[
Lists Home |
Date Index |
Thread Index
]
My position is/was that XSD is tightly integrated into
so many roles in VS that replacement by RelaxNG won't happen
at MS. It is the utility value, not the cost value, that drives
a network effect. The connectivity value of the network is
the solution space of VS multiplied by the desktop presence.
I don't expect MS to add RelaxNG support.
So we are in agreement. On the other hand, if it is
trivial to do and trivial to maintain, it can be worth
it. The learning curve for XSD is steep. The learning
curve for RelaxNG is not so steep. As John Cowan pointed
out, there are some validation chores Relax can do that
XSD can't. So the utility value of each compared to
the other without considering the VS environment is
useful to know when deciding which to apply to a given
application language design task or when creating
a reference design that is vendor-independent.
One can inquire of groups such as the SVG WG, why they chose
to use RELAXNG. That is, of the things XSD can do and
not do, of the things that RelaxNG can do and not do, what
utilities are driving the adoption of either beyond vendor
lock in?
I expect those who have done that diligence and found RelaxNG
the better solution for their design and who work in the
VS/MS environment to look to third parties for the support.
It's a niche seeking its own externality to drive growth.
len
From: Dare Obasanjo [mailto:dareo@microsoft.com]
> Best use isn't enough UNLESS there is repeat use of the tool.
> The market for schema tools to repeat users is what size if
> you start to split that up into 'best used for'? In other
> words, is a marginal utility market big enough? Or should
> the solution be a niche vendor who supplies support
> into the .Net framework? Can that be done and still get
> all the nice productivity features of VS?
Adding RELAX NG support to the .NET Framework wouldn't be that hard. In
fact you could do that trivially by porting Commons.Xml.Relaxng from
Mono to .NET. However when I think of XSD support at Microsoft I don't
just consider validating XML parsers. There's Object<->XML mapping with
the XML Serializer which is an integral part of the XML Web Services
story, Relational<->XML mapping with the ADO.NET DataSet and separately
in SQLXML, there's BizTalk's support for XSD, there's Office whose
entire XML support is based around XSD, and then there's actually
having the Visual Studio IDE understand how to work with RELAX NG.
So given all these ways XSD is used in various Microsoft products is
your position that all you'd expect from Microsoft if it supported RELAX
NG is adding a validating parser to the .NET framework and some syntax
highlighting + directed editing in Visual Studio?
That sounds highly unlikely.
|