Microsoft Continued Standards Abuse, PDF and the XML Paper Specification

Andy Updegrove and Bob Sutor have both well covered the announcement that Microsoft is now going to submit its PDF competitive "XML Paper Specification" to ECMA for standardization, which means they'll no doubt run it to ISO after ECMA so as to claim an international standard. 

The economic point of standards is to encourage and enable multiple implementations of the thing specified providing choice and risk mitigation to customers.  A "standard" with one implementation is a vendor specification regardless of standards body imprimatur -- the economics benefits the vendor by enabling an ecosystem around their product.  So a successful standard needs to have multiple implementations.   

But there is another quality shared by successful standards.  Successful standards are based on existing practice and experience across a technology area.  There is a richness of trials, product successes and failures, over a period of years that feeds into the development of the specification that gets standardized once the right level of abstraction begins to become apparent.  You don't standardize new things because they need the ability to change and evolve.  You don't standardize a new product, because while standards invariably have mistakes, you don't deliberately want to standardize a new product's bugs. 

As Andy points out, Adobe's handling of PDF was indeed heavy handed.  But by the time PDF was launched by Adobe 10+ years ago, there was a wealth of broad industry experience in document formats, markup languages, page description languages, printer description languages and the appropriate levels of separation between them.  The format has stood the test of time with little change.  There are multiple readers and writers in the marketplace.  With the completion (finally) of the ISO standard, one can claim PDF is a successful standard. 

For Microsoft (or anyone else) to introduce a new and potentially innovative technology in the space should be encouraged.  Just because there is an accepted standard way of doing things should not mean society (through the marketplace) should ignore the opportunity to evolve a technology space and solve new related problems.   To publish the specification to encourage an ecosystem to explore and evolve the space (to the developing vendor's advantage) is also to be encouraged because that's good business practice, AND it provides the real-world economic test of the marketplace.  It also allows the vendor to reap the economic rewards of solving a new problem space for customers.    

But that's as far as something new should go.

Once again we have Microsoft dumping its new product specification into ECMA, with a scope that says the standard must align with the product.  They originally presented the format as their PDF killer as if that was somehow a feature customers need, so we also  understand they're not evolving anything, they're just competing with an accepted standard. 

I'm a big fan of Heinlein's (or Hanlon's) Razor, but regardless of whether this is malice or stupidity on Microsoft's part, it's very wrong from a standards perspective. 

We need just to say "no".

Comments

[Stan]

If the ISO lets in 2 standards then why not 102? The question will go from "Is there and ISO Standard?" to "Which ISO Standard?". If it does then that is the end of ISO.

[Jo Walsh]

So much criticism is directed at OOXML over this. But do other ISO standards emerge from a series of "product successes and failures, over a period of years"?

ODF, for example, is based on a one-product specification, OpenOffice.org's internal XML format. OO.org is maintained by Sun at considerable expense, it is not a collectively maintained open source project, Sun and IBM exercise influence over the ODF committee at another pay-to-play industry consortium. Consortia change their rules to encourage Microsoft's membership.

To "punish" MS for its behaviour by seeking to block its standard at ISO is admitting that ISO certification is used as a political tool, a business tool. Is this far from the truth? It is a crisis of compromise in the standards system, rather than the actions of any specific players, which most needs criticism and change.