Semantic Web Interest Group IRC Scratchpad

Welcome to the Semantic Web Interest Group scratchpad generated automatically from discussions on IRC at port 6667 channel #swig by the chump bot, instructions in the chump user manual. Please use UTF-8 charset on IRC and pastebin for code or data more than 10 lines long.

Nearby: IRC logs (Latest) | semantic-web list | W3C Wiki (Recent changes) | delicious swigbot

last updated at 2002-03-01 23:42
Musicbrainz file tagger (Win32)
danbri: A quick review of for Windows. I'm trying it on a box with only one MP3 file.
danbri: Installs very easily. Simple browser-like UI. Feed it MP3s and you can describe them using 'Artist','Album','Title','Track num', 'Trm'(computed audio fingerprint) and Track ID.
danbri: I found a few minor problems. The tabbing order is backwards (ie. if you hit tab, it behaves as if you'd hit shift-tab in most apps).
danbri: The 'save' dialog is greyed out, inexplicably (until you realise that you need to have entered your MusicBrainz userid/passwd in the preferences). It doesn't tell you this until you hit 'write ID3 tags to file', which is backwards: you shouldn't need a password/account to change your local file.
danbri: Metadata issues: it doesn't make it clear how to express "don't know" or "not applicable"; I suspect this will encourage people to enter things like "Unknown", dirtying the database. It isn't clear what 'track ID' is (seems to assume one is describing the files on a CD, which my laptop doesn't have).
danbri: CD ROM issues: I don't have a CD-ROM, yet when I tried running a second instance of MB Tagger, it complained "lookup error: cannot read CD-ROM table of contents". This might be my fault to some extent, as I told it to accept Tagger as the action for CD inserts (even though I don't have a CDROM). Perhaps it assumes that a 2nd instance running happens only when a CD is inserted?
danbri: User accounts: the msg I get when I hit 'save to ID3' talks about a browser page telling me how to sign up; I don't know what this means, there's no page obvious from here, perhaps the msg text was copied from a web-based tagging tool?
danbri: That's all I can think of right now. I should find out how to get an MB account and try it on some more files.
danbri: Oh, it could be clearer about what 'save' means; the button (I think) means 'save metadata to web service', not 'save inside ID3 tags', which has a different tag. Also, it'd be nice to save a copy of the metadata locally in an rdf/xml file, thouhg I accept that's a rather geeky preference (and might raise data consistency issues if the ID3s were subsequently edited).
danbri: Interesting tool anyway, shuold be really useufl :)
danbri: (Anyone want to port it to Linux?)
danbri: Includes trmgen, for audio fingerprinting music files (I was looking for a more recent version though...)
dajobe: changes "In line with the recent decision from the RDF Core working group on the use of xml:base in RDF 1.0 description"...
dajobe: ..."PRISM descriptions SHOULD NOT use it."
danbri: Bookmarked as an interesting metadata application for images. The metapixel-prepare script outputs summary metadata describing the general colour content of the image.
JibberJim: Sample SVG graph output from "Simplistic RDF Editor" -
danbri: A quick test (may not parse yet) to see what it'd look like to put some key fragments of SVG inside the RDF graph.
danbri: Downside is that it involves coming up with a new, RDF-happy SVG syntax (or a subset). We can't simply "use the SVG namespace"
danbri: Upside is that RDF tools could easily dump out interesting SVG files based on this info, without needing an XML parser, XPath, DOM etc to deal with SVG syntax.
libby: the truth is out there...
libby: ...maybe
gk: This reminded me of RDFweb
libby: a great site for information about chinchillas
danbri: Chinchilla extract - a fundamental enabling technology to facilitate Semantic Web rollout and adoption
libby: veronika and sabrina are finding out about the chump. they reckon it would be useful in school
danbri: So soft... Sooooo... sooooffttt.
libby: aw
libby: rocky the rock rabbit rockrabbits are almost entirely unrelated to chinchillas
making SVG annotations to SVG images
chaals: when a graphic annotation is made to a graphic, it should have the graffiti effect, rather than having to be a seperate picture
chaals: which essentially means including the original picture as an <image> element in an annotation of type SVG
danbri: Are you talking about Amaya-Annotea annotations, or more general sense?
chaals: To do it to a png/gif/jpg it is more a case of gluing things on top of each other
chaals: in terms of annotea annotations (but not necessarily in Amaya)
danbri: Ah, right. I'd like to understand how much of the annotations payload goes in the RDF, versus pointers off to external content. There are several models already, exploring SVG complicates things further :)
chaals: The idea would be that the image inside the annnotation payload consists of the scribble over the top of the original image, and a single image element to include it visually
libby: "XMP support for embedding metadata in application files to easily repurpose, archive, and automate files in a workflow"
libby: thanks Yves for telling me about it
libby: see also XMP document from adobe
libby: "XMP encompasses the following: framework, schema, XMP packet technology, and the XMP Software Development Kit, which is available as an open-source license. XMP is based on the W3C's open standard for metadata, known as the Resource Description Framework (RDF)."
libby: :)
AaronSw: A new RDF query language (more of an RDFPath sort of thing) from 4Folks.
AaronSw: via XMLhack
Nocturne on an RDF Schema for topic maps at KT 2002
larsbot: I'm holding a nocturne at KT 2002 where we will try to design an RDF Schema for topic maps.
larsbot: The goal is to see to what extent topic maps can be realized as an RDF application, and also to see to what extent TM semantics can be useful in RDF.
larsbot: Contact me if you have questions.
danbri: Latest in the popular Semantic Web Vapourware series... I'm trying to use Batik, the Apache Java SVG library, to use SVG as metadata about the "who and what" of photo image substructure. Idea is to use the SVG2JPEG rasterizer to output interesting views of the original picture, eg. blank out everything except one person, zoom in on a feature etc.
danbri: Nearby: nice piece on Batik, original SVG-FOAF Amaya writeup by danbri and chaals, and RDFWeb photo metadata writeup...
danbri: Current status: I've implemented one line! Here it is (or see Makefile): java -jar batik-rasterizer.jar rdfweb-chaals-and-dan.svg
danbri: It outputs a slightly wrong image in PNG (or JPEG in this case), in that for no good reason it crops to the top left bit of the image (ie. charles).
danbri: I think this is because the SVG is hand edited (hand fixed) from the Amaya-generated original. I added xmlns for xlink, and (possibly bogus) width and height values for the image element. I based those on the dimensions of the photo. A mistake?
danbri: Time to find an SVG validator, I guess... [rummages]
Created by the Daily Chump bot. Hosted by PlanetRDF.