[whatwg] Link rot is not dangerous

[Toby A Inkster]

On 20 May 2009, at 05:23, Tab Atkins Jr. wrote:

> Specifically, people can use a search engine to find information about
> foaf.  I know that typing "foaf" into my browser's address bar and
> clicking on the first likely link is *way* faster than digging into a
> document with a foaf namespace declared, finding the url, and
> copy/pasting that into the location bar.


FOAF is a very famous vocabulary, so this happens to work quite well  
for FOAF.

Consider Dublin Core though. Typing "dc" into Google brings up  
results for DC Comics, DC Shoes, Washington DC and a file sharing  
application called Direct Connect, all ahead of Dublin Core, which is  
the nineth result. Even if I spot that result, clicking through takes  
me to the Dublin Core Metadata Initiative's homepage, which is mostly  
full of conference and event information - not the definitions I'm  
looking for.

On the other hand, typing <http://purl.org/dc/terms/issued> into my  
browser's address bar gives me an RDFS definition of the term  
immediately.

Your suggestion also makes the assumption that there is a single  
correct answer that Google/Yahoo/whatever could give to such a query  
- that any given string used as a prefix will only ever be  
legitimately bound to one vocabulary. That is simply not the case:  
"dc" for example is most often used with Dublin Core Elements 1.1,  
but still occasionally seen as a prefix for the older 1.0 version,  
and increasingly being used with the new Dublin Core Terms  
collection. While Elements 1.0 and 1.1 are largely compatible (the  
latter introduces two extra terms IIRC), Dublin Core Terms has  
significant differences. "bio" is another string commonly bound to  
different vocabularies - both the biographical vocab often used in  
conjunction with FOAF, plus various life-science-related vocabularies.

-- 
Toby A Inkster
<mailto:mail at tobyinkster.co.uk>
<http://tobyinkster.co.uk>