Edward C. Zimmermann writes:
> Quoting Mike Taylor <[log in to unmask]>:
>
>
> > So I think the request parameter should be:
> > x-info-99-metadata=true
> > or
> > x-info-99-metadata=TRUE
> > or
> > x-info-99-metadata=t
> > or
> > x-info-99-metadata=yes
> > or
> > x-info-99-metadata=1
> > or something similar. I am guessing that there must by now be a
> > convention for representing "true" as a URI query parameter value --
> > does anyone know where it is (and, better still, where it's
> > documented?)
>
> To my knowledge there is (and really can't be but that's another issue) a
> "standard". A parameter can even just be named without even a value.
> Even in XML where we need an assignment (value) we can have #IMPLIED so
> just naming it might be "implied" to mean its true (with the state of not
> naming it being false). Now while we might have a syntactical model for
> x-info-99-metadata='yes' or x-info-99-metadata='oui' or x-info-99-metadata='ja'
> or x-info-99-metadata='1" etc. we need to assign, in our models, some meaning
> that that's not always clear. Is x-info-99-metadata purely boolean? Is "1"
> the same as "Oui" or "True"? What about "Jawohl"? "Yup"? Thus, I'd argue, we'd
> be best off with the implied case! Its makes the value clear and immediately
> also implies the states :-)
We already discussed this and decided that using a parameter with no
value, while aesthetically pleasing, would likely trip up some
toolkits.
So we've decided to use a value to mean "true" and the issue is what
value that should be. Not one of the most earth-shaking issues we've
ever had to decide, I admit :-)
_/|_ ___________________________________________________________________
/o ) \/ Mike Taylor <[log in to unmask]> http://www.miketaylor.org.uk
)_v__/\ "Doesn't anybody live in one place any more?" -- Carole King,
"So Far Away"
|