[uf-discuss] citation microformat encodings

[Ryan King]

On Jan 24, 2006, at 3:04 PM, C. Hudley wrote:
> On 1/24/06, Tantek Çelik <tantek at cs.stanford.edu> wrote:
>> On 1/24/06 12:50 PM, "Ed Summers" <ehs at pobox.com> wrote:
>>>
>>> I must admit to feeling a bit confused about how to proceed. Could
>>> those of us who are interested in seeing openurl components in a
>>> microformat create some pages that illustrate how it could be used?
>>> Would this confuse current efforts or add to them?
>>
>> The key point missing is this.
>>
>> Microformats are based FIRST on human publishing *behaviors* on  
>> the *web*.
>>
>> And ONLY THEN do we look at what previous attempts at formats have  
>> done to
>> see if they can help address the problem that has been specified  
>> by the
>> examples documented from the Web.
>
> Fair enough - it says as much on the wiki.  But, was vcard being
> published as such on the web?  What's confusing me is the point at
> which folks become willing to translate a non-web spec to an hSpec.
> Did you go through the whole process translating vCard into hCard?  Or
> iCal and hCal?  Or were these short-circuited somehow because doing so
> was obviously a good idea?

Its important to note that "the process" grew out of working on XFN,  
hCard, hCalendar, rel-tag, rel-license and hReview. So, no, "the  
process" wasn't followed in an explicit sense. However, the work was  
done in mostly the same way, though with a smaller community.

Please understand the historical context of these formats.

> If you did follow the whole process both times, it seems totally fair
> to go through the whole process for translating OpenURL profiles to
> hCitations, or wherever else the process might lead.  And tell us so,
> and we'll study the list archives etc. and get smarter quickly to make
> sure we can help more at every step.
>
> If you didn't go through the whole process, then I have a different
> question to ask. :)
>
>> Existing formats are most useful for the
>> terms and vocabulary they have chosen.
>>
>> One point on OpenURL - as far as I can tell, all the information  
>> about the
>> citation is captured in the URL.
>
> Did you see the recent examples posted here that pull the OpenURL
> profile fields into html classes?  We have translated the OpenURL book
> and journal citation profile keys into HTML class attribute values.
>
>   http://microformats.org/discuss/mail/microformats-discuss/2006- 
> January/002781.html
>
> Note the class name pattern, aside from the COinS bit (the Z3988 class
> value title element, which essentially replicates the more obvious
> class attribute values).  Would it be appropriate for us to put those
> examples into citation-brainstorming?

Sure. This would be classified as a "strawman proposal" - a (possibly  
premature) proposal meant for brainstorming. Its worth recording, if  
only so that we don't repeat the proposal later.

>> The problem is that this is NOT the way people publish citations  
>> typically
>> on the web.
>
> We could litter the wiki with 76 different examples of how journal
> publishers mark up HTML for citations.  They are all inconsistent and
> incompatible.  Because of that, they mostly also use OpenURL to link
> out.
>
> Assuming you don't really want 76 different examples, we could pull
> out a handful
> of these that are better than others.  Some are already there.

Exactly. No need to put 76 different examples, but rather a  
representative sample.

> Forgive us for a bit of frustration, having worked through years of
> inconsistent journal publishing patterns in the 1990s, and then a
> four-year standards-setting process for OpenURL (which formally came
> out in 2004), and then having to start over again here.  We're willing
> to work through the process, it just isn't clear what barrier must be
> crossed before it might be even possible for somebody to consider that
> "perhaps this problem is already solved, maybe we could translate the
> answer."  Or if there is, indeed, no shortcut.

We don't mean to set the bar *too* high. However, we've found that  
the process for creating microformats yields good results by helping  
us to think about the problem in a productive manner. Please don't  
take offense at the general skepticism towards less familiar standards.

>> In short, OpenURL is *not* human friendly and does not convey human
>> *visible* information.  In addition, by its dependency on a  
>> specific "link
>> server" in order to be of any use at all, it does not encourage
>> *decentralized* development.
>
> Neither is vCard, nor iCal, human friendly.   We know!
>
> That's why we're here.  We have been working for two years to free
> OpenURL from the dependency from specific link servers and we want it
> to be useful for decentralized development, and microformats are
> obviously the best place to be for both. :)

Great. Let's work on this! If there's any confusion as to how to  
proceed, please ask, pester and argue (ok, don't argue too much)  
until its clear.

-ryan
--
Ryan King
ryan at technorati.com