Lists Home |
Date Index |
Michael Rys wrote:
> I am not close enough to their thinking to know
> how much they want to have people program towards BAML
If they didn't want people to use BAML, then why did they
expose the BAML serialization interfaces in the Longhorn
documentation? The fact that these interfaces are public certainly
seems to indicate to me that someone expects/wants them to be used...
> [Michael Rys] Only if people are going to start
> defining interop on this level.
People started doing interop with binary encodings long ago. I
won't bore you with references to X.400 and X.500 since you might say
that they are "tightly-coupled" applications. So, let's look at
something like the NCBI (http://www.ncbi.nlm.nih.gov/) the US National
Center for Biotechnology Information, that among other things runs the
GenBank library of genetic sequences. Long ago, they settled on ASN.1
defined binary encodings for interop. The GenBank format provides for
general interop between a large variety of applications in the genetic
sequence space. This is only one of many "interop" applications that
currently use ASN.1 based binary encodings. (Note: The recent effort
to produce XML encodings for NCBI data has resulted in a minimum 10x
expansion of file sizes...)
> XML is the one format that is defined as
> the "interoperable standard format".
No. XML is not *the* "interoperble standard format." It is
only one of them.