[uf-discuss] Regarding POSH and misuse of the microformats logo

Jon Tan microformats at gr0w.com
Sun May 6 05:00:42 PDT 2007

Keith Alexander wrote:
>> I think POSH is useful as a
>> conceptual tool for reifying the definition of microformats:
>> POSH Patterns: semantic practices resulting in meaningful markup
>> Microformats: HTML-based data formats
> I think (at least) 3 distinctions need to be made:
> - 'vanilla' semantic HTML (using non-presentational html markup 
> appropriate to the content it describes)
> - HTML-based data formats (actually, this is what I was looking for a 
> term for when I suggested 'POSH')
> - Microformats (HTML data formats that have gone through the 
> Microformats Process - a canon of html data formats )

Relating this discussion to real world problems we're trying to solve 
for a second, there may be a wider context: In Nov, 2005 I was trying to 
search for office space in Bristol, UK and getting garbage from search 
results. The problem was the legacy gap left between non semantic 
markup, bad labeling and machines. The gap had been filled with noise 
from pseudo-aggregation sites, paid directories and other "Web 
marketing" services -- a problem faced by anyone trying to use Web 
searches to extract up to date, usable aggregated data from poorly 
marked-up Web sites. That problem still persists.

At the time, my experience threw up some thoughts (and a rather verbose 
article[1]) on semantics and specifically what I called semantic 
information design ethics, shortened to "SIDE" for brevity. I see POSH 
and microformats as unique but connected components of a solution to 
free data from Web pages and allow it to be aggregated, discovered and 

If POSH is concerned with HTML-based data formats, or if semantic markup 
initiatives generally would seek to contribute to a solution for this 
current and real world problem then a further requirement has to be 
meaningful use of language in the document. I.e. Not only appropriate 
markup, but meaningful text itself to allows machines to recognise the 
page has a place in a matrix of aggregated data for a given search term. 
(This also applies to gateway pages on sites to proprietary datasets.)

FWIW, I still talk to businesses and individuals who's legacy sites 
don't even have descriptive page titles or n'er a <h> tag in sight. In 
ignorance they are considering buying some SEO or Web marketing activity 
to compensate. It may seem a little simplistic from the lofty heights of 
semantic enlightenment, but there are a huge number of less-enlightened 
colleagues and their clients who would benefit from an initiative in 
this area -- not to mention the browsing public.

A simple checklist of changes (including POSH) that they could make 
today to their markup, use of language etc could have significant 
benefits for everyone and make inroads to solving this real world 
problem. Being cynical, the bottom-line benefits for are obvious should 
be motivation enough for 90% of sites. More fundamental uF, IA, UI and 
accessibility checkpoints could also be included that they could work 

At the moment, without including the wider context for POSH or 
specifying real-world benefits for implementers POSH seems on the road 
to just another geeky acronym that's (almost) cool for those in the 
know. Outreach is critical and, "features, benefits /and/ incentives" 
need to be specified to do that.

Apologies for the lengthy reply. Thanks,
Jon Tan


More information about the microformats-discuss mailing list