I suppose it's worth fleshing out what I mean by modularization a bit
more, because I think it's all that's neccesary to infer type.
Suppose we have a core citation format, such as Tim White shows,
containing the following properties/classes. hCite
Author (hcard)
Title
Date
Collaborators
Description
Catalogue Number
Then we have properties that are specific to books/journals
Pages
Volume
If these properties are present, then we know that this item is
probably not say.. a photo or a painting, and contains all the
properties which allow it to be pased the same whether it's a book or
a journal. Combine it with hCite and suddenly we have bookCite
The properties specific to artwork might be:
medium
dimensions
add them to hCite and we have artCite
Then suppose we have properties specific to a photo
Aperture
Fstop
Camera
We add those to artCite and suddenly with have photoCite, demonstrated.
<cite class="hcite">
<p class "author fn">Ansel Adams</p>
<span class="title">Siesta Lake</span>
<span class="camera">8x10 view camera</span>
<span class="medium">Gelatin Silver Print</span>
</cite>
From the presence of "camera" we can glean that this is an instance
of photocite. But a generic parse can still glean the Author and
Title properties. A domain specific parser has the extra data it
needs for cataloguing, or whatever other task required. The domain
specific parser could safely ignore hCites lacking any of the
properties required for photoCite. Etc. etc.
In short, one core format that everything can understand, with
properties available for domain specific applications. The careful
categorization and "branding" of each module helps to keep things
simple for site authors.
Basically I'm basing this off the "modularization of xhtml". For
instance, most site authors only need the basic modules for xhtml.
They have no need for something like the Ruby module. But there's a
large portion of the audience that does, and when they need that,
it's available as a module.
Does that make sense?
On Mar 29, 2006, at 7:17 PM, Tim White wrote:
Well, this is a lot to process at the end of the day. Here's just a
few
of my initial thoughts.
First, and I've asked this before, what are we trying to do? For me, I
just want a *simple* way to mark up books, be it a title, title &
author, or slightly more.
We are NOT replacing OpenURL, etc.
We are NOT building library/scholarly citation records
(in my opinion)
Those already exist and, as has been shown on the list, are very
complicated. They also serve a specialized audience and I don't think
reflect the 80/20 of general users.
The format should be as simple as possible.
As for type attributes (ie, class="book"), Bryan Suda and I had a
lengthy discussion a while ago about that. I too believed it was
necessary, but came to see that it is purely extraneous metadata. Look
at a sample citation, something like:
R. Buckminster Fuller. Operating Manual for Spaceship Earth, Pocket
Books, 1970, pp. 13, 14.
No where does it tell you what this is. We infer (from the blog
post in
this case) that it is a book. Or, we look it up via Amazon or library
card catalog to find that it is a book.
Think of hCard. For organizations do we include a type identifier?
I.e.: <span class="org webdevelopment">Webs - R - Us</span>.
A simple format also makes the MF usable for more than books. Works of
art have been mentioned. Just use the same layout:
Edvard Munch. "The Scream", 1893.
It still has a creator, title and date.
--- Alf Eaton <[EMAIL PROTECTED]> wrote:
OK, so a minimal microformat for a citation could look like this:
<x class="citation [type]">
<x class="title">Item title</x>
<x class="creators"><hcards></x>
<x class="container citation [type]"><hcitation for the
container></x>
<x class="pages">n-n</x> [and anything else specific to this
particular type of citation]
</x>
This seems to be on the right track; similar to what I had in mind.
At work, we have need of a citation microformat and are going to be
using mark up like this for now:
<div class="citation">
<span class="articleTitle">"Accelerated Aging: Human Progeroid
Syndromes."</span>
<span class="author">Author Name</span>.
<span class="pubTitle">Encyclopedia of Aging</span>.
<span class="volume">Vol. 1.</span>
<span class="pubCity">New York:</span>
<span class="publisher">Macmillan Reference USA,</span>
<span class="pubYear">2002</span>.
</div>
It's not perfect, but it fits our needs. Transforming that:
<cite class="hcitation">
<span class="articleTitle">"Accelerated Aging: Human Progeroid
Syndromes."</span>
<span class="author vcard"><span class="fn">Author
Name</span></span>.
<span class="pubTitle">Encyclopedia of Aging</span>.
<span class="volume">Vol. 1.</span>
<div class="publisher vcard">
<span class="locality">New York:</span>
<span class="org fn">Macmillan Reference USA</span>,
<abbr class="dtpublished" title="2002">2002</abbr>.
</div>
</cite>
I know it isn't perfect, but it's based on reusing existing MF, and (I
hope)in keeping with the principles.
(In looking back at it, wouldn't it be possible to do only on vCard,
perhaps way up in <cite>, that would encompass the author and
publisher? Those who know parsing (Brian S.) -- does that screw up
hCard parsing?)
~ Tim
<a href="http://www.tjameswhite.com";>www.tjameswhite.com</a>
<a href="http://www.spreadfirefox.com/?
q=affiliates&id=12227&t=1">Get Firefox!</a>
__________________________________________________
Do You Yahoo!?
Tired of spam? Yahoo! Mail has the best spam protection around
http://mail.yahoo.com
_______________________________________________
microformats-discuss mailing list
microformats-discuss@microformats.org
http://microformats.org/mailman/listinfo/microformats-discuss
_______________________________________________
microformats-discuss mailing list
microformats-discuss@microformats.org
http://microformats.org/mailman/listinfo/microformats-discuss
