Meetings/2008-10-29

= Meeting 2008-10-29 =

Details

 * When: Wednesday 29th October 2008 from 1900 GMT
 * Where: #okfn irc channel on oftc.net (connect via Mibbit)

Agenda

 * Microfacts
 * COMMUNIA Workshop in Amsterdam (post-event update)
 * OKCon
 * Workshops
 * CKAN
 * Microfacts

Participants

 * Jonathan Gray
 * Rufus Pollock
 * Michael Chelen
 * Joshua Gay
 * Matthew Lee

IRC Log
#!irc 19:03:14 *	Now talking on #okfn 19:03:43 	hi all 19:04:25 	hello 19:05:13 *	rgrp (~rgrp@cpc3-cmbg6-0-0-cust206.cmbg.cable.ntl.com) has joined #okfn 19:06:22 	hi everyone 19:06:35 	hi rgrp 19:07:21 	hi mchelen 19:07:31 	how is everyone doing? 19:07:53 	thanks for announcing rgrp: i've been zipping around a bit... 19:08:53 	no problem, i know you have been busy :) 19:09:00 	i'm doing ok 19:09:21 	rgrp: just went over open everything and its all looking ready... 19:09:40 	good thanks, how about you? 19:09:52 	jwyg: good to hear 19:09:55 	if its ok you'll be in conversation with charles leadbeater and glyn moody for a bit after you've all talked... 19:10:17 	which i hope will be fun! :-) 19:11:02 	sounds cool 19:11:08 	mchelen: good to be back in uk. where are you based again? 19:12:06 	rgrp: one thing i was wondering, in ckan, is there any way that once can do faceted browsing of tags? 19:12:15 	e.g. 'uk' + 'gov'? 19:12:22 	or 'uk' + economics'? 19:13:38 	not at present but it would not be hard to implement basic faceted browsing 19:13:46 	i.e. anding 19:14:06 	i'm going to try doing some ckan stuff this week and i could implement this if necessary 19:15:16 	that would be fantastic! 19:15:39 	i think as there are more packages this could be really useful 19:15:39 	what exactly would you have in mind? 19:15:48 	what would be optimal search approach 19:15:54 	jwyg, cool, where have you been? in east coast US, which makes figuring timezones a bit difficult :D 19:16:19 	rgrp: well at the back of my mind has always been something like tag clouds 19:16:21 	rather than adding this to tag search i'm guessing we want to implement this on package side of things 19:16:25 	? 19:16:30 	not quite sure what this would mean here 19:16:38 	tag clouds are visualization right ... 19:16:42 	so a weighted representation of different package tags 19:17:02 	more frequent = bigger 19:17:26 	and the possibility to select multiple tags to narrow search.. 19:17:31 	(not fully thought through) 19:17:45 	this is more on the visualization side of things though 19:17:51 	true.. 19:18:05 	i guess there are 2 things.. 19:18:11 	1 would be the visualistion.. 19:18:20 	but in terms of browsing being able to select 'uk' 19:18:22 	we already pull out info about number of packages per tag 19:18:31 	then narrow by selecting, e.g., gov 19:18:43 	or statistics... 19:19:20 	what language is being used? there are some good viz libraries available 19:19:28 	new ckan upgrade looks really smart by the way :-) 19:19:48 	mchelen: been in berlin, then meeting in amsterdam.. now back in london.. 19:19:56 	mchelen: we are using python as backend, frontend stuff is plain old html + css + some javascript (for tag autocompletion) 19:21:27 	perhaps we can speak to liz about this when she's here.. 19:21:41 	she also really wants to speak to you more about microfacts.. 19:22:13 	that would be good 19:22:30 	there's some tag cloud code here: http://chir.ag/phernalia/preztags/ 19:22:31 	rgrp, for tag clouds probably some javascript or basic python will work, python imaging library can do more advanced images too 19:22:46 	jwyg, cool sounds interesting, hope they went well! 19:23:25 	mchelen: http://blog.okfn.org/2008/10/22/third-communia-workshop-marking-the-public-domain/ 19:23:55 	mchelen: it was great! :-) 19:24:08 	mchelen: do you know any good js libraries for tag clouds 19:24:21 	i think the idea here is to do something 'live' or interactive in which case we will need js 19:25:50 	jwyg, that looks great, public domain is kind of the perfect form of open source. what were the results of the conference? 19:27:08 	mchelen: i was really pleased that there is a new working group, which will include preliminary work on public domain calculators... 19:27:10 	rgrp: I didn't know you wrote a markdown2latex python script. That's great. I created my own little hacked up markdown to latext setup (in perl) to create the FSF bulletin. 19:27:16 *	joshuagay is now known as jgay 19:27:41 	mchelen: http://okfn.org/wiki/PublicDomainCalculators 19:28:13 	jgay: yep that was me 19:28:19 	hey jgay! :-) 19:28:24 	hey jwyg 19:29:41 	rgrp: speaking of latex, I've been in love with http://modtex.org/ + http://wikitex.org (see http://wikisophia.org/wiki/Wikitex for example use) 19:30:01 	due to a mistaken decision to 'write-my-own' mini text-langugae (basically a minor extension of markdown + asciimathml I needed to convert a load of academic papers from html to latex at one point so ... 19:30:11 	jgay: do you think it could be good to touch base with fsf europe people regarding copyright lawyers for http://okfn.org/wiki/PublicDomainCalculators ? 19:30:11 	rgrp, maybe this jQuery plugin would work? http://plugins.jquery.com/project/dynacloud 19:32:56 	mchelen: looks good 19:33:25 	unfortunately we're mainly using yui and/or mootools 19:33:34 	rgrp, http://code.google.com/p/tagcloud/ looks similar, also html can still be pretty interactive with python to serve it up :) 19:33:39 	(too many js libs ...) 19:34:34 	haha yeah, well maybe there is something like that available from mootools or yui? 19:35:21 	jgay: looks interesting though I have to say once you start using a lot of latex why not just write latex directly .... 19:35:49 	the only benefits of wiki markup are if you do a lot of ordered lists ;) 19:37:07 	i'm talking here about reasonably size bits of texts ... 19:37:18 	i'd always prefere to write my blog posts in e.g. markdown 19:39:24 	rgrp: Here is an example. Check out flexbook.ck12.org -- we have a wiki import, where authors can write the books in mediawiki markup and then import them into the flexbook ssytem and then send them to print. 19:39:43 	jwyg, that pd calculators project looks interesting, where is the development of the code portion focused? 19:40:22 	rgrp: so, one vision I have is imagine a nice mediawiki set-up. Authors write books using various aspects of mediawiki, e.g., semantic mediawiki extension. You get the power of that, but, then you want to be able to go to high-quality print at times. 19:41:02 	jgay: no it's a great idea though my point i suppose I'm saying is why start out with mediawiki then 19:41:09 	sure if people have already written stuff like go for it 19:41:16 	but why not just wikify latex stuff ...? 19:41:32 	i guess problem maybe with real time rendering? 19:41:47 	(by wikify i mean version + put on a website ...) 19:44:01 	wiki markup is probably easier for people to write than latex, although latex is quite powerful 19:44:07 	(sorry to correct my terrible grammar here that should have read: sure, if people have already written stuff in MW then you need to convert but if not why not begin in latex or something close?) 19:44:46 	mchelen: i guess that is the simple answer though once you are adding the fancy maths or chemical diagrams or ..., is that still the case? 19:45:25 	what's normally a hassle with latex is getting it 'compiling' plus its intolerance of syntax errors 19:47:57 	saying all this i realize that jgay is really correct here 19:48:03 *	mattl (~mattl@terminus-est.gnu.org) has joined #okfn 19:48:21 	we definitely do need some good way of compiling from  to latex (or any other print system) 19:48:33 	hey rufus 19:49:58 	hey mattl! 19:50:21 	is it mandatory to have a four letter nick in here? 19:51:14 	hope not.. :-) 19:52:02 	alright... 19:52:14 	i had better shoot... 19:52:29 	rgrp: I think that the issue is that latex has a steep learning curve and it is explicitly designed for print formatting. Also, it's easier to go from some markdown or RDF or something else into latex than it is to go the other way around. 19:53:08 	rgrp, equations need more than standard wiki markup, and its easier for a broader group if both are available =) 19:53:12 	Also, in the long-term, i don't think it's "latex" that we want. I think that ultimately we want TeX and Metafont rendering, but, it would be nice if we had a concerted effort of building a new interpreter on top of that, with a more modern syntax. 19:53:54 	if latex is generated from the input wiki markup, then it doesn't limit the user in any way 19:54:35 	mchelen: sure, but, ultimately we are continuing to encourage the creation of wiki stylesheet after style sheet, with people constantly reinventing the wheel. It woudl be nicer to have a better library/API set-up. 19:54:40 	jgay: yes, I'd already argued myself round to that point of view :) (though see http://plastex.sourceforge.net/ for an example of a system that parses latex back into whatever you want) 19:56:02 	i agree that we don't really want "latex". What would have been great would have been to use HTML + mathml + svg etc as a base and then build on top of that (be that with wiki languages, asciimathml etc etc) 19:56:20 	rgrp: WOAH! 19:56:27 	unfortunaely css3 with its better print support has been a long time coming and the font support for mathml is still poor (stix has finally come out i think) 19:56:34 	plastex looks great (on the surface). I'm excited 19:57:14 	i have only used it a bit but it seemed pretty good 19:57:27 	better head off... rgrp: speak soon! 19:57:30 	bye all! 19:57:30 	jgay, that would superb, what is the markup language that would address those issues? 19:57:34 	rgrp: another point is that one of the strongest features of latex is also one of its crutches. Fonts don't scale linearly. 19:57:38 	jwyg: before you go 19:57:42 	rgrp: nods 19:57:59 	so, for instance, say I want to create an SVG file as my output. 19:58:00 *	t__ has quit (Ping timeout: 480 seconds) 19:58:03 	Right now I have to render DVI to SVG. 19:58:29 	So, automatically I've created a lossy format 19:58:42 	is HTML a useful way to write books though? 19:58:45 	It's funny to think that SVG is lossy 19:59:37 	It's because SVG just does a linear scaling, whereas Metafont doesn't do linear scaling. 19:59:45 	rgrp, maybe latex or html would be accessible through a good gui? and let someone hand edit the code if they choose 20:01:09 	ideally we would have some sort of transformation language that understands how to correctly render a vector graphic for typesetting at different scales. 20:01:21 	mattl: i don't think html per se is a good way to write books but it would be a great base layer 20:01:35 	over the top of that you put you gui/easy markup language etc 20:01:45 	i think this would be preferable to latex in many ways 20:02:12 	Markdown seems like a good way to do it. 20:02:13 	but, to do that right now we would have to essentially implement metafont/tex in something like xslt. . . . I imagine most programmers would rather become farmers before accepting such a task :-) 20:02:33 	jwyg: check out new look of http://www.microfacts.org/ (just this minute updated ...) 20:04:47 	mattl: yes i agree, though you do need e.g. mathml for maths (or rather something like latexmathml, asciimathml on top of mathml) 20:05:23 	meh. ignore maths. nothing good ever came of it. 20:05:23 	rgrp: hey that's great. I have two overlapping projects I'm starting to pick up again. one is a problems and solutions collection. The other is a timeline project. Where people can combine various facts, images, citations, etc, to produce timelines on various themes. 20:05:49 	rgrp: all of which are projects to help me explore the idea of Annotations better. 20:06:52 	mattl: forget math. How about chemistry, chess or Go, diagrams, plotting, and music scores 20:07:05 	mattl: scroll through http://wikisophia.org/wiki/Wikitex 20:07:10 	jgay: annotations is a hard problem to do well it seems 20:07:13 	rgrp, how about allowing user javascript, so that people can use whatever code they like 20:07:21 *	t__ (~t@84.22.107.1) has joined #okfn 20:07:32 	mchelen: what's user javascript? 20:07:32 	i mean, they're nice, but why can't they just be SVG? 20:07:44 	mattl: because you want people to edit them inline 20:07:59 	editing a music score in SVG is ugly. Doing it in lillpond is fun. 20:08:04 	so, write a better editor. 20:08:23 	so, lillypond > svg 20:08:30 	and import the SVG into your markdown 20:08:31 	mattl: sure, but you still want the underlying representation to be pleasant. 20:08:49 	for chess games, there's pgn or whatever. 20:08:54 	So you can create a WSIWYG editor that produces lillypond that then displays as SVG or PNG or what you like 20:09:05 	right. 20:09:27 	rgrp, many wikis allow only wiki markup, by allowing users to include html or javascript then it is more flexible 20:09:49 	mattl: right, so the setup on the link i sent you brings you a lot close to having that ability. The only step is to make an easy to use javascript program to allow editing 20:09:50 	rgrp: microfacts looks great! 20:09:56 	mchelen: yup :-) 20:10:10 	rgrp: the geochrono is really coming along! 20:11:00 	rgrp: lots of the factlets are giving 500 errors 20:11:09 	yep just about to fix that :) 20:11:15 	damn changes in genshi versions ... 20:11:27 	rgrp: but, I think it's great. That is exactly the kind of set-up I was going for ... doing somethign simple with microformats. Are you also providing an RDFa of each factlet? 20:11:41 	not at present but it could be added pretty easily 20:11:51 	you can get any factlet as json via the rest api 20:12:00 	500 errors should be fixed ... 20:12:11 	rgrp: ok, perfect 20:12:29 	rgrp: yup, 500 errors are now gone 20:12:41 	really like the shakespeare scans :-) 20:12:46 	rgrp: this is really great. I'm not working at ck12.org. I'm going to present this work to them asap. 20:12:53 	(I just started working here on Monday) 20:13:46 	great :) 20:14:10 	rgrp: when can we release :-) 20:14:18 	yeah, we'd had enough of jgay at the fsf. 20:14:26 	he kept talking about books instead of software ;) 20:14:30 	:-) 20:14:30 	rgrp: one thing I'll be very interested in doing is creating references with attributes on the reference. So that factlets can be created for age level, subject, contextual usage, etc. 20:14:48 	alright.. i really better go.. i'm supposed to be pulling my weight with cooking 20:14:57 	over and out! 20:14:59 	jwyg: np, good to see you 20:15:00 	jwyg: have a good evening :-) 20:15:19 	jgay: that would definitely be possible 20:15:47 	rgrp: is there a wiki page to discuss these kinds of things? 20:15:59 	we were looking at adding arbitrary attributes and relations between factlets (i.e. rdf!) but there is a trade-off between expressiveness and ability to do useful stuff (such as do visualizations) 20:16:04 	which ones? 20:16:23 	rgrp: I can't wait to tie it into laconi.ca :-) 20:17:02 	microfacts has its main probject page at 20:17:10 	http://knowledgeforge.net/microfacts/trac/ 20:17:11 	rgrp: to discuss using factlets, design decisions, brainstorming, 20:17:16 	jgay: thanks you too! 20:17:20 	bye mattl! 20:17:23 	bye :) 20:17:24 	bye all! 20:17:25 	rgrp: great, thanks 20:17:39 	rgrp: speak tomorrow.. 20:18:28 	but might be better to put brainstorming stuff on http://okfn.org/wiki/projects/microfacts 20:18:30 	My heart delights when I see that others share an intuition for how we should share and organize knowledge. :-) 20:18:39 	:) 20:18:47 	is there a ck12.org discussion group? 20:19:39 	mchelen: not yet ... we will be finding ways to make it a more public project ... I believe part of the reason they hired me is to help them with that process. 20:20:09 	right now ck12 hires authors to write works and produce CC BY SA textbooks. 20:20:19 	yay for BY-SA. 20:20:34 	mattl: was that a bot you wrote ;-) 20:21:02 	jgay: i think ck12 should produce an electronic device that every 15 minutes reminds me that not everyone or everything is a bag of fail that i'm not the only one who cares about freely licensed creative works. 20:21:18 	jgay, an email dlist or user forum would be great to help people learn about the project 20:21:56 	mchelen: yes, I sent an email just a few minutes ago about dev, discuss, and announce lists 20:22:53 	is there some way to download the collection, other than viewing each book individually? 20:25:08 *	mattl is now known as lttam 20:25:33 *	lttam is now known as mattl 20:25:41 	was trying to evade a ban. 20:25:42 	mchelen: so, the flexbook site is pretty young and I'm not completely sure of all the functionality. 20:26:53 	mchelen: I'm honestly not sure of all of the capabilities of the flexbook system. It's primary use right now is not for editing or mixing and matching, but rather as an importer and print system. 20:27:43 	so right now it can import wikimedia formatted books and generate a printable PDF. Many of our authors are writing books in a semi-private mediawiki set-up (authors.ck12.org) and then when we go to print we have the flexbook system import the book into flexbook and print it from there. 20:28:57 	mchelen: The idea is that we have a lot we can learn from other tools, and flexbook should be designed to integrate with other production systems. So, the design strategy is to build the flexbook system in paralell with finding and improving other production and knowledge-organization systems (wikis, open office documents, etc). 20:30:18 	anyhow, I gots to get going. I need to prepare for a project meeting for us to figure out our roadmap for using LaTeX :-) Wish me luck! 20:30:25 	rgrp: Thanks again! 20:30:26 	later mattl 20:30:34 	jgay: great to talk with you 20:30:38 	mchelen: I'll send an update to the OKFN mailing list when it makes sense 20:31:19 	jgay, interesting, its good focus on getting content into the system properly, maybe more of those files could be exposed for download? 20:31:20 	we should definitely collaborate and I'd love to chat more about microfacts, annotation etc 20:31:28 	jgay, sounds good, thanks 20:31:29 	what's a microfact? 20:31:44 	jgay: (some (rather) old thoughs on annotation: http://www.rufuspollock.org/2006/12/19/web-based-annotation/) 20:32:01 	mattl: a small pice of in info (an event, a person etc) 20:32:26 	rgrp, have you heard of the firefox addon zotero? 20:32:37 	oh, you're not going to go all sem web on me, are you rgrp? 20:33:15 	mattl: no :) we didn't go down the semweb route, its plain old webapp with restful api serving json data 20:33:27 	at the same time i'd be very happy to have more rdf in there ... 20:33:44 	mchelen: yes, very much so. been using it for a couple of years now ... 20:34:11 	rgrp: have i ever told you about my fear and loathing of the term restful? its quite unreasonable. like my dislike for groklaw or fear of unknown restaurants. 20:34:12 	rgrp: yup, I have that page bookmarked from long-ago. I'll have an update soon on a new version of stet. I had a google summer of code student hack on stuff over the summer to redo stet as a JQuery extension with braoder support for microformats (specifically hReview). 20:34:41 	i also think i confused jquery and json until about a week ago. 20:34:47 	rgrp: with my new job I can put more time and energy into these kinds of projects because its part of my job descrioption. 20:35:05 	mattl: ok just call it a web api :) 20:35:40 	yay. 20:35:41 	jgay: that is great, i wanted to use stet when it was first released but was rather put off by the 'this is very alpha and probably won't work on your machine' warning 20:35:51 	hah, yeah. good old stet. 20:36:04 	jgay: does stet still require RT? 20:36:39 	mattl: the current vision of stet is a complete rewrite. It's split into two parts: 1) a way of processing annotations on the front end, 2) a way of storing annotations. 20:37:00 	rgrp, how do you like that the zotero system? requiring a browser addon is a bit cumbersome, but still works on most platforms. recent updates like server sync and sharing could be pretty interesting 20:38:17 *	jgay ponders the idea of a grease monkey latex editing system :-) 20:38:37 	it is pretty nice though i prefer to keep with bibtex (i use zotero to extract data and then export to bibtex usually) 20:38:42 	ok 20:38:45 	got to go 20:38:53 	hope to speak to people more next week or onlist :) 20:38:56 	o&o 20:38:57 *	rgrp has quit (Quit: leaving) 21:09:49 	hmm, woudl a microfact be an hFact ?? 21:10:33 	uFact 22:25:23 *	mattl has quit (Quit: Ex-Chat) 23:09:50 *	jgay has quit (Quit: Ex-Chat)