Lists Home |
Date Index |
Let me expand more on the "requirement info set" that an interpreting
environment needs to proper handle the document.
a) every Datuments or web application needs to include some meta data like
for instance, its dependencies with specific versions, modules,
b) We know that relying on the browser to figure this out is a mess most of
Actually, this kind of meta data can be included in every HTTP transaction
headers but it is very hard to encode HTTP headers in URL or authoring
tools. You'll need to expand the HTTP server basic functionality with some
code to do so. Thus, HTTP headers to specify the required capabilities are
often a show stopper.
The other way is to include the meta information into the document itself
without damaging the document's structure. RDF could be a good candidate and
a client side engine can extract, decipher, interpret this RDF fragment to
check is every PRE-CONDITIONS are met and suggest some mitigation strategies
or re-direct to a downgraded version otherwise.
For instance to take Peter's example, the CML document would include an RDF
fragment to indicate the required capabilities and redirect to another
document containing only images for the other environments in case these
requirements are not met. The user may get the choice to go for it anyway by
not following the suggestions and trying it anyway.
Another way is to add a new parameter (i.e command) to be interpreted; for
example, a parameter pointing to an RDF document containing the
requirements. This way, the requirements are decoupled from the document and
can easily be updated.
Didier PH Martin (from Didier's lab)