[uf-new] New proposal - replace custom web analytics markup with microformat(s)

[Liam Clancy]

Hello µF community,

I have been working on a particular problem and believe that the  
creation or adoption of microformats would not only solve it, but be  
of benefit to a great many web developers and help push the adoption  
of microformats within enterprise companies.

The generic use case I have is quite simple:

1. Large company buys an expensive Web Analytics type product from a  
vendor such as WebTrends, Omniture, Coremetrics or other (there are  
lots of smaller vendors in the EU)
2. As part of the implementation the large company must add good  
quality metadata to *every web page* on their site(s) to provide the  
purchased product with enough information to exercise its many features.

Currently the common means of adding metadata to the web pages is  
through proprietary and custom JavaScript mark-up that is time  
consuming, error prone, opaque and expensive to manage for the large  
company in question, e.g.

  WT.pn = 'Homepage';  // WebTrends markup
  s.pageName = 'Homepage'; // Omniture markup
  pageTracker._trackPageview('/Homepage'); // Google Analytics

Until recently there was little incentive for the vendor companies to  
improve on this, as the custom mark-up and its implementation provided  
a barrier to product switching and generated revenue through  
implementation consulting.

I've been lurking here for quite a while, and 18 months ago I was able  
to start working full-time on better solutions for these product  
implementations and soon came to the conclusion that microformats were  
the way to go. I have a post outlining this thinking in more depth here:

  <https://jshub.org/blog/2009/10/19/designing-hpage/>

We setup <http://jshub.org/> to promote this and other ideas to  
improve quality of data collected on web sites, and also to put the  
declaration of this data back into the hands of the content creators,  
avoid side-channels and establish new best practices in a similar way  
as current SEO and Accessibility techniques evolved. The idea here is  
to encourage the publishing of semantic content that is measurable by  
design.

One of our primary projects at jsHub.org is to establish a vendor  
neutral format for the data mark-up (and the Open Source tools to  
support it) and we have received very favourable feedback from all the  
vendors we have met with, including the biggest vendors  (WebTrends,  
comScore, Coremetrics, Google) on the use of microformats.

As part of this consultation we published a draft microformat we  
called 'hPage' (not to be confused with an earlier proposal to this  
list by the same name):

  <https://jshub.org/hPage/>

This proposal was developed using the published microformats process  
as much as possible, and it has always been our intention to share it  
with the microformats community for review and refinement and develop  
a final specification that can be adopted.

I apologise in advance for the initial unilateral development and  
labelling of this draft as a 'microformat', our circumstances have  
been very much the same as that of hNews, and I would like to ask for  
any and all feedback on the current proposal, and to start  
transferring our research and other content to the official  
microformats wiki for further development.

There are a number of initial questions that I believe I can anticipate:

1. We understand that this may never be a 'proper' microformat - our  
usecase is rather generic, but the implementation of a solution has  
specific constraints to be commercially viable. I and the other  
members of the jsHub.org team have a lot of experience in this due to  
previous work for the vendors (which we will be sharing), however I  
would like to point out the 'Design principles' section of the spec  
where we attempt to address some of the conflicts in approach and how  
we try to mitigate them:

  <https://jshub.org/hPage/#design_principles>

2. Its called 'hPage' to create an association with the Web Analytics  
vendors measurement of a 'Page View', which is the primary use of the  
proprietary mark-up it is intended to replace. May vendors have an  
enormous number of potential variables that can be set for use across  
a large product range and we don't believe that one microformat can  
replace all of these, only a common subset. We expect that there will  
be additional proposals needed, but we are starting with the most  
obvious current practice and trying to address the next required  
development in the Web Analytics industry - that of tracking multiple  
pages of content in one document (AJAX updates). This is covered in  
the section:

  <https://jshub.org/hPage/#multiple_pages_of_content_in_a_single_html_document 
 >

3. As with hNews we suspect that  there are potential benefits to  
integrating with or extending hAtom and other microformats and would  
like to first start with seeing how we can revisit our initial in-the- 
wild examples using your collective experience. Our objective is to  
establish 'a' microformat, not necessarily 'our' microformat for use  
by clients and vendors alike.

We are working full-time on this and other related projects at  
jsHub.org and I appreciate your time and am very much looking forward  
to hearing your opinions.

[If anyone is London, UK based I am generally out and about if you  
would like to meet up]


Liam Clancy
-- 
http://jshub.org/
http://metafeather.net/