Hi Gary, hi David,

thanks for the pointers! At least it's reassuring to hear I haven't
missed something obvious.

@Gary: I think I need to be a bit careful about not getting involved in
too many topics at once, so I'll pass for now. I'll be happy to chime
in though, and maybe I will have more capacity at some point.
I am also certainly not an expert in xml - what I know is derived
mostly from having to deal with with Maven pom.xml's...

@David: Thanks for the link! I really like the idea of having one
abstract, format-agnostic schema that can then be converted to the
concrete formats. I assume tag/value and rdf are not in scope? Since
those are (to my knowledge) quite different from json/yaml/xml.
I don't have much input for the xml schema so far; the only thing I
noticed are the list representations mentioned in the previous mail.

Regards,
Nico

On Mon, 2022-10-31 at 16:25 -0400, David Kemp wrote:
> Hi Nico,
> 
> I've been working on abstract schemas for SPDX v2 and v3.  The
> rationale for using an abstract schema is that it can mechanically
> generate concrete schemas for multiple data formats, including JSON,
> concise (machine-optimized) JSON, CBOR, and XML.  I don't currently
> have XML encoding rules built into the current tool but they should
> be fairly easy to create, given examples of the desired output style.
> 
> The benefit of using an abstract schema is that encoding rules only
> need to be developed once and then can be applied to all information
> models, allowing updates from v2.2 to v2.3 and v3.0 without any
> serialization-specific work.
> 
> Have a look at 
> https://github.com/davaya/spdxv3-template-tool/blob/main/Schemas/spdx-v2_2.jidl
> to get an idea of the abstract structure - it has a top-level
> Document type, and a packages property containing multiple
> PackageInfo elements.  Note that JSON does not have visible types, so
> type names in the schema are fairly arbitrary.  In XML, element types
> are visible, so adjusting the type names is possible without
> affecting JSON data.
> 
> I'm interested in applying the information model to both JSON and XML
> data, since that is its purpose.  If you have some napkin sketch
> examples of what your preferred XML might look like, that would guide
> development of encoding rules to generate them while preserving JSON
> conformance.
> 
> Regards,
> David
> 
> On Mon, Oct 31, 2022 at 2:05 PM Gary O'Neall <[email protected]>
> wrote:
> > Hi Nico,
> > 
> >  
> > 
> > I’m glad you asked – I put out a very preliminary XML schema, but I
> > don’t feel qualified to create even a draft of an authoritative
> > schema.
> > 
> >  
> > 
> > We could use some help here.
> > 
> >  
> > 
> > Here’s an issue tracking this in the SPDX Spec: 
> > https://github.com/spdx/spdx-spec/issues/615
> > 
> >  
> > 
> > Please feel free to pick this up.
> > 
> >  
> > 
> > I would like to generate the XML schema in a similar fashion to the
> > JSON schema – I can definitely help with the Java code to generate
> > once the schema itself is determined – but feel free to contribute
> > to that as well.
> > 
> >  
> > 
> > Here’s a start at an XSD generator utility: 
> > https://github.com/spdx/tools-java/blob/master/src/main/java/org/spdx/tools/schema/OwlToXsd.java
> > .  Of course it will need to be updated once we know what the XSD
> > is to look like.
> > 
> >  
> > 
> > Gary
> > 
> >  
> > 
> >  
> > 
> >  
> > 
> > From: [email protected] <[email protected]> On Behalf
> > Of Nicolaus Weidner via lists.spdx.org
> > Sent: Monday, October 31, 2022 7:16 AM
> > To: [email protected]
> > Subject: [spdx-tech] Official spdx xml schema?
> > 
> >  
> > 
> > Hi all,
> > 
> > I was wondering whether there is any official (or semi-official)
> > schema for the xml file format. My current assumption is that it
> > follows the json schema wherever possible.
> > I have to say "wherever possible" because some differences cannot
> > be avoided. I am aware of the following:
> > 
> > There needs to be a root element, so there is a top-level
> > <Document> element containing all properties
> > List representations look weird to me: For example, one can
> > currently find multiple successive <packages>...</packages>
> > elements in one document (e.g. 
> > https://github.com/spdx/tools-java/blob/92e1c5d29eacb0081139b0c05e5e6270b231788c/testResources/SPDXXMLExample-v2.3.spdx.xml#L254-L276
> > ). I would prefer single <package> elements wrapped in an enclosing
> > <packages> element at this point. Alternatively, at least each
> > single element should be named using singular even if there is no
> > enclosing element.
> > 
> > So, back to the original question: Is there any official definition
> > for the xml schema? Or does one just look at the json schema and if
> > in doubt, look at some examples (which seems dangerous to me,
> > because examples can contain errors...)?
> > 
> > Best,
> > Nico
> > 
> > 
-- 
Dr. Nicolaus Weidner * [email protected]
TNG Technology Consulting GmbH, Beta-Str. 13a, 85774 Unterföhring
Geschäftsführer: Henrik Klagges, Dr. Robert Dahlke, Thomas Endres
Aufsichtsratsvorsitzender: Christoph Stock
Sitz: Unterföhring * Amtsgericht München * HRB 135082



-=-=-=-=-=-=-=-=-=-=-=-
Links: You receive all messages sent to this group.
View/Reply Online (#4823): https://lists.spdx.org/g/Spdx-tech/message/4823
Mute This Topic: https://lists.spdx.org/mt/94686073/21656
Group Owner: [email protected]
Unsubscribe: https://lists.spdx.org/g/Spdx-tech/unsub [[email protected]]
-=-=-=-=-=-=-=-=-=-=-=-


Reply via email to