[
Lists Home |
Date Index |
Thread Index
]
> Schematron has a standard reporting format already:
> http://www.schematron.com/iso/P20.html
Yes, and its easy to over-ride that standard format with one that is
specific to your needs. This is what we do for message validation and
have *all* validators implement the same output format, so +1 for
schematron :-). I was exploring the possibility of JNVDL doing
something similar. Jirka positions JNVDL as a 'low level library' and
offers some possibilities in the (to come) 'Relaxed Validator' product
wrapping, so maybe comments such as these will resonate more strongly
with that effort.
(Note: http://www.schematron.com/ seems to be down at the moment ???)
> Standards very often on specify half the necessary picture:
Agreed, but sometimes this IMO is an inhibitor to adoption. We have
had this discussion before on another thread about the desire for a
reference implementation of a standard to help 'kick start' it use and
help it gain traction in a market place full of 'good ideas' that fail
at the '... for practical use' stage. Some see this as important some
don't. I'm in the camp that say it is, just from the experience of
trying to convince project managers to allow investment in
technologies that are often better than an 'insipid' mainstream vendor
packaged offering but are preceived to carry a greater operational
support risk. Many (most ?) large corporates are fairly risk averse
and conservative in this regard. The more complete the implementation,
the easier it is to demonstrate its value. As soon as I have to say
'.. well we had to write a bunch of additional custom code to fill in
some of the blanks..' I can see the shutters go up.
Fraser.
On 12/05/06, Rick Jelliffe <rjelliffe@allette.com.au> wrote:
> Fraser Goffin said:
> > I was kinda hoping that JNVDL was going to do a bit more for me (like
> > providing an aggregated view of all of the validation errors from all
> > of the validators called, in a standard vocabulary (maybe even
> > allowing me to define what that vocabulary should be in some way) -
> > and, if I want it, the results for each namespace individually).
>
> I think the WG would be happy to consider adding a standard message
> reporting format, if someone contributed it. One thing holding us back is
> the desire to have running code, but I guess this is a quite well known
> issue.
>
> What would be great is if XML-DEVers could cobble together something, for
> example a set of requirements or a draft. Would you prefer the messages be
> given in terms of line numbers or XPaths?
>
> Schematron has a standard reporting format already:
> http://www.schematron.com/iso/P20.html
>
> > When
> > I think of COTS tooling I don't normally expect to write half of the
> > implementation myself ;-).
>
> Standards very often on specify half the necessary picture: XML & SGML for
> example.
>
> Cheers
> Rick Jelliffe
> Member WG1/SC34/JTC1
>
|
