[
Lists Home |
Date Index |
Thread Index
]
Tatu> This approach is obviously less generic than NVDL, but
Tatu> seems reasonable for efficient validation within
Tatu> context of stax (pull) processing.
I am not sure if your approach is less generic. At a first
glance, NVDL looks complicated but I would not be surprised
if your approach can be safely extended for handling NVDL.
NVDL implementations have been built using SAX. However, it
is not easy tell, by reading the NVDL spec, whether NVDL is
implementable using SAX. To the contrary, most people will
suppose that it is not. The same thing might apply to
pull-based implementations.
Having said this, a single <namespace> element of NVDL
having more than one <validate> element might be a problem
for pull-based NVDL implementations. We have to allow a
single start tag to be *pulled" by more than one validator.
Cheers,
--
MURATA Makoto (FAMILY Given) <EB2M-MRT@asahi-net.or.jp>
|
