- From: Wendy A Chisholm <wendy@w3.org>
- Date: Fri, 01 Feb 2002 13:26:14 -0500
- To: w3c-wai-er-ig@w3.org
[18:29] <nick_kew> Check For Fixes Using EARL? [18:30] <sbp> yeah [18:30] <sbp> it probably makes more sense to check for breakages. Just reverse "passes" and "fails" in the test case [18:31] <sbp> Hmm... no, I like to check for things being fixed too. If I've issued a bug report, I want to know when it's been dealt with [18:31] <nick_kew> haven't had time to read it yet, but looks interesting [18:32] <sbp> it's just a very simple example of what should be possible. I think that the current corpus of EARL examples is a bit weak [18:32] <nick_kew> reading it now [18:33] <nick_kew> Ah, I see. That could be acompanied by document hashes to indicate what's changed [18:33] <sbp> of course [18:33] <nick_kew> (but I guess that's tool-specific) [18:35] <sbp> yes. And it depends on what you're evaluating - in this case, it's quite clear that the tool is "A-Prompt". No need to hash the source code, although that might come in handy if there were a new version every hour or so, with no version numbers, and mirroring lag [18:35] <nick_kew> Put that in a database, and add notification functions, and it becomes a full-blown tool [18:35] <sbp> :-) [18:35] <nick_kew> OK, Valet is evaluating webpages [18:36] <nick_kew> - and the hashes serve as a structural metric [18:37] <sbp> how are you hashing pages? Are you RegExping them, or using an HTML parser of some kind? [18:37] <sbp> I decided to just RegExp them (examples in the EARL API), but I was told to use an HTML parser. I think using a RegExp is better, because it's less programming language dependent [18:37] <nick_kew> grep on nsgmls output [18:38] <sbp> oh. A RegExp of sorts, then [18:39] <nick_kew> - that's particularly easy, because each line of output is an element/attribute/etc (the first char tells you which), so it's trivial to get the structure [18:40] <sbp> yep. and nsgmls is quite popular, which brings a bit of repurposability to the code <wendy> agenda+ nick's new page valet <wendy> agenda+ sean's earl api <wendy> agenda+ comments from EARL tech review <wendy> agenda+ tools review from ATRC/Chris Ridpath <wendy> zakim, agenda? [19:03] <Zakim> I see 6 items on the agenda: [19:03] <Zakim> 1. learn how to use Zakim [19:03] <Zakim> 2. get xena in here [19:03] <Zakim> 3. nick's new page valet [19:03] <Zakim> 4. sean's earl api [19:04] <Zakim> 5. comments from EARL tech review [19:04] <Zakim> 6. tools review from ATRC/Chris Ridpath <wendy> so, y'all ready to discuss this stuff? [19:05] <nick_kew> yep. [19:09] <sbp> thanks for hash.pl <wendy> agenda- 1 [19:09] * Zakim notes agendum 1, learn how to use Zakim, dropped [19:10] <nick_kew> Yeah .. I've explained it to several people; thought posting source easier [19:10] <JibberJim> yeah nick_kew - now if I can just find Apache, onsgmls and a box to run them all on :-) [19:10] <nick_kew> They'll all run on windoze [19:11] <JibberJim> I know... [19:11] <JibberJim> I run them all, but not together. <wendy> nick, was there more about page valet that you wanted to talk about? <wendy> i just checked it out again. i'm glad to see that the wcag errors now have a text indication of priority level. <wendy> although, WCAG1, WCAG2 might get confused with WCAG 1.0 and WCAG 2.0 instead of WCAG priorities. [19:14] <nick_kew> Well, it's under active develoipment, and wants review/feedback [19:14] <JibberJim> I confused that straight away Wendy! [19:14] <nick_kew> :-) <wendy> there's my feedback. jim or sean, did you have further feedback for nick? [19:14] <JibberJim> I think he's had all mine. [19:15] <sbp> I've already talked with Nick about it - it's very neat, and I only had some layout comments [19:15] <JibberJim> I know I've got to work on the WYSIWYG version. [19:15] <sbp> oh, and I mumbled something about XSLT to EARL <wendy> why's the EARL gone from http://valet.webthing.com/xslt/ <wendy> ? [19:15] <nick_kew> Mainly I'd like to get some of you folks to review it seriously [19:15] <nick_kew> wendy, follow the link to the stable version * wendy follows [19:16] <JibberJim> My main concerns are the number of matches you get on well authored pages. [19:16] <JibberJim> (warnings about scripts, css etc.) [19:16] <sbp> oh, why is the XSLT output not XHTML? [19:17] <nick_kew> Indeed. In principle I'd like to offer different levels of fussiness, but still thinking through how to do that [19:17] <sbp> in fact, *how* is it not XHTML? Unless you're using <xsl:text>? [19:17] <nick_kew> sbp, which view are you using? [19:18] *** sbp has quit IRC (Connection reset by peer) [19:18] <nick_kew> oops <wendy> the EARL is interesting. [19:18] *** sbp (~sean@m66-mp1-cvx4c.pop.ntl.com) has joined #er [19:18] <JibberJim> and I'm the one with the modem lead taped to the computer. [19:18] <nick_kew> That's only the old prototype <wendy> why is it this way: <wendy> _:mid6 <wendy> rdf:subject _:mid6s ; <wendy> rdf:predicate earl:fails; <wendy> rdf:object _:mid6o . <wendy> _:mid6s <wendy> earl:testCase <http://www.cnn.com/> ; <wendy> earl:date "Thu, 31 Jan 2002 00:16:30 GMT" ; <wendy> sp:line "72" ; sp:column "100" . <wendy> _:mid6o <wendy> earl:expectedResult """ <wendy> required attribute "TYPE" not specified <wendy> """; <wendy> earl:id sp:Validation . <wendy> --- <wendy> ? [19:21] <nick_kew> wendy, I can't remember all that well - long time ago I put that up <wendy> ok. sbp's question about xhtml? <wendy> sbp - which view are you using? [19:22] <sbp> ignore me <wendy> ok. [19:22] <nick_kew> But it's a prototype: I'm open to further suggestions. Any updates now will be based on visval and include accessibility stuff <wendy> so, visval will generate HTML, XML, or EARL? <wendy> (in the future)? [19:23] <nick_kew> If there's demand for EARL :-) * wendy drums up a chorus, "we want EARL! we want EARL!" [19:23] <JibberJim> Wendy seems to be alone here ... [19:24] <nick_kew> Wendy, can you pull any strings to get it reviewed and/or recommended where people will notice it? <wendy> no, really, there's a whole chorus. [19:24] <sbp> sorry, I was listening to Leadbelly - would have chanted "keep EARL from the Gallows Pole" <wendy> nick, if it generates EARL, I will promote it's use. <wendy> I'm talking with three other developers who also want to generate EARL [19:25] <nick_kew> OK, you're on. EARL by the weekend. <wendy> I will point them to that. [19:25] * JibberJim notices Nick didn't say which weekend. <wendy> At the tech plenary, I'll be pushing earl, if page valet generates it, I can use that as a demonstration. [19:25] <nick_kew> EARL/XML or EARL/N3 ? [19:25] <JibberJim> XML! <wendy> jim - you had done some work with annotea and earl, right? [19:25] <nick_kew> - or is that immaterial <wendy> i'm heading to MIT to meet w/annotea folks re: earl annotations [19:26] <JibberJim> I thought we agreed on this at a previous meeting [19:26] <sbp> EARL/NT! [19:26] <nick_kew> ? <wendy> (refer to the scenario at http://www.w3.org/2001/08/AnnoteaOxygenDemo.html) <wendy> yes, XML. [19:26] * sbp notes that EricP is going to follow up on www-annotation with a thread about EARL [19:26] <JibberJim> I've looked at both Wendy, I never got them working together, and my Annotea client broke with a recent protocol change it's in my to do list. <wendy> re: ericp - cool. [19:27] * sbp wonders if CWM can be made to POST <wendy> jim - any chance we could try some stuff out in the next couple of weeks? what can i do to help? [19:27] <nick_kew> Future plans certainly include EARL, when it permits a user to feed it a spec against which to test [19:28] <nick_kew> (eg to support XAG) <wendy> nick - cool! [19:28] <nick_kew> - discussed that with sbp a day or two ago [19:28] <sbp> and for that, we need to work on the TPDL aspect of thing [19:28] <JibberJim> Well I need to get my Annotation client back working, I'm not sure of the difficulty that was a hack which broke, and now I need to be more real RDF which is good, but more time consuming. The EARL portions if I can get that working should be relatively easy. [19:28] <sbp> s/thing/things/ [19:29] <JibberJim> As to things you can do to help Wendy, is come do my day job :-/ [19:29] <nick_kew> :-) <wendy> :) <wendy> i haven't seen your annotation client, jim. did you have anything out and about for people to play with? [19:30] <nick_kew> (wendy lost her nose?) [19:30] <JibberJim> It was part of my web browser http://jibbering.com/snufkin/ <wendy> yes, i'm actually a frog. * wendy checks out snufkin [19:30] <nick_kew> hmm, I must have been hallucinating at Brighton <wendy> jim - i'll check out snufkin tomorrow. [19:32] <nick_kew> hi aaron. Can you bring in xena? <wendy> er, unless you need to work out some bugs w/annotea first? <wendy> i mean the annotea client. [19:32] <AaronSw> hi nick. let me take a look <wendy> agenda? [19:33] * Zakim sees 5 items on the agenda: [19:33] * Zakim 1. get xena in here [19:33] * Zakim 2. nick's new page valet [19:33] * Zakim 3. sean's earl api [19:33] * Zakim 4. comments from EARL tech review [19:33] * Zakim 5. tools review from ATRC/Chris Ridpath [19:33] <JibberJim> It won't have an annotea client tomorrow, I'll look to do the update when I get the chance. <wendy> ok. jim. <wendy> agenda- 2 [19:33] * Zakim notes agendum 2, nick's new page valet, dropped <wendy> sean's earl api? anyone played with it? <wendy> i've still got to install python 2.2 and play with it more. [19:34] <AaronSw> he hasn't told me the location of the secret plans... i mean source [19:34] <sbp> www-archive. Hardly secret! [19:35] <AaronSw> see, he didn't tell me! <wendy> nick - have you looked at the api? does it look like you could use that to generate earl for page valet? [19:35] <sbp> you're subscribed! [19:35] <AaronSw> I don't read email anymore. <wendy> also, sbp you had made several suggested changes to earl 1.0 [19:35] <nick_kew> More likely elsewhere in Site Valet <wendy> anyone have reactions to those? [19:35] <nick_kew> But I haven't "done" the Python learning curve [19:35] <sbp> but we haven't even made EARL 1.0 yet :-) <wendy> :) [19:37] <nick_kew> It's on my TBD list, but for Valet I'll want both C++ and Perl versions :-( <wendy> sbp what are the "other less-used features of EARL" <wendy> ? [19:38] <nick_kew> sbp, keep nagging me every few days, and I'll get around to it [19:38] <sbp> stuff like earl:purpose [19:38] <sbp> Nick: will do [19:38] <nick_kew> - eventually [19:39] <sbp> you'll have to nag me to nag you <wendy> what does: "loose note" and a "normative definition" set of <wendy> properties <wendy> mean? [19:39] <AaronSw> Zakim, nag nick_kew every few days [19:39] <Zakim> I don't understand 'nag nick_kew every few days', AaronSw. Try /msg Zakim help [19:39] <sbp> well, rdfs:comment seems to be used for two main things at the moment [19:40] <sbp> * as a loose note "here's some crap about the property", * as a definition "this is what this proeprty is, and is for" <wendy> ah, ok. [19:40] <sbp> and I'm sick of splurshing them into one property object value under the guise of rdfs:comment [19:40] <sbp> so I decided that two seperate properties would be a neat idea [19:40] <sbp> then you can do a string:concat to form an rdfs:comment [19:41] <AaronSw> agenda? [19:41] * Zakim sees 4 items on the agenda: [19:41] * Zakim 1. get xena in here [19:41] * Zakim 2. sean's earl api [19:41] * Zakim 3. comments from EARL tech review <wendy> does definition get into the realm of the TPDL? [19:41] * Zakim 4. tools review from ATRC/Chris Ridpath [19:41] <AaronSw> agenda- 1 [19:41] * Zakim notes agendum 1, get xena in here, dropped [19:41] <sbp> although if you declare them as sub properties of rdfs:comment, you might get some odd results [19:41] <sbp> that's what the "D" stands for <wendy> :) <wendy> right, so that might be something that would be dropped as well? <wendy> then comment could be just for the loose stuff? [19:42] <sbp> erm... [19:42] <sbp> TPDL defines test cases, not RDF properties <wendy> or, is this a general issue you have with rdfs:comment - that it gets overused other places as well? [19:42] <sbp> yep, it's a general issue [19:42] <sbp> I've wanted something like this for a while, for almost every schema that I hack up [19:42] <sbp> it might even be on the RDF issues list <wendy> right, the D should be in TPDL not EARL - therefore drop it from earl is where it sounded like u were heading. [19:43] <sbp> I don't want n:definition added to EARL - it's purely a documentation property [19:43] <sbp> in fact, it belongs in RDFS [19:43] <sbp> so RDF Core should deal with it <wendy> what about earl:excludes - you say to remove it b/c it is shorthand for enumerating text point IDs and that it is easy to do w/CWM. <wendy> However, it seems like a really useful shorthand. <wendy> although, you say "applying them is difficult" <wendy> unless you know how many to exclude. [19:45] <sbp> yes. I have a test case that I've been working on [19:45] <sbp> I end up with... [19:45] <nick_kew> erm, why is CWM relevant to EARL spec? [19:45] <nick_kew> or rahter "easy to do with CWM" <wendy> good q [19:46] <sbp> I'll answer that in a sec., hang on [19:46] <sbp> [[[ [19:46] <sbp> <http://example.org/2001/05/bob#WCAG10P1> a :TestCase; [19:46] <sbp> :id wcag:tech-alt-pages, [19:46] <sbp> wcag:tech-auditory-descriptions, [19:46] <sbp> wcag:tech-avoid-flicker, [19:46] <sbp> [...] [19:46] <sbp> wcag:tech-redundant-server-links, [19:46] <sbp> wcag:tech-scripts, [19:46] <sbp> wcag:tech-simple-and-straightforward, [19:46] <sbp> wcag:tech-synchronize-equivalents, [19:46] <sbp> wcag:tech-table-headers, [19:46] <sbp> wcag:tech-table-structure, [19:46] <sbp> wcag:tech-text-equivalent; [19:46] <sbp> :testCriteria [ [19:46] <sbp> [...] [19:46] <sbp> :excludesId wcag:tech-table-headers, [19:46] <sbp> wcag:tech-table-structure, [19:46] <sbp> wcag:tech-text-equivalent ] . [19:46] <sbp> ]]] [19:47] <sbp> CWM is relevant because it is a good implementation of FOPL over RDF, and representative of the type of querying functionality that RDF gives you <wendy> one of those "you get it free from rdf and rdf imps, therefore don't sweat it in your lang" type of answers? [19:48] <sbp> one of my "I love CWM" type of answers <wendy> :) [19:48] <nick_kew> which is all very well in a world of good RDF imps <wendy> :) <wendy> ribbit [19:48] <sbp> yes. EARL APIs should be able to work around it [19:49] <sbp> but we'll see. I haven't got around to that in the Python EARL API yet [19:49] <sbp> It needs a name. Ribbit is as good as anything <wendy> b/c we have earl:testcase that shows what it does conform to, right? [19:49] <nick_kew> visval is C++. As I see it, using EARL there will imply rather more groundwork <wendy> ribbit: lol [19:50] <sbp> well, I managed to work out something pretty quickly in Python. There's not much code there <wendy> it depends on what visval is going to do. if visval is going to. What if it generates EARL, but another tool (written in cwm) does the reporting of the results? [19:51] <sbp> as long as you can handle a bit of list programming, and have some semblance of an RDF API, it shouldn't be much of a problem. And perhaps there are C++ APIs out there that you can build on [19:51] <sbp> .google C++ RDF API [19:51] <xena> C++ RDF API: http://nestroy.wi-inf.uni-essen.de/rdf/sum_rdf_api [19:51] <nick_kew> I'm struggling with the fact that OpenSP uses a homebrwe String type (to deal with big chars - i18n), and there's no regexp library for it:-( [19:51] <sbp> no RegExp library? Eek [19:51] <nick_kew> precisely [19:52] <JibberJim> There must be an i18n regexp library [19:53] <nick_kew> Can't find one. I *think* that's why Java's so popular for XML - better libraries support [19:53] <sbp> but Wendy's right: if you're only outputting for the moment... [19:53] <nick_kew> (native 16-bit char) [19:53] <nick_kew> Yeah, outputting easy <wendy> nick or sbp - have tied hash to anything? <wendy> s/have tied/have you tied [19:55] <sbp> pardon? <wendy> the hash hack exists, but is it tied in with anything? i.e. page valet? [19:56] <nick_kew> Hash, yes, Valet's problem reporting+management database [19:56] <sbp> Ribbit has it, but doesn't use it as yet <wendy> :) [19:57] <nick_kew> When you click on the Valet logo, and select "comment on this page" <wendy> ok. i'll have to play w/that. <wendy> agenda? [19:57] * Zakim sees 3 items on the agenda: [19:57] * Zakim 1. sean's earl api [19:57] * Zakim 2. comments from EARL tech review [19:58] * Zakim 3. tools review from ATRC/Chris Ridpath <wendy> any further discussion about sbp's api? <wendy> sbp - you've more stuff you're adding to it? [19:58] <sbp> yeah <wendy> when do you expect to release a new version? [19:58] <sbp> indefinite <wendy> ok. <wendy> agenda- 1 [19:58] * Zakim notes agendum 1, sean's earl api, dropped <wendy> comments on EARL tech review - jim responded re: cc/pp. <wendy> i don't have a good answer. anyone else? [19:59] <JibberJim> Pet topic of mine that. <wendy> :) [19:59] <sbp> what, that he thinks CC/PP sucks? does that need an answer? Seems like a statement rather than a question... :-) <wendy> do you agree that cc/pp sucks? [20:00] <sbp> no comment [20:00] <nick_kew> Is it really anything more than a fancy name for HTTP content negotiation? [20:01] <JibberJim> I don't think it sucks, I can see it being useful in the mobile world for distributing content, it's just not for accessibility ... [20:01] <nick_kew> (or what should be HTTP, if you follow KISS) [20:01] <sbp> agreed <wendy> EARL is more than accessibility. <wendy> it's stating if something conforms to a spec or not. <wendy> e.g. page a is valid XHTML. [20:02] <JibberJim> True, but it's not what CC/PP is useful for. <wendy> therefore, if partnered with cc/pp, it could be deduced that page a could be used on the following devices... [20:02] <sbp> CC/PP is just a bit that you may conceivably add onto the Test Subject to say where you got it from <wendy> or something like that. [20:02] <sbp> that's it <wendy> onto the testSubject? [20:03] <sbp> TestSubject, yep [20:03] <sbp> and you can go from there. As you say, we don't stop people from using EARL however they like <wendy> what about the other way around - someone has cc/pp preferences set, one of those is an EARL assertion about the types of content they want. [20:04] <JibberJim> Absolutely if people want to use CC/PP in earl, good luck to them. [20:04] <sbp> how would have an EARL assertion about the types of content they want? isn't that what CC/PP is meant to do? [20:04] <sbp> s/would/would you/ [20:05] <sbp> er... [20:05] <sbp> s/would/would they/ [20:05] <sbp> you can say that some access mechanism works or it doesn't work [20:06] <sbp> and you could follow that, I suppose [20:06] <nick_kew> Accept: text/html;graphics=no,sound=yes ? [20:06] <sbp> :-) [20:06] <nick_kew> (e.g. for mobile device) [20:06] <sbp> opinions about HTTP extensions tend to Vary [20:07] <sbp> wow, that was awful. Sorry [20:07] <nick_kew> aaargh! [20:07] * JibberJim quite liked it... I obviously need to get out more. [20:07] <sbp> heh, heh <wendy> it seems to me that if a client has some cc/pp and approaches a server for a match, [20:08] <nick_kew> Accept: spb;q=0.01 <wendy> that the server could check out its earl and say "here's something that matches exactly" [20:08] <nick_kew> Wendy, http content negotiation does exactly that! <wendy> we don't get any benefits? any logic? <wendy> content negotiation has this level of detail? [20:09] <nick_kew> Convince me we do get any benefits [20:09] <sbp> benefit: all your GETs get slowed down a lot, so the bandwith use of the Internet goes down [20:09] <JibberJim> I don't think there's much point of this, at least 2 of us see nothing but negative in CC/PP and I think it's probably 3. <wendy> agenda? [20:10] * Zakim sees 2 items on the agenda: [20:10] * Zakim 1. comments from EARL tech review [20:10] * Zakim 2. tools review from ATRC/Chris Ridpath <wendy> agenda- 1 [20:10] * Zakim notes agendum 1, comments from EARL tech review, dropped <wendy> any comments on chris' tools review? [20:10] <nick_kew> Content negotiation is based on several headers. If not enough detail, it could be revised [20:10] <sbp> when's it going to be finished? <wendy> dunno. [20:10] <JibberJim> Refresh my memory? <wendy> http://lists.w3.org/Archives/Public/w3c-wai-er-ig/2002Jan/0010.html [20:11] <nick_kew> I've asked them to include Page Valet, but it sounds like they won't find the time <wendy> that's gonna be a lot of earl to play with. <wendy> to be presented in march 2002. <wendy> probably not much to say. [20:12] <sbp> the ATR output is a little borked [20:12] * JibberJim Is still waiting for the page to appear... <wendy> in what way? [20:12] <sbp> which is odd, because I'm sure that Chris fixed the problems etc. <wendy> hmm. maybe you could follow up w/him? [20:12] <sbp> in what way: tools have odd IDs, it uses earl:Pass instead of earl:passes... [20:13] <sbp> yep, I should do. Give me an ACTION item [20:13] <nick_kew> Can Chris Ridpath be subpoenad to appear here sometime? <wendy> action sbp: help get ATR output unborked [20:13] <sbp> I think the correct term is "deborked" :-) <wendy> :) <wendy> deborked it is. <wendy> i could try to get him to a meeting next week. [20:14] <sbp> how about: "Help Chris debork the ATR output"? [20:14] <nick_kew> .google ATR <wendy> sbp - whatever will help you. [20:14] <xena> ATR: http://www.atr.co.jp [20:15] <sbp> .google ATR Chris Accessibility [20:15] <xena> ATR Chris Accessibility: http://www.w3.org/WAI/ER [20:15] <sbp> argh <wendy> ATR http://www.aprompt.ca/ATR/ATR.html <wendy> plus http://lists.w3.org/Archives/Public/w3c-wai-er-ig/2001Sep/0012.html <wendy> ok. so i'll try to get chris here next week. <wendy> i think that's it for today unless there other things y'all want to discuss. <wendy> i'll send this log to the list for minutes. <wendy> ribbit [20:16] <sbp> :-) [20:16] <nick_kew> What does ATR run on? [20:16] <JibberJim> I'm going to be building a new Repair Tool in the next couple of weeks, would you rather I did that or looked at EARL and annotations? <wendy> windoze [20:16] <sbp> Methanol <wendy> lol [20:17] <nick_kew> sbp, I prefer ethanol <wendy> jim - what's the repair tool gonna do? [20:17] <sbp> Nitromethane is good stuff, too [20:17] <nick_kew> wendy, I prefer anything but <wendy> personally - i'm interested in EARL and annotations... [20:17] <sbp> for fast-cars, that is... I wouldn't advise drinking it [20:17] * sbp too [20:17] <JibberJim> Evaluation Tool really. A development on Snork that can be properly deployed. <wendy> agenga -1 [20:18] <JibberJim> A tool that can cope with and report on accessibility/ cross browser nature of scripts. <wendy> agenda- 1 [20:18] * Zakim notes agendum 1, tools review from ATRC/Chris Ridpath, dropped <wendy> agenda? [20:18] * Zakim sees nothing on the agenda [20:18] <sbp> Jim: write-up? <wendy> cool. then we're done. [20:18] <nick_kew> Jim, that's breaking new ground. Go for it! [20:18] <JibberJim> Ah, but Wendy wants EARL and annotea... <wendy> i'll let you all fight it out. [20:19] <nick_kew> sbp+jim: A UK accessible web campaign? [20:20] <sbp> Nick: sounds interesting. I noticed something about that on #validator, but lost my logs <wendy> see you all next week. jim - whatever you decide would be good - as long as it does something w/earl. :) [20:20] <nick_kew> Ah, I buttonholed Jim; you weren't talking to me <wendy> all tools should do something w/earl. :) [20:21] <sbp> pardon? I am not someone to whom "button hole" is a verb [20:21] <sbp> Wendy: heh [20:22] <nick_kew> As I said, I wrote to the Beeb's "in touch". [20:22] <sbp> yep, I recall [20:23] <nick_kew> But then in my usual brainfart mode, I thought why stop there? How about a more extensive campaign, a .org dedicated to it, etc [20:24] <JJ> Any thoughts on Annotea or Snork (I need a proper name...) [20:24] <nick_kew> ribbit [20:25] <nick_kew> OK, if it still seems a good idea in the morning, I'll post something more coherent [20:26] <JibberJim> Is the only interface to EARL Seans new API? [20:26] <JibberJim> Nothing that outputs it in a friendly manner? [20:27] <nick_kew> Didn't Jim write an Earl client? [20:27] <sbp> yeah. That was a great demo [20:27] <JibberJim> Yeah but it didn't understand real Earl... [20:28] <nick_kew> (nick wishes he had the time to start doing interesting things with Konq like Jim does with IE) [20:28] <JibberJim> I was thinking more generic than web, I mean if ATR gives us some EARL on all these tools, you'll need something to look at them. [20:28] <JibberJim> Snufkin even has some real users! [20:28] <JibberJim> (other than me I mean.) [20:29] <nick_kew> Did I ever tell you about the first s/w I released free under the Webthing name? [20:29] <JibberJim> and the BNP? [20:29] <nick_kew> evidently yes :-( [20:30] <sbp> EARL => HTML: all I can think of is XSLT, or API [20:31] <JibberJim> Isn't RDF-XML quite annoying to parse though with XSLT. [20:32] <nick_kew> Why's that? (not tried it) [20:33] <sbp> yeah, it is quite annoying [20:33] <JibberJim> I'm not sure my xslt is pretty ropey (I prefer DOM.) I just heard that. [20:33] <sbp> but you can get by, because many people will output some fairly regular output [20:34] <nick_kew> I can see lots of namespaced elements might be irritating [20:34] <sbp> well, XSLT is TC, so anything should be possible. But XSLT is so annoying to debug [20:34] <JibberJim> We never discussed if we use earl: or reuse others... [20:34] <nick_kew> I thought we did both? [20:35] <JibberJim> Maybe I missed it, what was the decision? [20:35] <JibberJim> was it earl:email or foaf? [20:35] <nick_kew> pass. sbp? [20:36] <sbp> you're right: we didn't discuss it. I should have asked to add it to the agenda :-) [20:36] <sbp> I just wondered if we should bung lots of terms into the EARL namespace, or use ones already defined [20:36] <JibberJim> I think already defined so long as they are well defined. [20:36] <sbp> it's a bit of a thin line, because for each, there are pros and cons. It's not a particularly interesting debate, though [20:37] <sbp> yeah, that's my general opinion, too [20:37] <JibberJim> Although it makes hack type parsers more complicated. [20:38] <nick_kew> yeah. Perhaps a limited set of namespaces should be included in earl spec? -- wendy a chisholm world wide web consortium web accessibility initiative seattle, wa usa /--
Received on Friday, 1 February 2002 13:25:56 UTC