RE: Announce: A brief history of SOAP

Just watching the thread but decided to jump in this time :-)

> I'm not disagreeing with the idea that metadata should be 
> standardized -- I
> think it should.

Standardizing metadata is the base for new things to come out. The least it
can do is define global layers of meta-communication.

> All I'm saying is that a given implementation shouldn't be required to
> interpret the metadata, if it's available in some other form (like
> programmer docs).

No it shouldn't, but making best practices public would be as valuable as
the spec itself, something like "implementation specs about the spec" but
not necessarily coming from the original spec authors. 
Designing implementations is (in the spec) is harmful IMHO, and it does make
the process of growing the spec slow and a lot more complicated that it has
to be. Also, it adds into a spec's complexity towards developers thinking
about adopting  it.

> Specs which require the (programmatic) interpretation of standardized
> metadata place a large burden on the implementors of the spec, IMHO
> unnecessarily.

So very true. The "spec" is supposed to work on the data layer only.
Otherwise, you loose focus by thinking of possible applications. The spec
should be neutral (well, as possible). We are having the same problem in
HumanMarkup... some think of specific applications while others try to keep
the specs easy and with wide range of applicability.

Kindest regards,

Manos

"In which level of metalanguage are you now speaking?"

Received on Wednesday, 4 April 2001 08:18:03 UTC