Re: [xsl] question on standards and xml/json

Subject: Re: [xsl] question on standards and xml/json
From: "Wendell Piez wapiez@xxxxxxxxxxxxxxx" <xsl-list-service@xxxxxxxxxxxxxxxxxxxxxx>
Date: Sat, 13 Jul 2019 21:32:38 -0000
Brian,

I get you.

So in general, there are two ways, it seems to me, to deal with the
impedance mismatch. The first is to work around it with syntactic
elaboration. As you note, this makes ugly results in the target
language. The other is to constrain one's usage in the host language
such that it maps cleanly into the target language. This makes
prettier results, but of course only a subset of the source language
can be used.

The JSON-friendly XML subset you have in mind would enable the second
approach with minimal or no resort to the methods of the first
approach. Of course, there is no single such subset of XML. In the
project I linked to we have one we are working on, but it is not the
only one possible and (depending on your needs and what makes you mad)
possibly not the best.

To get closer to your question, however, it bears mentioning that we
are not thinking about the XML subset (that we are supporting)
explicitly, but rather backing into its definition, by implementing a
schema syntax to produce "equivalent" JSON Schema and XSD together for
"equivalent" data, subject to some definition of equivalence. (This
work was more or less impossible before JSON Schema was available,
which didn't stop people demonstrating that such a relation is
possible in principle. David Lee, G. Ken Holman, H J Rennau, Jonathan
Robie have all done work on this IINM.) Of course it's in this
definition of "equivalent" here that your question lurks. But it's
much easier when it's a design problem than when it's a theoretical
question.

Cheers, Wendell


On Sat, Jul 13, 2019 at 1:36 PM BR Chrisman brchrisman@xxxxxxxxx
<xsl-list-service@xxxxxxxxxxxxxxxxxxxxxx> wrote:
>
> Thanks Wendell,
>
> I think this is kind of similar.
> Basically, I work with a number of people/groups that are pretty much JSON-only.
> I commonly find the need to take some sort of XML that we've been
> using and provide them with an extract of that data in the JSON
> format.
> While it's not that difficult to convert XML into JSON with XSLT, I've
> been wanting to have some of my structures more 'ready to convert'
> already.
> As Liam mentioned, the reverse can be problematic because JSON may
> include XML-excluded characters, but my interests right now are mostly
> around "other people need data (currently in XML) that I'm working on,
> and I need to make it easier on me to get it to them.
> I've evaluated some xml->json conversion stylesheets, but I would
> guess some standard like: "If you wind up with *only* this subset of
> XML functionality, your XML will always be reasonably and directly
> accessible after conversion to JSON." would help me a lot.
> Things like:
> - no multiple-namespaces (ie, I can have namespaces in my XML, but
> they cannot be reasonably expressed in JSON, so anything in any other
> namespace will only exist for XML processing tools)
> - containers - can I mark up items to explicitly be containers such
> that they translate well to JSON arrays?  This frustrates me because
> XML doesn't require objects of ostensibly the same type (element name)
> to be stored in an array, but JSON does (using language loosely here,
> hope the meaning is pretty clear)
> - attributes vs text nodes... lots of ways to go there.. but perhaps
> choosing only one of those is the best option?
> - (maybe other stuff)
>
> In essence, I know there's certainly a way to express *any* XML in
> JSON, so long as we're willing to go far enough in producing ugly
> JSON.... but I've been looking for a standard such that if I express
> my XML in 'just such a way', it will provide JSON that ... doesn't
> make people angry... :)
>
> - Brian
>
> On Mon, Jun 24, 2019 at 2:25 PM Wendell Piez wapiez@xxxxxxxxxxxxxxx
> <xsl-list-service@xxxxxxxxxxxxxxxxxxxxxx> wrote:
> >
> > Good readers,
> >
> > Gerrit points out to me the link was broken. Which is odd, but here's
> > another one:
> >
> > https://pages.nist.gov/OSCAL/
> >
> > Apologies! -- Wendell
> >
> > On Mon, Jun 24, 2019 at 5:16 PM Wendell Piez wapiez@xxxxxxxxxxxxxxx
> > <xsl-list-service@xxxxxxxxxxxxxxxxxxxxxx> wrote:
> > >
> > > Hi,
> > >
> > > XML and JSON alignment is one of the problems we've been working on here
> > >
> > > www.usnist.gov/OSCAL/
> > >
> > > Warning: under development and subject to change. But lots of XSLT in there!
> > >
> > > Cheers, Wendell
> > >
> > >
> > > On Sat, Jun 22, 2019 at 11:07 PM BR Chrisman brchrisman@xxxxxxxxx
> > > <xsl-list-service@xxxxxxxxxxxxxxxxxxxxxx> wrote:
> > > >
> > > > Has there been any work to define an XML subset which is
> > > > simply/directly transformable into JSON?
> > > > I know there are many XML expressions that  are very difficult to
> > > > convert to JSON due to limits in JSON that make such a conversion very
> > > > messy, but I would guess that with a number of standardized
> > > > restrictions, that might be easier?
> > > > Just wondering... I'm seeing more and more situations where I need
> > > > that kind of 'more direct compatibility'.
> > > >
> > > > - Brian
> > > >
> > >
> > >
> > >
> > > --
> > > ...Wendell Piez... ...wendell -at- nist -dot- gov...
> > > ...wendellpiez.com... ...pellucidliterature.org... ...pausepress.org...
> > > ...github.com/wendellpiez... ...gitlab.coko.foundation/wendell...
> > >
> >
> >
> >
> > --
> > ...Wendell Piez... ...wendell -at- nist -dot- gov...
> > ...wendellpiez.com... ...pellucidliterature.org... ...pausepress.org...
> > ...github.com/wendellpiez... ...gitlab.coko.foundation/wendell...
> >
> 



-- 
...Wendell Piez... ...wendell -at- nist -dot- gov...
...wendellpiez.com... ...pellucidliterature.org... ...pausepress.org...
...github.com/wendellpiez... ...gitlab.coko.foundation/wendell...

Current Thread