[microformats-discuss] Re: Educating Others

Stephen Downes stephen at downes.ca
Wed Oct 5 11:30:52 PDT 2005


>On 10/5/05, Mark Pilgrim <pilgrim at gmail.com> wrote:
>>The regulars on this list have all been through the full standards
>>cycle many times.  We know about edge cases, we know about validators,
>>we know about standards.  We know.  We've been there.  We've all
>>decided that this way is better.  Not because it's easier or faster or
>>sloppier, but because *it leads to a better result*.  Really.  The
>>fact that it happens to be easier and faster is just a karmic
This is my observation as well, based on my experience in academic 
publication and learning resources metadata.

In these fields, two major standards exist: Dublin Core, which will be 
familiar to many of you, and Learning Object Metadata (IEEE 1484), which 
will not. As you may know, Dublin Core began as a very simple 
specification, and for the most part remains so, though the 
supplementary DC-Terms was later added. By contrast, IEEE-LOM, which 
began life as a consortium backed a priori specification (IMS LOM), 
started and continues to be a complex specification.

While it is true that IEEE-LOM is well-used, particularly via an 
application profile called SCORM (implemented for the U.S. military), it 
is primarily used by large corporations. And studies of its use (cf 
Friesen 2004, via http://www.edtechpost.ca/mt/archive/000567.html ) show 
that practitioners only use a small number of core data elements; other 
elements are either not used or misused. Dublin Core, meanwhile, has 
become a staple for the Open Archives Initiative and is now used to 
describe millions of academic papers worldwide.

Danny Ayers wrote:

>If it leads to a better result, great. But I haven't seen any
>compelling evidence for that yet. Yep, it works so far, but then so
>does Atom, which followed a more traditional route.
Atom had the advantage of following RSS and is in many ways a clone of 
RSS (and in many ways is not). Without venturing to start a religious 
war, I would nonetheless say that whether Atom 'works' is still open to 
debate. I would ask, is the functionality that Atom provides over and 
above RSS widely used in a non-corporate environment (ie., outside the 
major blogging engines)?

>One thing I do find a little curious. A large proportion of systems on
>the Web have a long tail distribution - there's value in the
>thousands, millions of little thingies, not only the top 10. So I'm a
>little surprised to find so many people around these parts prepared to
>write off edge cases as so much cruft. As the Web grows, surely the
>sum of specific cases approaches the general case..?
This is why the core specification should remain extensible, not why the 
little-used features should be included in the specification.

While the little-used features, cumulatively, add up to be as important 
as a core feature, the little-used features have the disadvantage of 
being all *different* from each other. You can't take a hundred 
little-used features and craft one 'general' version of them all.

Microformats should stay simple. Each microformat should aspire to no 
greater complexity than Dublin Core or RSS. And the number of 'core' 
microformats should remain low. Define just enough to demonstrate the 
utility of the format, and document and clarify extensions (including 
additional microformats) as they become more widely used. Don't try to 
do it all up front. Once the 80 percent line is felt to have been 
attained, call a moratorium to standards development and begin 
developing applications and documentation. There can always be a phase two.

-- Stephen

p.s. I can't take it any more -- I've corrected the spelling of 
'Educating' in the subject line.  ;)

Stephen Downes  ~  Research Officer  ~  National Research Council Canada
http://www.downes.ca  ~  stephen at downes.ca         __\|/__ Free Learning

More information about the microformats-discuss mailing list