[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: RelaxNG examples
Hideki Okita wrote:
Andy,
Thanks for your comment.
As I said in the first mail, I agree that XSD is difficult
to read "directly".
Normally, to develop applications using XML data, developers
at first construct classes and generate XSD description from
these classes by some XSD generation tool.
In addition, they use GUI to draw classes and XSDs.
And, when they include and validate the XML schema like the
NETCONF protocol schema, they use integrated development
environment (IDE). They do not read directly the XML schema.
Which expression is best for the WG documents?
Do not we need to write drafts more visually?
I am not comfortable with the idea of standardizing text that
many people to not "directly" understand, and just trust that
all toolmakers will get everything exactly right and exactly
the same. I don't use a GUI tool to write XSDs. Is the IETF
going to mandate that I use some tool, just like it is
effectively mandated that xml2rfc be used to produce an
Internet draft?
Best regards,
Hideki Okita
Andy
Andy Bierman wrote:
I do not rate XSD ease-of-understanding as good.
I rate it as awful, even unacceptable.
Human factors are very important here because humans, not machines,
need to understand the data models that are being proposed and
developed for standardization in IETF WGs.
IMO, most of the humans working in the NETCONF WG do not understand XSD
well enough to throw down hill 5 feet. This is a significant
process problem, since it is difficult to reach consensus
on complex technical details if very few people actually
understand what they are reading.
A comment was made at the data modeling meeting that W3C is now
doing development in RelaxNGc, and converting it to XSD for normative
reference in the final standard. How clueful. How about if we do
the same thing? We don't care if the conversion isn't pretty.
(Nothing about XSD is pretty, so it's a non-issue. ;-)
Note that there is no guarantee that people will understand
RelaxNG either, but most people find it more readable and writable
than XSD.
Andy
So I think we choice "xsd", for operator who can make netconf tools and
NMS programmers.
-------
Yoshifumi Atarashi
Hideki Okita wrote:
From implementors' viewpoint,
RelaxNG as the schema is not a good choice.
Major development environments such as Java (Apache),
Visual Studio and others has no support for RelaxNG.
Surely RelaxNG make it easy to read "directly" the NETCONF schema.
However, finally, it can decrease the development efficiency.
I do not object to use RelaxNG as the optional means.
However, XSD should remain as the default schema in NETCONF drafts.
Best regards,
Hideki Okita
Hitachi, Ltd.
References:
RelaxNG Homepage:
http://www.oasis-open.org/committees/tc_home.php?wg_abbrev=relax-ng
Apache Xerces2 Java Parser Homepage:
http://xerces.apache.org/xerces2-j/
Simon Leinen wrote:
On the ad-hoc meeting on data modeling issues, the possibility of
using RelaxNG for data modeling work was raised, and some people asked
for examples. Here are some that are somehow related to network
management.
The Compact RelaxNG Schema for NETCONF's protocol operations (at the
time) that Rob Enns posted to the list in November 2004:
http://ops.ietf.org/lists/netconf/netconf.2004/msg00597.html
The (Expired) draft-schoenw-nmrg-snmp-measure-01.txt contains a
RelaxNG schema for a set of information from SNMP packet traces:
http://bgp.potaroo.net/ietf/all-ids/draft-schoenw-nrmg-snmp-measure-01.txt
--
to unsubscribe send a message to netconf-request@ops.ietf.org with
the word 'unsubscribe' in a single line as the message text body.
archive: <http://ops.ietf.org/lists/netconf/>
--
to unsubscribe send a message to netconf-request@ops.ietf.org with
the word 'unsubscribe' in a single line as the message text body.
archive: <http://ops.ietf.org/lists/netconf/>