How ’bout one elegant namespace?
Dave responds….. My responses below…. in a whisper…..
Marc Canter responded to my crucifix post, with a list of metadata he’d like to see included with podcasts. It’s interesting that some of the data he asks for is already defined by the RSS 2.0 spec.
For example, “What is the podcast about?” can be handled by the description element, that’s actually been in RSS since 0.91. What keywords could be associated? That’s the category element. When was the podcast recorded? That could be pubDate (but maybe he wants them to be separate dates, in which case we’d have to go to a namespace, not a problem). What are categories? Hmmm, maybe I misuderstood keywords. What’s the diff betw categories and keywords? Isn’t keyword just a simple case of category? (Please don’t yell at me.)
The only two that clearly require new elements are: “Who is on the podcast?” (There can only be one author element per item, but podcasts can and often do have more than one.) And length (which may be hard to get from the operating systems, although it clearly would be valuable; there’s no RSS element that corresponds to length).
Another place we can and will go for metadata for podcasts is the ID3 information that’s carried along in MP3s. No reason you should have to download the MP3 enclosure to get that info.
Here’s a screen shot of the ID3 info for a Morning Coffee Notes podcast.
I’m not sure what Jon Udell, Tantek Celik or Adam Bosworth would say (3 respected XML guru dudes) - but I’m gonna act like a naive marketing guy and first ask “why can’t ALL this stuff be in one elegantly desigend namespace called ‘audio blog’ or something like that (notice I avoided that ‘Apple’ term. Dave knows I just hate fucking Apple.)
Anyway - that said - here’s some specific feedback on Dave’s reponse….
Who’s in the podcast. This brings up the issue of Digital Identity. There are a number of standards out there - ranging from an rdf schema called FOAF (which Tantek could show you in simplifed XML form called XFN and I would pitch you a subset standard called FOAFnet.) There’s also something called i-names - so we all can keep track of each other and some light-weight identity infrastructure called Sxip Networks, which was created by a buddy of mine named Dick Hardt. Any effort at standardizing digital identity should embrace ALL these standards. I don’t take digital ID lightly. Afterall - what software ISN’T about people?
What is the podcast about? Description tags to me are the body copy of a feed. Sure we could fit a verbal discussion on what the podcast is all about, but I’d prefer a namespace like the Dublin Core. Afterall - meta-data kind of implies structured data. The description element should be for the transcription or annotation or quotes or body copy kind of stuff. IMHO.
What keywords or tags should be associated with the podcast? I have to disagree with you here on your asertion that categories are keywords (more on that below under OpenTopics.) Up here I’d like to say that DEL.ICIO.US and Flickr have really picked up the ball and run with their usage of keyword/tags as a way to aggregate info and photos within an open community. It ROCKS! There is also the ENT namespace that Matt and Paolo came up with and they use in their k-collector aggregator. And folks have connected DEL.ICIO.US with FOAF - so there’s LOTS of interest in using keywords as anchors to connect us all together.
OpenTopics
Which brings us to what Category is the podcast in. This is where we depart on usage of terms. You use the term category as a place to put keywords or tags into. But ‘most’ people (I’m using this as carefully and gently and politely as possible) use categories to mean some sort of taxonomy. A system for categorizing information. Category, sub-category, you know - like Directories. I know you’re the outline and OPML king, so I don’t have to explain to you the difference between the actual keywords or tags - and the categorization of things.
So in this case - categories of podcasts would be
“political”, “food”, “technology”, “blogosphere”, “sports” or “sex”
…while keywords or tags about these categories - which would get attached to individual photos, podcasts or links - would be stuff like:
“tantric sex”, “baseball”, “moblogger”, “Wifi”, “spaghetti” or “red state”.
I know this sounds nitpicky, as we’re just down one level of a hieracrhy, but the way it works in the real world is that individual tags get added to ’stuff’, and that stuff gets put into particular folders or directories. Call it what you will - but they’re two separate beats. One is a category, the other a keyword. In my venacular - I’m not sure how else to call it.
The fun part happens when aggregators sort on keywords and allow anyone to contribute their stuff. That’s what’s going on at Flickr and DEL.ICIO.US.
Podcast was recorded on what date? Time? I’m glad to see you think we should add something for this. This seems pretty important to me and a way to gateway to OpenEvents and shared calendar kind of stuff (”who me ALL the podcasts recorded on this day”) BTW Robert Kaye has a complete namespace for music related meta-data at Musicbrainz.
I hope that this shows we can have cordial, polite, well articulated discussions on these important issues.
Finally which one am I? The short pudgy guy or the tall pudgy guy? I seem to recall Dave and I are about the same height - but I’m more pudgy.
Back in the day - many people confused us as brothers.
And good luck to BigDave - while we’re at it!