[uf-new] video-info-examples page clarifications

Manu Sporny msporny at digitalbazaar.com
Tue Sep 11 21:31:13 PDT 2007


This is a partial continuation of the following thread:

http://microformats.org/discuss/mail/microformats-new/2007-September/000823.html

I wanted to make some clarifications about the video-info-examples page
before we go down a rabbit hole. Or rather... this is an attempt not to
go down a rabbit hole :)

Mary Hodder wrote:
> So the names you use here:    http://microformats.org/wiki/video-info-
> examples#Properties    don't all make sense to me.  Where did they
> come from?  Some of them I've never seen used before.

They are names that I picked from the previous media-info-examples pages
and discussions. Some of them are new because of freshly discovered
properties on web pages that were not identified in the first analysis
pass of video services.

> Published?  that doesn't mean anything without the explanation.  Why
> can't you just use "date" so that we know what you are talking
> about?  No need for long explanations if we do that.

We could use "date". What does "date" mean? Does it mean "published
date", or "released date", or "posted date"... because all of those are
different for the video examples that are currently being analyzed.

I used "published" because it has a very specific meaning in Microformats:

"published" - Publish date of a weblog/microcontent entry[1]

That being said, I went ahead and changed "published" to "post date",
and added "release date" to the set of properties that websites could have.

> Why do you need more examples?  In the new page you just made here:
> http://microformats.org/wiki/video-info-examples
> Why not combine?

I've taken your suggestion and combined the individual publishing
examples of media-info-examples with the video-info-examples page. That
added 10 more examples to the analysis. I think we still need more.

The reason I say this is that somebody arguing against hVideo could say
that we do not have a statistically significant data yet. You said that
there are over 1,000 sites that you collect data from. If we wanted to
state something with a confidence level of 95% and an confidence
interval of 5% in a population of 1000, we would have to take 278
samples. We currently have 33.

There's a middle ground here. I'm trying to make sure that it is much
harder for somebody to dismiss the research that we are doing by falling
back on the "you didn't analyze enough examples" argument. Thus I don't
think that 33 examples are good enough to make our point.

> I'm not really sure what you are getting at with your list of
> properties then?

The list of properties is defined so that everybody understands what
each property means. This is important because we don't want to have any
misunderstandings about the definition of a property name. So, if a
website on the video-info-examples page has a 'trackback url', we can
look its definition up and know that it means the following:

"The trackback URL lets you know who has linked to the current blog
entry."[3]

Perhaps I have done a bad job with the property definitions on the video
examples page? The goal was to create a common vocabulary with specific
definitions for analysis without getting too attached to the terms.

> We know people, all over the web, across the 27 million videos Dabble
> has indexed, as well as the rest of the videos we don't have but are
> mostly hosted by those same 1000 hosters, that this list rings true:

I trust that you have the data and have done the analysis, so would it
be possible to release that data and analysis to this list so that we
may integrate hard numbers into the wiki? Adding 1,000 examples that
have already been analyzed for the Microformats community would be
fantastic!

> By putting price and purchase URL right at the top of the definitions
> you are giving me the impression that you care most about that.

Well, that certainly wasn't intended. I've added the following text to
the properties section, does the following help?

"These properties are in alphabetical order and in no way represent the
frequency of their use in the examples. The property names are also not
final and probably will not be used when the Microformat vocabulary is
decided. Deciding the vocabulary of the Microformat is not performed at
this stage of examples collection and analysis. These property names and
definitions are listed here in an attempt to keep the current and future
example analysis teams using the same definitions for property names."[3]

Here's to hoping that this e-mail helped clarify more than it muddied :)

-- manu

[1] http://microformats.org/wiki/existing-classes
[2] http://www.surveysystem.com/sscalc.htm
[3] http://microformats.org/wiki/video-info-examples#Properties



More information about the microformats-new mailing list