Semantic Web Interest Group IRC Chat Logs for 2004-01-15

01:22:58 <tav> tav is now known as tav|sleep

02:02:55 <D[a]vey> D[a]vey is now known as Davey

02:23:47 <darobin_> darobin_ is now known as darobin

04:02:28 <D[a]vey> D[a]vey is now known as Davey

04:42:36 <tav|sleep> tav|sleep is now known as tav|offline

04:42:45 <tav|offline> tav|offline is now known as tav|sleep

04:46:13 <nick_kew> nick_kew is now known as niq

06:05:49 <Davey> Davey is now known as D[a]vey

08:22:34 <D[a]vey> D[a]vey is now known as Davey

09:10:47 <mdupont> hi

09:29:59 <CaptSolo> hi all

09:30:59 <mdupont> hey CaptSolo!

09:32:28 <CaptSolo> what's new mdupont?

09:32:47 * CaptSolo read through the SW tutorial in N3

09:33:01 <CaptSolo> good stuff, very easy to understand

09:33:10 <CaptSolo> and actually things look simpler in N3

09:33:27 <CaptSolo> (i was using RDF/XML for the most part)

09:36:01 <deelan> n3 rocks

09:36:40 <deelan> also check turtle: http://www.ilrt.bris.ac.uk/discovery/2003/11/ntriplesplus/

09:36:59 <mdupont> n3 is great

10:09:54 <swh_home> swh_home is now known as swh

10:12:27 <mdupont> i use cwm, euler and redland for procesing

10:16:17 * libby waves, looking for an image meeting...

10:18:51 * GregElin waves "over here"

10:19:53 <libby> heya gregelin

10:20:03 <GregElin> hi

10:21:33 * GregElin looks for a few links

10:24:34 <GregElin> Libby, I was looking at your rdf tool yesterday...the one with the blue background. But I thought I had seen another, too.

10:24:53 <libby> I had the same one without a blue background :)

10:25:14 <libby> I was applying my feeble design abilities

10:25:56 <GregElin> Was that the only one?

10:26:04 <libby> yeah I think so

10:26:07 <GregElin> (and I can I get URL one more time?)

10:26:36 <libby> yep http://swordfish.rdfweb.org/discovery/2003/06/codjsform/shell.html#

10:26:37 <GregElin> I'm also looking at http://kanzaki.com/docs/sw/img-annotator.html

10:26:56 <libby> yeah that's very cute

10:27:14 <GregElin> What do you think he means by the two fields, Area description and Description that depicted?

10:27:30 <libby> morten sent me this: http://www.wasab.dk/morten/2003/10/photos/england/3/edit.html

10:28:45 <libby> there's a generated bit of rdf somewhere - that might explain a bit

10:29:53 <libby> hm the 'area description' is a dc description of the area - i.e. atextual description

10:30:16 * GregElin added links to wiki.

10:30:22 <libby> the 'description that depicted' is another dc:description

10:30:28 <libby> not sure about that

10:30:34 <libby> it's a nice interface though

10:31:14 * GregElin cogitates...trying to follow a small thread that's been growing in his head...

10:32:51 <GregElin> Do you have a link to morten's interface taht created all the data? I like the effort at parsing the rdf and mixing icons to present info

10:33:50 <libby> I'm not sure how he creaed it initially. some is direct from the exif data in the image and from a gps. in that link above you can edit several images at once

10:35:53 <GregElin> I see: http://www.wasab.dk/morten/2003/11/annotate/

10:36:55 <libby> yep. on the previous page, you can tick boxes and annotate multiple pictures at once

10:37:03 * GregElin continues to cogitate...but can handle hearing your thoughts.

10:38:33 <libby> to a certain extent the exact form of the UI will depend on the things we want to talk about, which we haven't decided yet...I'd suggest running through a list of them briefly, maybe add to it later

10:39:31 <libby> so imo, we need:

10:40:41 <libby> - for the image itself: dc descripton, creator/maker, date, license

10:41:41 <GregElin> dc=dublin core?

10:41:46 <libby> - areas of the picture: depicts people, maybe objects, also dc:description perhaps

10:41:47 <libby> yep

10:42:11 * deelan likes links posted so far

10:42:52 <libby> - the whole image: the event depicted, perhaps the location (or this could go in the event)

10:43:33 <GregElin> another link: http://www.fotowiki.net/w3photodev/fotonotes.php?p=photos%2Fjackelinfriends.jpg

10:43:35 <libby> - the metadata itself: creator/maker, license, date annotated

10:43:51 <libby> I'm sure that's not complete but might server as a start

10:43:54 <libby> serve

10:44:51 <GregElin> better url: http://www.fotowiki.net/w3photodev/jackelinfriends.ftw

10:44:58 <libby> maybe we shoudl chump all these links?

10:46:00 <GregElin> I think tis good to separate location from event. Both location and event can be respectively recurssed (e.g., New York > New York Sheridan > Lobby of New York Sheridan)

10:46:21 <GregElin> Location v. geolocation v. 'place' is a discussion on its own!

10:46:42 <GregElin> sure

10:47:44 <GregElin> Issues we need to figure out:

10:48:23 <GregElin> "layering" e.g., representing multiple annotations of the same area (if 20 people annotate the same thing;

10:48:34 <GregElin> Can you comment on an annotation?

10:48:52 <libby> I guess you could

10:49:09 <GregElin> "annotator" do we track who annotes? Is a log in required?

10:49:11 <libby> I'd suggest leaving till later though, might be fiddly

10:49:17 <GregElin> (or desired?)

10:49:29 <libby> I would suggest we track somehow. you sugggested a bloggy system

10:49:40 <libby> I think layering is essential

10:49:45 <GregElin> "editing" interface for editing existing annotation

10:50:15 <GregElin> "storage method" indicating/selecting if annotation is stored in jpeg or somewhere else

10:50:57 <GregElin> Browsing/Searching, default navigation interface for finding your way around the collection of images.

10:51:23 <GregElin> Does each annotation have its OWN URL?

10:51:35 <libby> might be nice to be able to add a dc:description to the image even if you don;t say what is depicted

10:51:50 <libby> <GregElin>Does each annotation have its OWN URL?

10:51:58 <GregElin> support rectangles only or polygons?

10:52:18 <GregElin> what do you mean, Libby? Example?

10:52:20 <libby> that's how I';ve implemented it. then you can add information abotu the annotator at the document level, and separate out different layers that way

10:53:11 <libby> that also makes it easy to remove an annotation if it is objected to (which is pretty important I think)

10:53:32 <libby> very important :)

10:54:45 <libby> in terms of a browse interface, I have : http://swordfish.rdfweb.org/discovery/2001/08/codepict/

10:55:27 <libby> the idea being that for any omage, you display any metadata that you can find and link to similar images, by person, by wordnet thing, by location, by event etc

10:55:37 <GregElin> I see that first screen as a "search" more than a browse

10:55:51 <libby> yeah

10:56:04 <GregElin> okay..until I do search that is...

10:56:38 <libby> yeah I meant stufff like: http://swordfish.rdfweb.org/discovery/2001/08/codepict/aboutimage.jsp?uri=http://swordfish.rdfweb.org/photos/2003/05/24/2003-05-24-Images/6.jpg

10:56:42 <GregElin> That make sense, a simple search form at top of page (or side of page) with found images represented.

10:56:50 <libby> though an initial search page by name is very useful I've found

10:57:08 <libby> (it used to be by email, but by name is much more popular)

10:57:28 <libby> it's all fairly straightforward browsing-wise I think

10:57:47 <GregElin> I like the feel of clicking an name and seem more photos of that person...

10:58:31 <libby> actually that image has been annotated twice, so there's an example of how that might work

10:58:41 <GregElin> Oh yes...initial search page good, but needs to either be single field (like google) or have some default images shown (e.g., recent) shown too, else it is intimidating.

10:58:43 <libby> (badly in this case - duplicate names)

11:00:56 <libby> re <GregElin>I think tis good to separate location from event. Both location and event can be respectively recurssed (e.g., New York > New York Sheridan > Lobby of New York Sheridan)

11:01:15 <libby> that's true, but the relationship between an event and a location isn;t very clear

11:01:47 <libby> i.e. locations might not be depicted by an image, they are where the image was taken

11:02:07 <GregElin> agreed

11:02:19 <libby> moretn does this:

11:02:21 <libby> <dc:spatial rdf:parseType="Resource">

11:02:21 <libby> <cyc:near rdf:parseType="Resource">

11:02:21 <libby> <geo:lat>40.442673</geo:lat>

11:02:21 <libby> <geo:long>-79.945815</geo:long>

11:02:21 <libby> </cyc:near>

11:02:22 <libby> </dc:spatial>

11:02:25 <libby> which I quite like

11:02:40 <GregElin> problem with event...is it "WWW2001 Conference" or is it "Key Note Speech"?

11:02:49 <libby> anyway, that's another discussion maybe. the key thing is that it's at the image-level not at the part of image level

11:02:58 <libby> it could be both

11:03:02 <Davey> Hmm, is there any easy way to tell if a URI is a node id or not?

11:03:18 <GregElin> that is a geolocation in my book...but remember all 10 floors of a hotel would have that same geo!

11:03:27 <GregElin> node id?

11:03:27 <libby> yeah, true

11:03:34 <libby> sorry davey, I don;t know

11:03:39 <libby> it's ok :)

11:04:18 <GregElin> Actually, Davey...it could be a relevant question here, too.

11:04:18 <libby> I think greg that we might have to settle for rather non-specific location information, but leave it open to more detailed geo location afterwards

11:04:27 <GregElin> what is a node id?

11:04:35 <libby> i.e. make sure it's extensible in a more specific way later

11:04:37 <Davey> GregElin: its on-topic but not to do with the image stuff

11:05:11 <libby> it's an rdf syntax contruct greg, for identifying a thing in a documet to refer to it repeatedlky in teh same documnent

11:06:07 <Davey> GregElin: well, libby has <foaf:currentProject><dc:title>foo</dc:title></foaf:currentProject> which results in a "foaf:currentProject http://xmlns.com/foaf/0.1/currentProject bNode5" (for example) triple. How am I supposed to know if the "URI" object is the subject for another triple, or (as I expected) a URI to a page about a project

11:06:07 <libby> the thing is greg, if we keep it simple like morten's example, we could just do a drop-down list of the 13 conferences, and add geo info automatically

11:06:39 <libby> well a bnode is a bit different fro a nodeid

11:06:53 <libby> a bnode may be identified by a nodeid or it might not

11:06:57 <Davey> libby: right, a bnode seems to work like an automatically generated ID, from what I'm seeing.

11:07:03 <Davey> yeah, that too :)

11:07:06 <libby> yep

11:07:26 <GregElin> bNode?

11:07:33 <Davey> blank node

11:07:34 <libby> the api shoudl give you that information really davey - should have some way of working it out internally

11:07:51 <dajobe> libby, line 1 of http://swordfish.rdfweb.org/discovery/2001/08/codepict/scutterplan.rdf is not xml ..

11:07:56 <Davey> libby: ok... I can that, I think. But at the time I want, its too late.

11:08:12 <GregElin> blank node in terms of something identified, but nothing there when you go to it?

11:09:22 <libby> no blank node in that it doesn;t have a uri

11:09:44 <libby> so greg, has that progressed things a bit for you?

11:09:48 <GregElin> Libby, if we were using rdf, OWL, or a foaf...and we wanted to note that a photo was from www1999. We could we note that AND note an additional event, like "first night dinner"? Would that just result in two lines in the rdf? What about OWL? Foaf?

11:10:09 <libby> hm

11:10:10 <Davey> libby: but if you specify a nodeid, it can be a blanknode AND have a URI... *boggle*

11:10:18 <GregElin> regarding interface or learning about nodes?

11:10:24 <libby> interface :)

11:10:39 <GregElin> Yes. Think I have an idea.

11:10:53 <libby> you can certainly say it depicts two events in RDF

11:11:41 <libby> what we would like to be able to say is taht this event is in some sense 'part of' another event, so that you could say that it annotated panel A and the software would infer that it also depicted www1999

11:11:58 <libby> the mindswap ontology has that property

11:12:24 <libby> however from an interface point of view I think going from WWW1999 to the subevents would be easier

11:12:34 <libby> *if* we had the information about the panels and talks etc

11:12:36 * Davey concurs

11:12:44 <Davey> if I understand you at least :)

11:12:52 <libby> which might in fact be rather difficuklt to get because it will involve contacting 13 organisations

11:13:16 <Davey> libby: no websites with historical data?

11:13:33 <libby> - that's why I'd suggest just doing structured information about the conference at first, then see if we can be more specific later

11:13:49 <libby> how do you mean davey?

11:13:59 <Davey> well, don't the websites keep the old data around?

11:14:17 <libby> yeah we could do that but we'd have to scrape it

11:14:27 <libby> - all different html etc etc

11:14:39 <Davey> why not just extract it by hand? heh

11:14:51 <libby> danbri's sent an email to the organising committee so maybe we won't have to

11:15:04 <Davey> OK

11:15:27 <libby> often (I think) the more interestign pictiures are not of panels or speeches, but more informal ones. there are exceptions to that of course

11:15:47 <Davey> I tell you something, pawscon has certainly generated quite a stir.

11:15:56 <libby> that's cool :)

11:16:07 * Davey generates new webalizer stats

11:16:23 <libby> gregelin, have you given any thought as to storage of this rdf data for searching etc?

11:16:35 <Davey> http://www.pawscon.com/~stats/ for anyone interested :)

11:16:35 <dc_rdfig> A: http://www.pawscon.com/~stats/ from Davey

11:16:54 <Davey> A:|Webalizer stats for PaWS Conference Website

11:16:55 <dc_rdfig> Titled item A.

11:17:13 <GregElin> davey, how many people are you expecting, btw?

11:17:20 <Davey> GregElin: 300-500

11:17:37 <Davey> libby: is Bristol Uni interested in presenting anything btw?

11:18:13 <GregElin> Libby, I've made a note that we have to figure that out...

11:18:31 <libby> heh, I don;t know davey

11:18:37 <GregElin> Fotonotes stores stuff in jpeg. But I think we want to also store RDF outside of jpeg.

11:18:58 <libby> we would have to if you want to search right?

11:19:10 <Davey> grabbing the data from JPEG every time would be quite costly...

11:19:37 <libby> one option might be to 'harvest' the data form the jpegs, put it in a database and search like that

11:19:50 <libby> using jena, or something else...

11:20:24 <GregElin> yep. I like that

11:20:39 <GregElin> btw...do you know if OWL and foaf also support multiple locations?

11:20:52 <libby> it decouples the annotation and searching interfaces

11:20:59 <GregElin> I think Event (conference year) and subevent (whatever) makes the simplest form

11:21:12 <GregElin> yes

11:21:40 <libby> we've been looking at locations in foaf only really indirectly

11:22:01 <libby> mortenf uses cyc:near which is quite nice

11:22:02 <GregElin> OWL?

11:22:34 <GregElin> what is "cyc" an abbreviation of?

11:22:50 <libby> owl is a way of describing vocabularies. what we need is a particular vocab that relates things to locations

11:23:02 <libby> so foaf has 'basednear' for people and places

11:23:04 <tav|sleep> tav|sleep is now known as tav

11:23:21 <libby> cyc isn;t an abreviantion, it's an upper ontology

11:23:42 <Davey> actually, its a namespace qualifier, it doesn't have to mean anything </pedantic>

11:23:43 <libby> - http://www.opencyc.org/

11:24:21 <GregElin> back to nodeid for a moment...can't an idenified area of a photo be thought of as a node of the photo?

11:24:55 <libby> I think this needs more discussion

11:24:56 <GregElin> Thanks of that info.

11:25:01 <libby> don;t get sidetracked by nodeid

11:25:11 <libby> - it's a syntactic thing, can fix later

11:25:47 <libby> I was going to suggest a discussion about the rdf/owl relationship between an image and part of an image

11:25:57 <GregElin> Not distracted by it...something I've been thinking about for a long while. Photos, in a sense, are a flattening in time of an N-dimensional space.

11:25:59 <libby> because there's so much interest in this area

11:26:27 <GregElin> that would be a good discussion...maybe after we get an interface up?

11:26:58 <libby> we might need something minimal to link an image with it's parts

11:27:36 <libby> you need to be able to get from an image to its parts and back again

11:27:45 <libby> almost certainly one of the proposed vocabs does this

11:28:07 <Davey> I think that putting the RDF in the image itself is a good idea in that, its a self contained entity, you can copy it to another server and keep the RDF with it. Harvesting it though,would be necessary for a speedy/searchable system.

11:28:52 <GregElin> I think you described it perfectly, Davey.

11:29:06 <Davey> yay me!

11:29:13 <Davey> and to think, I don't even *know* RDF :D

11:29:35 <libby> so in jim ley's tools (and I think mindswap's too) you can click on a part of an image and get more pics of the person outlined

11:29:47 <GregElin> Right libby...so back to Davey's original question...is there something in a URL that tells you if you are looking at whole photo...or looking at a part of a photo (e.g. a 'node).

11:30:11 <libby> hm, I don;t think that's was davey's question !

11:30:12 <Davey> I had a question that said that? I only asked about nodes cause of my FOAF parser :)

11:30:17 <GregElin> And in Fotonotes you can address and get just the part of the picture, rendered on the fly.

11:30:22 <libby> we shouldn;t be looking in a url for that information

11:30:38 * libby looks for a good example

11:30:56 <GregElin> Davey: Hmm, is there any easy way to tell if a URI is a node id or not?

11:31:18 <Davey> that was my question :)

11:31:36 <libby> it's not reqally relevant here

11:31:44 <Davey> the only way I can see is to check that no subject, predicate or object has it, and its not a fully valid URI at that.

11:31:59 <GregElin> Which I construed (perhaps out of being naive) to be "Should we be able to tell if a URL is for a part of something?"

11:32:13 <libby> ok, gregelin, davey, look at 2.2 Describing part of an image and what is depicted by it in http://esw.w3.org/topic/ImageDescriptionRdfExamples

11:32:25 <libby> that's not the same question greg

11:32:37 <libby> it's confusing. davey was talking about a syntax question

11:33:07 <libby> we need to be able to see what's part of an image, but that's not something we shoudl get from the url

11:33:22 * GregElin smiles cryptically...

11:33:37 <libby> in jim's example, theres a url (or some id) for the image and another one for the part of the image idetified by the points

11:33:51 * Davey *knew* he shoulda asked in #foaf. Hehe

11:34:06 <libby> and they are connected by 'haspart'

11:34:23 <libby> I'm not suggesting we necesaarily use that vocab, but those are the sorts of things we need

11:34:27 <GregElin> I'm interesting in being able to address/point-to/work with just a part of an image independently of the entire image.

11:34:53 <GregElin> That's true. The image *does* have part which depicts Libby.

11:35:22 <GregElin> And a visual instance of "Libby" is also included within the image.

11:35:31 <libby> right. jim ley and danbri did stuff like that with svg. Jim's vocab mapos to svg which is a nioce way of doing that

11:35:50 <GregElin> a harddrive is part of a computer...

11:36:24 <GregElin> but the harddrive can also be referenced indepdently of the computer...and can be placed in a different computer...or in a firewire box.

11:36:31 <libby> e.g. http://jibbering.com/rdf/co-search.1?person=mailto%3Adanbri%40w3.org :)

11:36:53 <libby> (SVG)

11:37:21 <GregElin> Now you tell me! That just killed my browser.

11:37:24 <Davey> Davey is now known as D[a]vey

11:37:28 <libby> oops! sorry

11:37:39 <GregElin> what would I have seen?

11:37:51 <libby> a bunch of outllines of dan

11:38:10 <libby> i.e. just the parts of pictures with dan in

11:38:15 <GregElin> Outlined within an image?

11:38:26 <GregElin> or as their own thumbs?

11:38:42 <libby> initially outlined in an image and then clipped and redisplayed

11:38:44 <GregElin> (or jpegs?) And

11:39:04 <GregElin> And do those redisplayed thumbs each have a URL?

11:40:17 <libby> can you see tiffs?

11:40:32 <GregElin> My point is this: a blog has posts and a post is directly addressable.

11:40:40 <libby> right

11:41:03 <libby> in this case, the images themselves don;t necessarily have urls right?

11:41:04 <GregElin> Tiffs? Sure. Especially if emailed. Not sure browser will...

11:41:26 <GregElin> Doesn't both a blog and a blog post have URLs?

11:42:16 <libby> http://swordfish.rdfweb.org/~libby/danbri.tiff

11:42:22 <libby> yep

11:42:39 <libby> sometimes the urls will be fragments of a larger page

11:43:01 <GregElin> Isn't it interesting to say that libby@www1999photoDs10.jpg or greg@www1999photoDs10.jpg? Doesn't that unleash the same power that libby@w3c.org does?

11:43:22 <libby> greg, I don;t think it matteers if the fragments are urls or uris; a unique id though is important, yep

11:43:41 <libby> and I think both vocabs do that

11:43:59 <libby> perhaps yep

11:45:08 <libby> so if you're planning on embedding information in the jpeg, that assumes people will upload the photos

11:45:21 <libby> heya bryce

11:45:34 <GregElin> Which brought me back to davey's question...does it make a difference if the URI of a node can be represented in a URL. (^_^)

11:45:56 <libby> 'as a url'?

11:45:59 <D[a]vey> that *still* wasn't my question, stop bringing me into this :P

11:46:13 <libby> 'as a url' rather than 'in a url'?

11:46:35 <libby> or 'by a url' rather than 'in a url'?

11:47:01 <libby> hm, the answer is, I don;t know

11:47:05 <GregElin> i know I read the question out of context...but I think the specific foaf version of a more abstract issue --identifying things within some type of nodal space with URLs-- which also matters to photos.

11:47:33 <GregElin> Um, I guess I don't know if I separated a in a url from a url.

11:47:37 <libby> an html fragment identifier can;t refer to part of an image as far as I know - that just does n't make sense

11:47:46 <GregElin> basically, I'm thinkin permalink...

11:48:10 <libby> ok, well we could snip the part of the picture, put it up somewhere, and give it a url yep

11:48:59 <D[a]vey> libby: an HTML fragment identifier COULD refer to a piece of RDF that is about part of an image

11:49:12 <libby> yes but that's not quite the same

11:49:22 <GregElin> if html pages have fragments...why not photos, too? But I can let the point go for now. It is a point of differentiation for Fotonotes, though...the idea of being able to *easily* address and work with fragments.

11:50:48 <libby> there may well be somethingi in svg that allows this, which is why a mapping to svg is nice.

11:50:53 <GregElin> (Sorry and thanks Davey!)

11:51:03 <libby> I don;t know enough about it

11:51:10 <GregElin> Maybe...I'm not sure if I saw it in SVG when I read it.

11:51:20 <libby> I suggest punting this to a discussion of the vocab itself

11:51:23 <GregElin> Anyway...back to the interface. I think we did make good progress.

11:51:45 <libby> do you think you'll have enough time to implement all this greg?

11:52:30 <libby> also - can you set up jena or somesuch to harvest the rdf data once we have it, or maybe you need help with that?

11:52:32 <GregElin> I'm imagining a simple interface...with minimum fields named specifically for our context of W3 conferences...and the ability to expose an advance form that gives more (abstracted) field controls...

11:52:46 <libby> right, sounds good

11:53:08 <GregElin> I have enough time to work on the annotating of the selected area, and annotating a single image overall...

11:53:43 <GregElin> ...but I'd love somebody else to pick handle (a) generally browsing through collection and (b) search & search results. Maybe use your stuff?

11:54:05 <libby> I'm not sure mine is too reliable :(

11:54:15 <libby> but jena can probably do it - a java rdf toolkit

11:54:16 <GregElin> i was also thinking about Gallery as a possible front end for browsing...or maybe something that UMD has done.

11:54:35 <libby> it's just a bunch of queries really

11:54:37 <GregElin> I can probably figure jena out, especially with a little help.

11:54:52 <GregElin> Yes..but there still is designing the form...

11:54:55 <libby> <JibberJim>svg has a fragment id which can point to individual elements or an area of the image.

11:55:17 <GregElin> cool.

11:55:23 <libby> yep, seems like a lot to do

11:55:43 <libby> greg, you mean to browse by date?

11:56:16 <GregElin> I think browse by date would be great.

11:56:23 <libby> yeah makes sense

11:56:40 <GregElin> That's the kind of thing it would be great if someone else helped accomplish.

11:57:18 <libby> well, fairly strightfoward

11:57:25 <GregElin> I don't want to re-create the good work that already exists on browsing and organizing collections ofimages.

11:57:51 <GregElin> I'll get to putting together a mock up of these ideas.

11:58:02 * libby channelling jibberjim:

11:58:03 <libby> <JibberJim>I'd like to see an interface not be the key thing, I'm much more interested in having a server (perhaps annotea, perhaps something simpler) than many tools can post RDF to, and just agree on the syntax of that RDF.

11:58:03 <libby> <JibberJim>Pushing everything on Greg doesn't help, we have 4 or 5 interfaces to generating the data, we should all be able to talk it, interopability is the thing I think we should be demonstrating.

11:58:26 <libby> post is an interesting idea

11:58:35 <libby> agreeing on the RDF ++

11:58:51 <libby> greg, do we have a server we can use for this?

11:58:54 * bryce is cathing up... didn't expect so much convo while he was reading email!

11:59:13 <libby> I can help but I'm only used to linux-like systems, be lost on windows

11:59:32 <GregElin> Where is Jibber?

11:59:35 <D[a]vey> D[a]vey is now known as Davey

11:59:42 * libby could help w jena, maybe annotea or similar...

11:59:48 <GregElin> Jibber...can you see my Aunt Zelda? does she forgive me?

11:59:55 <libby> he can only talk 121 for some reason

12:00:03 <libby> he can see the logs :)

12:00:40 * libby likes jibber's post idea very much, but it means relinquishing some control over what's in the data

12:00:56 <libby> could still delete etc etc

12:01:11 <GregElin> Jibber...I want to make sure we have a simple default interface. Simpler the better. Then, if we work out the storage of photos correctly, all of us can create other interfaces.

12:02:01 <libby> greg, how about I set up an example post server like annotea or similar for people to test with

12:02:30 <libby> fotonotes type thing could still be a 'default' on the site, with links to others

12:02:33 <GregElin> post server...as in put photos there? or as in put annotea annoations there and point to photos on other server?

12:02:41 <libby> the latter

12:02:59 <libby> hm, though that would mean separating out the photos form the metadata

12:03:05 <libby> and linking to photos

12:03:26 <GregElin> Go for it! Ben suggested we get a sampling of photos ASAP, like 100-300 and begin to work with. I can put them up on fotonotes as a place for those images.

12:03:52 <GregElin> As per davey...I think we want metadata in image...and outside.

12:04:00 <libby> <JibberJim>I think storage has to come first, I don't see the value in producing a "semantic server", we know we can do that it.

12:04:01 <libby> <JibberJim>no, you could post photo with embedded metadata

12:04:09 <Davey> inside for image, outside for application.

12:04:35 <GregElin> I can post data inside photo right now. We can modify to stick rdf inside, too.

12:04:37 * libby will chat with jibs, see what he means

12:05:02 * libby better get on w other stuff

12:05:10 <libby> nice chatting greg :)

12:05:18 <GregElin> Does anyone have an existing 'scrape' tool to get photos and data from a remote site? Can we use your stuff for that libby?

12:05:39 <GregElin> nice chatting with you, too, libby.

12:06:00 <libby> I can grab links to photos

12:06:07 <libby> hm

12:06:49 <libby> did you see I got some? not many though

12:07:03 <libby> - http://swordfish.rdfweb.org/discovery/2001/08/codepict/aboutevent.jsp?uri=http://www2003.org/

12:07:28 <mdupont> http://www.jpl.nasa.gov/index.html yeahh!!! spirit is on the move!!

12:07:57 <bryce> greg... i use httrack sometimes to copy an entire website. if we found a repository of images and were lucky enough to find the images in one dir, it would be really easy. i suggest browsing through some cc photo collections.

12:12:26 <bryce> bryce is now known as bryce_afk

12:12:43 <JibberJim> Hi

12:14:46 <JibberJim> Greg, I think the most important things to agree are the vocabulary for the annotations themselves.

12:15:29 <JibberJim> and the way of telling the central resource about metadata, either posting raw RDF, or posting a link to some remote rdf, or posting a link to a jpeg containing RDF.

12:21:34 <libby> what's the name of hawaii's airport?

12:22:19 <dajobe> honolulu presumably

12:22:19 <JibberJim> Actual interface design can then flourish underneath this level, While it will be great if you can create a central site with a good annotator, if we can other tools that can also interoperate, then we'll get a lot more data.

12:22:48 <libby> d'oh! thanks dajobe

12:23:00 <JibberJim> You have whwhwhwwhwh search for it!

12:23:05 * libby thinks great idea jim; shall I tryu and set something up? what would it need?

12:23:20 <libby> you need to know it's name first...or what it's near

12:23:36 <JibberJim> all it would need is an agreement on the interface :-)

12:24:05 <libby> right I meant a post server etc

12:24:07 <dmiles> it's code is HNL

12:24:16 <libby> thanks dmiles, found it :)

12:24:22 <JibberJim> It's exactly what the svg annotator already does (POST's RDF data to the server, which then also posts info about it to an annotea server)

12:24:47 <libby> would that also work with photos?

12:26:52 <JibberJim> you can POST a photo to (or ideally PUT it) the code is just take the posted file and save it.

12:28:46 <libby> right, that's basically what mine does too (it also bungs it in a db)

12:29:12 <libby> want me to host somethign temporary, or do you want to?

12:29:32 <libby> how can we get agreement on the form of the rdf I wonder?

12:30:35 <JibberJim> I vote for http://www.w3.org/TR/rdf-syntax-grammar/ :-)

12:31:01 <JibberJim> I don't want to host something, I've got a knackered hard drive remember!

12:31:14 <libby> sure

12:31:24 * libby doesn;t wan t to host permanently

12:31:38 <libby> maybe I should take it to the list

12:32:30 <JibberJim> - http://jibbering.com/rdfsvg/v.2.1 already does it though.

12:32:43 <libby> cool

12:35:13 <dmiles> say(ty)

12:35:15 <jllykifsh> ty

12:35:37 <dmiles> oh wierd but ok.. i noticed when libby said cool jllykifsh tried to run it

12:36:21 <dmiles> % ?- Warning: [Thread 2] Undefined procedure: cool/0

12:36:22 <dmiles> :)

12:36:25 <libby> heh

12:36:31 * libby finds http://www2002.org/gallery/

12:37:48 <libby> - but I don't know who the peopel are :(

12:37:50 <dmiles> awesome %?- error(permission_error(halt, thread, 2), context(halt/1, 'Only from thread `main\''))

12:39:06 * libby thinks - if we had the exif time data, we might be able to match pics with speakers

12:41:32 <bryce_afk> i agree with you, libby. even if we ask the user to confirm, it would be nice if they are given intelligent choices (perhaps a drop down list with the most likely choices floated to the top). of course, the time may have to be adjusted to deal with timezone. maybe the interface "learns" to deal with that for each camera's exif.

12:42:24 <libby> yeah

12:42:33 <libby> think we maybe won;t have time

12:43:05 <bryce_afk> maybe not. maybe for www2008 ;)

12:43:14 <JibberJim> Someone might have enough time in their interface

12:44:58 <bryce_afk> maybe it would be helpful to aggregate our ideas, even if we don't implement them. someone someday might. sort of a w3photo lazyweb (see lazyweb.org)

12:45:34 <libby> true, yes

12:45:44 <libby> on a wiki page?

12:46:35 <bryce_afk> maybe, or a forum, or a bug tracker. whatever is easiest to administer, i guess.

12:46:45 <libby> wiki :)

12:46:45 <bryce_afk> bryce_afk is now known as bryce

12:47:08 * bryce has been, until now, typing with his MIND.

12:47:58 <JibberJim> I've got some www2002 photos marked up Libby

12:48:11 <libby> cool, jim, where?

12:48:31 <JibberJim> on my website, I don't know...

12:48:51 * bryce adds "add wiki" page (or link to one) to todo list for w3photo website

12:49:59 <JibberJim> linked from http://jibbering.com/rdfweb/seeAlso.1 but which I couldn't say without my box or slowly going through

12:50:38 <libby> cool, ta

12:51:02 * JibberJim should've wrote all this stuff up when I actually did it...

13:08:32 <bryce> bryce is now known as bryce_

13:12:08 <libby> - http://swordfish.rdfweb.org/discovery/2001/08/codepict/aboutevent.jsp?uri=http://www2002.org/ :)

13:12:45 <Davey> Davey is now known as D[a]vey

13:14:42 <libby> hello martin :)

13:14:56 <MJDuerst> hello libby,

13:15:30 <MJDuerst> I'm hanging out here because DanC told me that there might be some people here who understand gnu automake/autoconf.

13:15:39 <MJDuerst> I was (and still am) somewhat surprised.

13:16:31 <libby> I can't help, sorry. dajobe maybe?

13:20:51 <dajobe> I do a lot of automake, conf etc.

13:20:58 <dajobe> but I'm jsut off to lunch, back in a bit

13:24:56 <bengee> is it already clear which vocab(s) should be used for a w3photo annotator? or will there be a new/integrated one?

13:26:46 <libby> no, we haven't decided yet bengee

13:27:26 <libby> I was hoping to reuse some of the stuff on http://esw.w3.org/topic/ImageDescriptionRdfExamples

13:28:21 <libby> this: http://www.mindswap.org/~glapizco/technical.owl was also suggested (though it focuses on images and multimedia not on geo, event, etc stuff

13:28:44 <libby> I think we will try and combine some vocabs anyway

13:31:19 <bengee> ah, I see. esw page is great. thanks

13:33:21 <GregElin> Hey...I just got caught up ont he last bit of conversations.

13:33:22 <libby> it's not be any means definitive. some of the contructs are used, some are just suggestions/proposals

13:34:42 <libby> I'm gonna suggest a meeting to decide on a vocab gregelin - sound ok?

13:34:58 <libby> I think getting a bit urgent

13:37:52 <libby> gregelin, thanks for the email. I like the way it looks. does the title/desc refer to the part of the image or the whole? also, is there a way of talking specifically about the people in the pic, i.e. a place for identifying the person by (hashed) email or homepage, weblog

13:37:53 <libby> ?

13:37:56 * libby will send mail

13:44:56 <GregElin> Libby, also the page...http://esw.w3.org/topic/ImageDescriptionUserInterface (replaces ImageDescriptionUI) for notes

13:45:08 <libby> ah cool, yep

13:45:24 * bryce_ notes the link...

13:45:51 <libby> useful page greg

13:45:57 <GregElin> In the segment annotation, I though title might be a title of the segment.

13:46:06 <libby> right

13:46:12 <libby> think it needs one?

13:46:14 <libby> maybe

13:46:20 <GregElin> Wasn't sure if we should just have name of person (thing) or separate title and who depicted.

13:46:41 <GregElin> Undecided, Libby. Just tried it to look at i.

13:46:43 <GregElin> it.

13:46:48 <libby> imo, a description, a name and some id for the person would be ideal

13:47:30 <GregElin> a name of segment? id of person, and description/story? Just three 'fields'

13:47:36 <GregElin> And annotatedby, right?

13:47:46 <libby> it's true that with the sube of all the people in the world who went to the conferences might be unique name, but for interop would be best to try and identify them with something 'more unique'

13:47:59 <libby> I meant name of person

13:48:01 <GregElin> I think there is the User Interface...and JibberJim's point about programming/data interfaces.

13:48:22 <libby> yep

13:48:28 <GregElin> Ah name of person, id of person (optional), description/story, and annotatedby.

13:48:40 <libby> yep

13:48:42 <GregElin> Wirks for me. Listen. I have to run and catch a bus.

13:49:09 <libby> except I'd treat an annotation event as a per-image thing, rather than part of image

13:49:16 <GregElin> I'll make changes. I'm all for you all figuring out the vocab stuff ...along with the ontology stuff that is happening.

13:49:28 <GregElin> Oh. One quick thought...not in UI mock up yet.

13:49:42 <GregElin> I agree regarding event.

13:49:57 <libby> - you could still get back to the annotator from the part of the picture fine

13:50:57 <GregElin> I was thinking we could intertwingle a simple listing of who is in a photo (via a field for the overall image that lists person 1 per line) and actual annotations. In other words...the list of depicted parties is either typed in OR created by parsing the selected areas.

13:51:06 <GregElin> They each support the other.

13:51:24 <GregElin> Didn't say that clearly...but I think we can create an interface that supports both.

13:51:30 <GregElin> Later all!

13:51:35 <libby> not sure what you mean...

13:51:37 <libby> :)

13:56:13 <eaon> eaon is now known as eaon|away

14:20:45 <bengee> heya andreas. still at isi.edu?

14:21:07 <aharth> yep. 6 weeks to go

14:21:28 <aharth> didn't know you're on #rdfig, good to see you :)

14:22:13 <bengee> 2nd day. took me some time to find an irc tool that worked with my proxy

14:22:29 <aharth> how's owlchestra coming along?

14:22:38 <bengee> and took even more time to finish my diploma thesis...

14:22:58 <aharth> do you have your degree now?

14:27:25 <dajobe> MJDuerst: I'm back, what's the problem?

14:28:57 <MJDuerst> hello dajobe,

14:29:37 <MJDuerst> My problem is that I wanted to find a good intro/overview to automake/autoconf/...

14:30:25 <MJDuerst> all the docu I have found until now seems to be pretty accidental.

14:30:34 <MJDuerst> but that may be the problem of the tools, rather than the docu

14:31:21 <dajobe> There's a book somewhere, and I think it has an online version

14:31:34 <dajobe>http://sources.redhat.com/autobook/

14:31:34 <dc_rdfig> B: http://sources.redhat.com/autobook/ from dajobe

14:31:47 <dajobe> B:|GNU Autoconf, AUtomake and Libtool

14:31:47 <dc_rdfig> Titled item B.

14:40:10 <swh> auto* are the bane of my life, they dont like me and I dont like them, but somehoe we have to get along

14:44:18 <dajobe> I seem to be doing better now, automake 1.4-1.8 work for me, autoconf 2.1x -> 2.59

14:58:29 <arnarl> arnarl is now known as arnarl|away

15:00:30 <MJDuerst> the impression I got with auto* is that it's a bit like MSWin: you can figure out how to get something done

15:00:42 <MJDuerst> (if you try hard enough or know the right people),

15:01:04 <MJDuerst> but don't hope for ever understanding the big picture

15:01:09 <MJDuerst> (there might be none)

15:02:05 <dajobe> it's a set of very specialised tools for developers

15:02:19 <dajobe> so it is not like a general operating system for millions of people

15:02:57 <dajobe> there is extensive 'info' documentation plus lots of examples in source code everywhere, if you prefer reading wha tpeople do

15:03:58 <dajobe> of course it does have flaws, but it's way better than most alternatives

15:17:07 <dajobe> yay, redland builds out of the box on OSX 10.3 (without fink). However no mysql and no berkeley DB means no persistent store.

15:29:11 <cmjg> autoconf, automake, etc are easier than the documentation makes them appear.

15:32:07 <sky> cmjg: I strongly disagree

15:38:14 <cmjg> that's your perogative. I'm not claiming they're particularly good

15:38:22 <dajobe>http://www.w3.org/TR/2004/REC-CCPP-struct-vocab-20040115/

15:38:23 <dc_rdfig> C: http://www.w3.org/TR/2004/REC-CCPP-struct-vocab-20040115/ from dajobe

15:38:45 <dajobe> C:|Composite Capability/Preference Profiles (CC/PP): Structure and Vocabularies 1.0 - W3C Recommendation 15 January 2004

15:38:45 <dc_rdfig> Titled item C.

15:47:41 <dngor> are easier than is a type of verb

15:47:46 <dngor> is easier than is a type of verb

16:54:12 <libby> BLURB:sources of WWW conference photos

16:54:12 <dc_rdfig> D: sources of WWW conference photos from libby

16:54:32 <libby> D:[WWW10 website|http://www10.org/photo/w10-gallery.html] - 2001

16:54:33 <dc_rdfig> Added comment D1.

16:54:42 <libby> .google WWW2003 photos

16:54:43 <datum> WWW2003 photos: http://rdfweb.org/topic/BudapestWebConfPics

16:55:17 <libby> D:[photos gathered by danbri|http://rdfweb.org/topic/BudapestWebConfPics] - 2003

16:55:17 <dc_rdfig> Added comment D2.

16:58:26 <libby> D:[dajobe's photos|http://photos.dajobe.org/gallery/www2002-hawaii], [semantic grid photos|http://www.semanticgrid.org/photos.html], [http://www2002.org/gallery/|gallery on the WWW2002 website]

16:58:26 <dc_rdfig> Added comment D3.

16:58:45 <libby> D3:[dajobe's photos|http://photos.dajobe.org/gallery/www2002-hawaii], [semantic grid photos|http://www.semanticgrid.org/photos.html], [http://www2002.org/gallery/|gallery on the WWW2002 website] - 2002

16:58:46 <dc_rdfig> Replaced comment D3.

17:01:49 <libby> D:[libby's from 2002 (bad quality)|http://swordfish.rdfweb.org/photos/2002/05/08/index.html], [etc|http://swordfish.rdfweb.org/photos/2002/05/09/index.html], [etc|http://swordfish.rdfweb.org/photos/2002/05/10/index.html], [etc|http://swordfish.rdfweb.org/photos/2002/05/12/index.html]

17:01:50 <dc_rdfig> Added comment D4.

17:13:30 <D[a]vey> D[a]vey is now known as Davey

17:19:38 <iwaim> iwaim is now known as iwaiAway

17:28:55 <libby> D:[Amy van der Hiel's from 2002|http://people.w3.org/amy/200205_/200205_.html]

17:28:55 <dc_rdfig> Added comment D5.

17:34:32 <bryce> bryce is now known as bryce_mamhlad

18:10:34 <libby>http://esw.w3.org/topic/ImageDescriptionUserInterface

18:10:34 <dc_rdfig> E: http://esw.w3.org/topic/ImageDescriptionUserInterface from libby

18:10:53 <libby> E:|some ideas for Image annotation user interface

18:10:53 <dc_rdfig> Titled item E.

18:12:04 <libby> E:see also [http://esw.w3.org/topic/ImageDescription|ImageDescription], [http://esw.w3.org/topic/WWW2004|WWW2004], [http://esw.w3.org/topic/ImageDescriptionList|Image description list]

18:12:05 <dc_rdfig> Added comment E1.

18:52:32 <bryce_mamhlad> bryce_mamhlad is now known as bryce_

19:28:12 <Davey> Davey is now known as D[a]vey

19:39:43 <deelan> oh my god, sbp!

19:48:55 <nym> hey guys long time

19:49:16 <nym> anyone here use sesame/redland?

21:17:43 <bryce_> cheers to Libby for http://esw.w3.org/topic/ImageDescriptionUserInterface

21:18:31 <libby> greg elin did it all I think

21:19:01 <bryce_> ok, scratch that. cheers to greg!

21:19:05 <libby> :)

21:23:08 <libby> E:by Greg Elin - see also [http://www.ilrt.bris.ac.uk/discovery/chatlogs/rdfig/2004-01-15.html#T10-16-17|today's discussion]

21:23:08 <dc_rdfig> Added comment E2.

21:51:04 <eaon|away> eaon|away is now known as eaon

22:32:16 * DanCon wanders by, noodling on rdf/xhtml, TagIssue:siteData-36, vaguely curious about ImageDescription ScheduledTopicChats

22:33:11 <mortenf> methinks libby is planning one, no date yet

22:33:23 <DanCon> hmm... do we have a regression suite for rdf/xhtml?

22:35:14 <DanCon> hmm... I don't like rdf-in-xhtml nor xslt2rdf as names...

22:36:21 <DanCon> hmm... esw.w3.org seems kinda slow

22:40:00 <DanCon> sandro? zat you? or just your machine?

22:40:57 <DanCon> oh spiffy; a vandal. http://esw.w3.org/topic/FrontPage

22:43:49 <sandro> It's me.

22:43:58 <sandro> not the vandal. :-)

22:44:37 <sandro> goofy bit of vandalism, yeah.

22:45:07 * DanCon restores FrontPage, notes the address of the vandal in AboutThisService, returns to work on EmbeddingRDFInHTML

22:55:54 <DanCon> I'm thrashing today... between an odd job for a local business, a big ebay transaction, a WG telcon, TAG foo, etc.

23:49:23 <karlcow>http://www.geocities.com/SoHo/Lofts/5614/f76.jpg

23:49:23 <dc_rdfig> F: http://www.geocities.com/SoHo/Lofts/5614/f76.jpg from karlcow

23:49:41 <karlcow> F:| It could be the Semantic Web by Folon, a french Artist :D

23:49:41 <dc_rdfig> Titled item F.

23:51:41 <karlcow> F:[Official Folon site|http://www.folon-art.com/]

23:51:41 <dc_rdfig> Added comment F1.