[uf-new] restarting the [citation] microformat effort

[Tantek Çelik]

On Mon, Aug 6, 2012 at 6:47 PM, Ed Summers <ehs at pobox.com> wrote:
> On Mon, Aug 6, 2012 at 9:23 PM, Tantek Çelik <tantek at cs.stanford.edu> wrote:
>> Greetings,
>>
>> I am restarting the citation microformat effort with a new brainstorm
>> proposal that is:
>>
>> 1. greatly simplified (as compared to previous citation proposals) in
>> an attempt to solve a simpler problem first (per the microformats
>> principles)
>> 2. uses only microformats2 syntax
>>
>> The goals here are to be modern, minimal, and immediately easily usable.
>
> I'm glad to see a simplified effort on citation.

Thanks. I've certainly stayed away from citations for a while as it
seemed like everyone who attempted citation related formats (whether
in microformats or elsewhere) ended up going down various unproductive
ratholes, so I figured I would end up at the same place(s). Enough
time has passed that it seemed reasonable to stubbornly try again to
see if we can achieve a different result by reframing (refocusing) the
problem. Here's to hoping.


> You said:
>
> """
> They didn't design a citation microformat that could be used as a
> building block, but rather, erred on the side of attempting to
> describe the myriad types of references to dead-tree resources.
> """
>
> This was a while ago, but if memory serves we went down this road
> because you (ahem) insisted that we pave the cowpaths of the various
> citation formats that are used on the Web :-) But that's water under
> the bridge I guess.

Could very well be, as I used similar (flawed) reasoning early on for
the re-use of the entirety of vCard and iCalendar's Event object,
instead of subsetting them to match (pave the cowpaths of) the 80/20
of what *data* (examples) are actually published. Happy to accept any
amount of blame on this :)

Subsequently (2008) I corrected this flawed reasoning in the process,
and addressed it in a process FAQ:

http://microformats.org/wiki/process-faq#Can_a_microformat_be_class_names_from_another_format_vocabulary

The explicit reference to Dublin Core in that FAQ makes me think that
the citation work at the time may have been the tipping point for that
particular insight.

> So in your web-to-web example, what would happen if you remove the anchor tag?
>
> <span class="h-cite">
>   <time class="dt-published">YYYY-MM-DD</time>
>   <span class="p-author h-card">AUTHOR</span>:
>   <cite class="p-name">TITLE</cite>
> </span>
>
> Would that be an acceptable web-to-off-web citation?

Yes, I think that could work well.

Next thing to do would be to try it with actual content being
published on the web (perhaps a bibliography, list of papers in a
resume, etc.) and see how it works/feels there and what seems to be
missing. The "p-publisher" property might also be useful in some
web-to-off-web citation use cases.

It's at least an incrementally useful semantic/structural improvement
over *only* using a <cite> element, and maybe that kind of small
evolutionary step is good enough to enable progress.

Tantek