Should the XML "recommendation" have an expiration date?

["Costello, Roger L." <costello@mitre.org>]

Hi Folks,

In the article "Vulnerable Compliance" the author Dan Geer suggests that standards [recommendations] should have an expiration date:

    If one interprets a standard as a kind of license, 
    then perhaps standards should come with an 
    expire-by date. Some attacks that are not possible
    in today's state of the world may become possible
    in the future and invalidate the design environment
    in which a standard was crafted. 

Is the design environment in which XML was crafted (circa 1998) invalid in today's state of the world?

There are a number of potential benefits to having an expiration date on XML:

1. A looming expiration date will force a lot of smart people to work on creating something better -- less vulnerable to attacks, faster, smaller, and more capable. (Necessity is the mother of invention.)

2. Constant churning and change seems to be consistent with the way nature and complex systems work. (I guess this isn't a benefit, it's more of an observation.)

3. It will create jobs.

4. It's fun to create new things and people like new things. There will be lots of conferences and workshops, attended by thousands of people.

There is, of course, a disadvantage:

    It will take significant effort and money
    to replace the existing tools and standards.

Do the advantages outweigh the disadvantage?

What do you think, should the XML recommendation have an expiration date? If yes, what date would you set as the expiration date? 

/Roger