[dsdl-comment] Re: FAQ: (1) NVDL versus XProc, (2) Performance of NVDL

Roger,

I have assumed that XProc and NVDL can play very well. In fact,
if we can establish reconstruction by NVDL (as opposed to
dispatching), we can use XProc to control pipelines such as
NVDL dispatching, XSLT transformation, and NVDL reconstruction.

SnRNV already implements experimental reconstruction. I will
soon send some examples of such reconstruction to this mailing
list.

As far as performance, different implementations will
exhibit different results. Here is a result of SnRNV.

I used the NVDL schema shown in A.1 of OOXML Part 5 (DIS).
I used a large document and added 3134 foreign attributes (mc:Ignorable),
which are dispatched to msv from SnRNV.

bash-3.2$ !s
sh c:/NVDL/SnRNV/msvnvdl.sh mc.nvdl document.xml
Open the NVDL file: mc.nvdl
Validate the instance file: document.xml
document.xml is a valid XML document.

real 0m1.044s
user 0m0.015s
sys 0m0.062s
bash-3.2$

For comparison, I tried sax.Counter of Java.

bash-3.2$ time !!
time java sax.Counter document.xml
document.xml: 172 ms (12171 elems, 11137 attrs, 0 spaces, 75420 chars)

real 0m0.772s
user 0m0.047s
sys 0m0.015s

IMHO, the NVDL overhead is almost nothing.

Cheers,

-- 
MURATA Makoto <murata@hokkaido.email.ne.jp>
--
DSDL comments
To unsubscribe, please send a message with the
command  "unsubscribe" to dsdl-comment-request@dsdl.org
(mailto:dsdl-comment-request@dsdl.org?Subject=unsubscribe)