the basic stuff really reliably and consistently. Looking at this I don't think there's a lot of new stuff, which is good ... hopefully they've just cleaned the UI and fixed more bugs. Users need to be able to do some simple semantic markup (<b> and <i> are just end user friendly substitutes for <strong> and <emph>) and get predictable results when they insert images. Leaving the rest up to the templates is the best approach.
[13:18:11] <Marc Worrell> indeed
[13:18:22] <Marc Worrell> I am even uncertain about the image stuff
[13:18:38] <Marc Worrell> what we proposed is to have blocks, which can also be a callout, or an image or whatever
[13:18:50] <Marc Worrell> that gives more control to the template in how to show the information
[13:19:33] <Marc Worrell> and I want to investigate if we can have something like "fill in the page", showing a template. A lightweight version of square
[13:20:18] <Marc Worrell> http://www.squarespace.com
[13:22:18] <Marc Worrell> their video: https://vimeo.com/45734056
[13:24:25] <simon.smithies> they have some good stuff there
[13:24:43] <Marc Worrell> yes, when we have some simple version of that then it would already be great
[13:25:08] <Marc Worrell> don't need to do everything in an interface - we have a template system for most of the stuff :)
[13:26:18] <simon.smithies> yes. all that stuff in the video is still going on within the content block, and ultimately still just rendering html like tiny mce does
[13:27:16] <simon.smithies> but looks like there's a basic block structure in place to being with, determined by another template I guess
[13:27:29] <simon.smithies> *begin with
[13:27:39] <Marc Worrell> indeed
[13:27:46] <Marc Worrell> I was thinking that we might do something along:
[13:27:53] <Marc Worrell> 1. Able to select a template
[13:28:04] <Marc Worrell> 2. Show the template, with placeholders
[13:28:12] <Marc Worrell> 3. Let an user fill in those placeholders
[13:28:23] <Marc Worrell> Like selecting images etc
[13:28:55] <simon.smithies> In many ways that approach is easier to manage than giving them a blank canvas which is what happens right now
[13:29:03] <Marc Worrell> yes
[13:29:21] <Marc Worrell> and most pages you don't want to change anything anyway
[13:29:28] <Marc Worrell> just some pages need special attention
[13:32:10] <simon.smithies> I think what you're proposing would be good step forward.
[13:32:38] <simon.smithies> A lot of work though unless there's already some kind of framework available for that
[13:33:13] <Marc Worrell> we have the blocks already
[13:33:19] <Marc Worrell> need a better way to input them
[13:33:21] <simon.smithies> I just realised that
[13:33:45] <simon.smithies> I already have templates that include multiple pages
[13:34:00] <Marc Worrell> and I want to keep the strict separation between content and presentation - otherwise you can't federate content between sites or change your site templates.
[13:34:38] <simon.smithies> good
[13:35:35] <Marc Worrell> importing wordpress content is a pain - because all that markup in the content
[13:35:52] <Marc Worrell> that markup is something we really shouldn't want
[13:36:44] Arjan joins the room
[13:37:57] <Marc Worrell> when you come across good examples of content editors, then please share them - we are looking into that area right now
[13:38:12] <simon.smithies> certainly
[13:38:26] <Marc Worrell> I have been checking full screen editors for markdown - a bit problematic on iPad, to say the least....
[13:39:01] <Marc Worrell> And tablets will be used a lot for quickly editing a web site...
[13:39:20] <simon.smithies> I think so, yes
[13:40:22] <simon.smithies> I will definitely start paying attention to editors again, because it seems to have a big impact on how easy it is to promote zotonic -- certainly with my customers
[13:40:50] <simon.smithies> I'm glad you're open to moving it forward
[13:41:34] <Marc Worrell> we are thinking of separating the admin and casual editing
[13:41:58] <Marc Worrell> so that the admin can stay a power tool - and people can more easily do the standard 80% in a simple editor
[13:42:31] <simon.smithies> and maybe the admin can opt for the simple tool when they're on their phone
[13:42:41] <Marc Worrell> indeed
[13:42:55] <Marc Worrell> that is the idea -> drop into the /admin when doing tricky stuff
[13:42:56] <simon.smithies> my problem is even my 'simple' users want to be able to do what they can do in word
[13:43:31] <Marc Worrell> and that is something we can help them with when they have a bit better feeling for the page/content
[13:43:56] <Marc Worrell> full word is not doable, too much presentation in the content
[13:44:01] <simon.smithies> once they understand the rules are different on the web, that's ok ... but any feature that looks like one from Word must work like the one in word
[13:44:19] <Marc Worrell> yes, you are right - those are a trap
[13:45:10] <Andreas Stenius> Marc, just made a quick google to see what came up. Do you already know about yui? (http://yuilibrary.com/ ) Never heard of it before, but it seems like a ton of features packed it there... (not sure if it's any good, though :p )
[13:45:10] <simon.smithies> the nice thing about squarespace approach is it seems to work quite differently, so user expectations may be different too
[13:45:29] <Andreas Stenius> (BSD licensed)
[13:45:47] <simon.smithies> a lot of fuss about http://www.firepad.io/ right now
[13:46:01] <simon.smithies> seems collaboration-focused thoough
[13:46:08] <Marc Worrell> yui is nice - kind of competing with all the other frameworks
[13:47:25] <Marc Worrell> Just checking http://socrates.io/#2s7ybuz
[13:47:30] <Marc Worrell> might be what we need
[13:47:40] <Marc Worrell> and works well on ipad
[14:01:07] <Maas> yui is used by flickr and quite a number of other sites.
[14:09:28] Arjan leaves the room
[14:28:19] Maas leaves the room
[14:31:34] Arjan joins the room
[14:31:48] <Marc Worrell> we are all here http://demo.firepad.io/#SSs7ZnFlSz now :p
[14:41:10] <Andreas Stenius> I got to go... be back later...
[14:41:17] <Marc Worrell> see you!
[14:41:25] <Marc Worrell> We had a good discussion :)
[14:41:29] <simon.smithies> sheesh I should go too
[14:41:37] <simon.smithies> we did indeed
[14:41:45] <Marc Worrell> I will wrap up and make a simple editor for end users :p
[14:41:46] <simon.smithies> cheers guys
[14:41:47] <Andreas Stenius> indeed. it's fun to see the document deveelop. Kind of like a mind map
[14:41:52] <Andreas Stenius> cheers simon.smithies
[14:42:00] <Marc Worrell> ciao!
[14:42:11] Arjan leaves the room
[14:43:19] simon.smithies leaves the room
[15:09:46] Maas joins the room
[17:03:19] Maas leaves the room
[17:04:16] Arjan joins the room
[19:05:38] Arjan leaves the room
[19:34:21] Maas joins the room
[19:35:23] Maas leaves the room
[21:19:35] Maas joins the room
[21:38:29] arcusfelis leaves the room
[21:43:50] Maas leaves the room
[21:47:25] Maas joins the room
[22:25:19] simon.smithies joins the room
[22:26:05] <Marc Worrell> Challenge: who can get one of these working smoothly in/with Erlang?
[22:26:06] <Marc Worrell> http://code.google.com/p/google-diff-match-patch/
[22:26:22] <Marc Worrell> For cooperative editing...
[22:30:23] <Maas> Interesting algorithms :-)
[22:32:06] <Marc Worrell> http://www.mobwrite.net/static/mobile.html?plop
[22:32:14] <Marc Worrell> with a simple textarea
[22:32:34] <Marc Worrell> and some explanations: https://www.youtube.com/watch?v=S2Hp_1jqpY8
[22:33:48] <Marc Worrell> the other methods need to go to great length to keep a list of all changes, and keep that synced between servers.
[22:33:52] <Marc Worrell> and clients
[22:36:05] <Marc Worrell> but i need a diff-match-patch nif or driver (guess driver is better - can take quite a bit of cpu)
[22:39:19] <Marc Worrell> The C++ version uses Qt - though I saw a patch in the forums for STL
[22:40:29] <Maas> Hmm, an erlang version would be nice too. But definitely not the strong points... maybe if it can be done with multi-core, but I guess not
[22:41:13] <Marc Worrell> I have looked into it andplayed a bit with the code - especially the patching is very complicated, using bitmaps...
[22:44:43] <Maas> Looking at python implementation. Funny, that deadline in the diff processing.
[22:45:03] <Marc Worrell> yeah, it can take a long time otherwise, it tries to find the best diff...
[22:49:20] <Marc Worrell> I like the fact that is always syncs - even on packet loss
[22:49:29] <Maas> It doesn't look like something erlang would be really good at indeed. Have you seen https://github.com/tomas-abrahamsson/tdiff?
[22:49:34] <Marc Worrell> so you can disconnect, edit and then later reconnect and sync
[22:49:44] <Marc Worrell> yeah, saw tdiff
[22:49:45] <Maas> That is a big win.
[22:49:53] <Marc Worrell> i think we need a driver...
[22:50:17] <Marc Worrell> or a nif and hope the best :p
[22:50:26] <Maas> Crash and burn
[22:50:29] <Marc Worrell> (short timeouts)
[22:53:10] <Maas> Well for nifs you can also start a separate thread. That is how I do it with sqlite.
[22:53:21] <Marc Worrell> that is good enough
[22:53:25] <Maas> but this is different
[22:53:40] <Maas> if you have lots of parallel sessions
[22:53:52] <Marc Worrell> is ok, just queue them all :p
[22:54:13] <Marc Worrell> most diffs/patches are < 1-10msec
[22:54:41] <Maas> having the algorithms in erlang is a win with lots of sessions.
[22:55:05] <Marc Worrell> true - but will be hard to get it working…. and fast enough on <12 core systems
[22:55:06] <Maas> not sure what you are aiming for... first proof of concept i think
[22:56:46] <Marc Worrell> and from there production :p
[23:03:55] <Marc Worrell> dat textarea ding werkt geweldig :)
[23:13:51] <Marc Worrell> Time to sleep :)
[23:14:24] <Marc Worrell> Issue 25 - google-diff-match-patch - QT Library? - Diff, Match and Patch libraries for Plain Text - Google Project Hosting <http://code.google.com/p/google-diff-match-patch/issues/detail?id=25>
[23:14:33] <Maas> indeed, truste..
[23:14:34] <Marc Worrell> http://code.google.com/p/google-diff-match-patch/issues/detail?id=25
[23:14:37] <simon.smithies> good night
[23:14:48] <Marc Worrell> thanks :) you too :)
[23:15:14] <simon.smithies> in a few hours ... eating breakfast right now :P
[23:16:11] <Marc Worrell> haha :)
[23:16:15] <Marc Worrell> middle of the night here
[23:16:39] <simon.smithies> Sounds like you were very productive while I was asleep
[23:16:50] <simon.smithies> I'll let you go :-)
[23:16:54] <Maas> multi timezone chat i love it
[23:31:35] <simon.smithies> question for anyone still up ... am going to upgrade erlang on that server of mine. Clutching at straws really, but who knows maybe it'll make the restart problem go away
[23:31:47] <simon.smithies> what version do you recommend?
[23:36:44] <Maas> Still running R14B03 on my systems... Wanted to look at upgrading this weekend.
[23:38:33] <simon.smithies> Zotonic is not yet ready for R16 - is that right?
[23:40:10] <Maas> The latest should be I think. Marc removed all parameterized modules a couple of weeks ago.
[23:42:10] <simon.smithies> Given I have problems already, I guess i'm looking for the safest option
[23:42:28] <Maas> What are you running now?
[23:43:04] <Maas> Wanted to go for the latest R15 release myself.
[23:44:39] Andreas Stenius leaves the room
[23:45:04] <simon.smithies> R15B01
[23:45:30] <Maas> R15B03-1. Mainly for the fixes in erlang http client code. R14 is doing fine.
[23:46:08] <simon.smithies> R14 is good for me too, on the old server
[23:47:11] <simon.smithies> does the -1 in R15B03-1 mean it's a beta or pre-release?
[23:48:11] <Maas> There was a small bug in the B03 release and they patched it. It is a strange number indeed.
[23:48:41] <Maas> It is a real release.
[23:49:01] <simon.smithies> I see. So it looks like the one to go for then.
[23:50:42] <simon.smithies> and you're happy with the erlang solutions builds?
[23:52:02] <Maas> Haven't tried it yet, but I prepared a vm for testing this weekend. Usually I compile my own.
[23:54:12] <simon.smithies> ok
[23:56:51] <Maas> So I have to see, I guess it should be allright, although I don't really know what compile flags they used.
[23:58:43] <Maas> For smaller vm's it is better to take the 32 bit builds
[23:59:14] <simon.smithies> even on a 64 bit linux?
Powered by ejabberd Powered by Erlang Valid XHTML 1.0 Transitional Valid CSS!