W3C home > Mailing lists > Public > xmlschema-dev@w3.org > September 2002

Re: Announcment - JBind a Java-XML Databinding Framework

From: Stefan Wachter <Stefan.Wachter@gmx.de>
Date: Fri, 13 Sep 2002 14:37:58 +0200 (MEST)
To: "Eric Jain" <Eric.Jain@isb-sib.ch>; xmlschema-dev@w3.org
Message-ID: <20845.1031920678@www55.gmx.net>

Hi Eric,

About adding behaviour: JBind does not intend to define the behaviour of XML
data inside an XML schema. It mearly allows to add behaviour by the seamless
intergration of manually written Java code with the generated code.

About extensions: A short term goal for JBind is to have all binding
information and extensions outside of the schema document in a separate document.

About unmarshalling large documents that do not fit into memory: Thats not
possible with JBind the reason is that the complete object tree is
instantiated (Maybe you asked that question because the usage of unconstraint XPath in
identity constraints on types?).

About the license: I submitted the JBind license to opensoftware.org for
approval. The license is a mixture of two already approved licenses: the
Artistic License and the Attribution Assurance License.

I hope you will have some fun testing JBind.

--Stefan


> > I would be happy if some of you could have a look on it and give me
> > feedback. I would be especially thankful for comments on the extensions
> > that I added to XML schema.
> 
> The idea of specifying logic in a schema is certainly compelling. On the
> other hand we use schemas not for generating code for internal use but to
> allow other organizations to work with our data without having to write
> all
> the object model and marshalling code themselves, *if they are not also
> using Java*, that is, in which case they can use the more powerful tools
> we
> provide. For any extensions to be of any use, they must therefore by
> widely
> available and standardized. Of course someone always has to make a
> beginning...
> 
> Another issue is that you probably won't be able to specify all behavior
> you
> may want to attach to your data with such extensions, unless the extension
> would in fact contain a full-blow programming language independent
> programming language. For the time being I would rather specify all
> behavior
> in plain English than having a half-half solution.
> 
> One question (before I actually try it out...): does your framework
> (un)marshal large files that don't fit into memory?
> 
> Also, is the licensing term really open-source compatible? Not that I
> object
> to anyone getting paid for their work, but we must distribute our data
> publicly and integrating features (even if optional) that require
> licensing
> in certain cases is problematic at best.
> 
> "4. You may use the Package, a Modified Package, or a non-commercial
> product
> that contains the Package or a Modified Package in a commercial
> environment,
> e.g. a commercial product, a commercial project, or inside a company only
> with the written permission of the Copyright Holder. Normally this
> permission is granted after a modest attribution to the Copyright Holder."
> 
> 
> --
> Eric Jain
> 
Received on Friday, 13 September 2002 08:38:33 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Tuesday, 11 January 2011 00:14:34 GMT