[microformats-discuss] Re: playlist microformat

Chris Messina chris.messina at gmail.com
Fri Sep 9 12:08:54 PDT 2005

I wanted to chime in briefly at the meta level (ayuck) to try to help
the hPlaylist discussion evolve productively, in a way that's most
supportive of the microformat mission.

* My first point is that microformats ought be micro. When I first
took a look at all the proposed classes at
http://gonze.com/microformats/xspfxmdp.html, my eyes bulged a bit. I
did a reality check and took a look at hCard and hCal and though they
contained many potential classes, they started out very simply.

As a test, as a web developer or blogger, I can remember the basic
classes fairly well, so I can actually use the MF when I'm
coding/composing. If I were to write up a playlist of tracks, I would
want the microformat to similarly serve my purposes.

* My second point is that microformats are supposed to be used like
building blocks (that's what the logo was designed to look like, FYI).
So in this case, there seems to be little value in reinventing XOXO
just to make a flat outline for media. This plays to my first point --
the hPlaylist/mPlayItem classing should be fairly focused and atomic.
Use existing microformats as much as possible and extend or add only
when absolutely necessary (i.e. you can't do what you want to do with
what already exists).

* On top of that, we're really discussing two separate tasks here. The
first is marking up a series of items in a specific order -- i.e., a
"playlist." The second aspect is marking up media items like movies,
mp3s, images and so on. These are typically binary objects that need
to be described externally and linked to remotely. Those items need
not be found in playlists to be useful, so really we're describing two
separate microformats.

Hence I've suggested that we use XOXO for the outline-level organizing
of the media items and come up with a standard way to mark up the
individual media items.

* And, as David has pointed out, the playlist spec should be
universally useful and not simply apply to mp3esque media files.
Although the work so far suggests that hPlaylist apply specifically to
mp3-style lists where ID3 + m3u might have sufficed before, we're
seeing combination playlists now that represent songs, music videos
and digital booklets
(http://www.flickr.com/photos/factoryjoe/41771361/). The hPlaylist
work should accommodate this -- as well as, for example, the work
being done on H20 playlists (h2obeta.law.harvard.edu).

Anyway, thought I would toss out these thoughts and observations in
the hopes that can begin moving this discussion closer to something
that feels more like an instantly usable (and comprehensible)
microformat and less like a macroformat for marking up metadata in

I'm really excited about seeing something emerge here, since I need
this for my work on social bookmarks!


On 9/3/05, David Wiley <david.wiley at gmail.com> wrote:
> I'm very interested in making sure that hPlaylist turns out to be
> generic enough that I can use it to sequence arbitrary multimedia
> assets. 

>Today there is no simple
> standard for expressing these sequences. The hPlaylist format will
> certainly be useful for this if it will support arbitrary media, which
> at first reading it appears to do.

More information about the microformats-discuss mailing list