Re: [Json] JSON Schema Language

[John Cowan <[e-mail]>]

On Tue, May 7, 2019 at 5:30 PM Phillip Hallam-Baker <[e-mail]>
wrote:

1) The lexical analyzer may distinguish integers and floats and this may
> cause a schema driven parser to reject legitimate number because it is
> encoded as a double. (e.g. 1.0)
>
> 2) There may be precision loss because the number is presented in
> exponential form (1.0E20)
>
> 3) There may be precision loss because the integer is larger than the
> floating point mantissa can hold.
>

JSON encodes absolutely no information about whether its numbers are to be
interpreted as exact or inexact.  1.23 may be an exact number equivalent to
123 divided by 100, and 1.0E20 may by the same token be the exact integer
100,000,000,000,000,000,000.   Per contra, 12 could be an inexact number
that happens to have been rounded to the nearest integer.  We are so used
to the idea that dots and Es *mean* inexact, and their absence *means*
exact, and that all our numbers are *inherently* bounded in range and (if
non-integral) precision, that we think these things must be requirements of
our data notation.  But they aren't.

I agree.. And I think most others here agree. The problem is that when we
> go down the 'schema' road, we tend to end up with schema languages that
> become baroque and more hassle than they are worth.


The conclusion I drew from the XML Schema debacle is that schema languages
get wrong-footed when they try to do two things at once: validating purely
syntactically (as Relax NG does) and validating for data binding,
especially to rigid static languages.

Pretty much every programming language out there today will cause all
> calculations to be performed in Intel 80 bit floating point until the data
> has to be stored in a program data value when the mantissa will be
> truncated at 52 bits and stored in a double.


Practically every programming language can handle arbitrary size, aribtrary
precision exact numbers, either as built-ins or through a library.


John Cowan          http://vrici.lojban.org/~cowan        [e-mail]
De plichten van een docent zijn divers, die van het gehoor ook.
      --Edsger Dijkstra