Have they asked us how we feel about creating interoperability
nightmares like this?
Seriously: the web specs have always been open. I didn't see much in
the way of proposals before they deployed their mechanisms. I saw the
Mosaic-L10N paper, and I referred them to the MIME specs and the
charset stuff. A long time ago. No response, apparently. (Note:
RFC1341 dates back to 1990).
If they currently enjoy interoperability via ad-hoc techniques, it's
by luck. Any vendor that promised interoperability based on these
techniques is unscrupulous. There's nothing to stop a vendor from
supporting these techniques. But customers should know that they are
depending on a particular vendor (or an experiment, in the case of
Mosaic L10N) when they use them.
If folks want to ensure interoperability, they should represent their
needs in the open standards forums like this one.
We have to weigh the cost of "destroying interoperability" now, since
there is a certain amount of deployment. But those concerns must be
secondary to reliable operation of conforming systems, if we ever hope
to get out of "bugward compatibility" mode and into reasonable amounts
of innovation.
Dan