Re: [xml-dev] Is the set of languages expressible using XML asuperset of the set of languages expressible using JSON?
> And XML isn't just strings, btw. For what is a "string"?
A text string is sequence of (Unicode repertoire) characters whose only common property is that they have some impact on glyph selection or spacing.
The only proper semantics of a string in standalone XML data content or attribute values is its influence on rendering and collation (sorting, matching). Anything else involves handwaving, metonymy, and passing off other layers (which certainly will exist) as "XML".
Standalone XML provides no way to express values: no datatypes not even boolean or numbers. XML provides no way to express facts: no surity level of the sources, no reliability of the sources, not even booleans.
XML is a just a notation for organizing Unicode strings in ways that may be useful. (Useful to other layers which may map strings to datatypes, or treat values as facts.)
And we should not be surprised if people who want a notation to directly represent values adopt JSON, which has enough delimiters and rules to represent boring values without fuss*
Cheers
Rick
* My Rapid Access Notation thought bubble goes beyond XML Schemas and JSON in the data values that can be expressed directly, such as many multi-value logic and ISO8601's date ranges, uncertainty and date wildcards.
https://schematron.com/document/2957.html
Rick Jelliffe <rjelliffe@allette.com.au> wrote:
> I think XML is not a notation for transferring facts.
If you peel away enough semantic crud, you're always stuck with
something that can't do what you want it to do. And by any token, no
technology will do well in transferring "facts" as we can't even
agree on what "facts" are, or what it means to "transfer" them.
But I don't agree that XML is this simplistic thing that can't do what
we want it to do. Of course it can. Just not currently in any agreeable
form. Whether something is encoded in string or sound or small cuts on
a leaf blown by the winds over the ocean, Wittgenstein's ghost will
always make sure these kinds of dishcussions go on forever. That's job
security right there.
And XML isn't just strings, btw. For what is a "string"? There is
semantic goodness even in the bare bones of the cursed child of SGML,
any notation carries significance even if the interpreters can't agree
on it. Getting back to the subject a bit, XML has a much higher semantic
payload than JSON. I think that lies at the crux of this very
discussion, both pro's and con's, and that semantic payload can very well
carry higher level semantics that carries higher level semantics, it's
just that the cake is so high and the interpreters so many and the
epistemological conundrums so complex. That's why we most often
return to data and numbers and some comfortable definition of data,
wrapped in as little as possible to see if we can accomplish ...
"something", and most often something very practical "something".
This is why XML and JSON both are successful. And maybe why JSON
is becoming the next "something" is because of the lack of burden, the
lack of control, the lack of fiddly human meaning, and so let's just focus
in of data. Plain, boring data. Because the other stuff is really hard.
Cheers,
Alex
--
Information Alchemist, tone modulator, swords master
thinkplot.org | linkedin.com/in/shelterit | sheltered-objections.blogspot.