[x3d-public] Agenda for X3D WG Open Meeting Wednesday 10th May 2017

[Yves Piguet]

On 10 May 2017, at 22:57, Roy Walmsley <roy.walmsley at ntlworld.com> wrote:
> An empty MFString array would have no values. Now that is an interesting one
> we didn't discuss today!!!
> 
> According to the current standard it would be simply [''], i.e. two single
> quotes or apostrophes. However, if we accept the changes discussed today,
> one of which was that an MFString field with only one value could dispense
> with the outer set of quotes and be encoded as an SFString, then that could
> present a dilemma. It would mean that it could be interpreted as a single
> value, which was the empty string, or an empty MFString array. These are not
> the same things, and so this case would require specific wording in the
> text, for clarity.

"this case would require specific wording in the text, for clarity": how? a special case? Is it really worth it wrt the current MFString encoding which isn't that bad (and was likely chosen precisely because of that 15 years ago)? Btw, rationale would be useful to keep a trace of decisions available to everybody.

> On 10 May 2017, at 23:03, Don Brutzman <brutzman at nps.edu> wrote:
> 
> On 5/10/2017 1:40 PM, Yves Piguet wrote:
>> Excuse me if I've missed it, but how do you specify an empty MFString array? If double quotes are optional, '' (empty attribute value) would mean an empty SFString, i.e. [""] if an MFString is expected, not [].
> 
> Excellent question.  Two answers for completeness.
> 
> a.  For empty MFString array, use default value

What if the default value is not empty?

> or an empty attribute.
> 
> Empty array value for ClassicVRML:  [ ]
> 
> 	<Text/>
> 
> 	<Text string=''/>
> 
> 	<Text string=""/>

'' and "" can also be the unquoted version of a single empty string, i.e. [""], unless you forbid explicitly that case.

Yves