>It will not address the transport protocal that is part of OpenMath
>(in my opinion, far and away the most important part), and that part
>will still be left open for OpenMath to solve. WRI volunteered to
>donate a solution (MathLink), but NIH rules there as elsewhere...and so
>they still don't have a solution.
I think it critical to _not_ attempt to address that issue in the markup
definition. About the same time the Web was being put together CMU put together
an "all or nothing" package called Andrew. The system was pretty unsuccessfull
getting other places to adopt it until they broke it up into AFS, what became
the basis of MIME and a number of other usefull modules.
There are many people looking at the transport issue and at the moment the lead
solution technically seems to be ACTOR based systems such as ActiveX.