Re[2]: [cgmo-webcgm] Issue: linkuri (chap 3 & chap 5) different

[Lofton Henderson <lofton@rockynet.com>]

At 12:23 PM 4/27/2005 -0400, Benoit Bezaire wrote:
> Tuesday, April 26, 2005, 8:40:12
> PM, Lofton wrote:
> [...]
> LH> It says:"Parameters. The data record is a SDR of one
> member,
> LH> containingthree strings (type SF, String Fixed). The first
> string
> LH> is the link destination, a URI, the second string (possibly
> null)
> LH> isa Link Title parameter, and the third string is (possibly
> null)
> LH> the Behavior parameter. Note that a null string is a
> zero-length
> LH> string, and is not the same as an omitted
> parameter."
>
> LH> I have always thought this to mean that all three parts must
> LH> be present, even if part 2 and/or part 3 are empty strings. 
> Do
> LH> you read it otherwise?
> Again, if a script writer ends up reading this definition instead 
> of
> the one in chapter 5... he may get confused. 

Then we have to be clear in both places that each (CGM instance/SDR, and
DOM/delimited-string) has its own technique.

> It's not that clear how
> you map the above wording to:
> obj.setAppStructureAttr( "linkuri", ??what goes here??
> );

You don't!  I.e., we explicitly chose not to use CGM SDR coding for
things like list-of-linkuri, or list-of-name.  We chose to use
"delimited string".  Therefore ... again, each place
should clearly explain its method and not tempt the user to "...map
the wording..."

Which I think we actually agree to...

> [...]
> LH> That would probably help.  We should clearly distinguish
> that
> LH> the delimited-string DOM encoding in section 5.5 and 5.7.6)
> is
> LH> distinct from the SDR encoding of the WebCGM instance itself
> (SDR,
> LH> as described in eachof 3.2.2.1 thru 3.2.2.10).
> Yes.
>
> LH> Btw, in 5.7.6 the example for linkuri is:
> LH>
> "http://w3.org"
> "W3C"  "_blank"
> LH> Shouldn't it be:
> LH>
> '"http://w3.org"
> "W3C"  "_blank"'
> LH> I.e., enclosed in APOSTROPHEs?
> Yes and no. Apostrophes are not in section 5.5

Oops, you're right.  Either apostrophes or quotes can be used to
enclose individual list items, but I was inferring one too many in the
grammar (agreed, there are no all-enclosing quotes or apostrophes in the
grammar).

> [...]
> obj.setAppStructureAttr( "visibility", "off"
> );
>
> But the quote are required to tell the ECMAScript engine that this
> is
> a string; the actual string value is: \o \f \f \0

Right, agreed now.  The linkuri example could be written:
obj.setAppStructureAttr( "linkuri",
"'http://w3.org'
'W3C'  '_blank'" );

The value of the argument does not include the outer
quotes.

> Yes, the table in insconsistent,
> but it's not clear if it is showing a
> set string or a get string.

To be consistent, should it have quotes for the interactivity example,
"on"?  (I.e., if we exclude the all-enclosing quotes for
delimited-string examples, because they are not part of the value, should
we also exclude the quotes on single strings, because they are not part
of the value?)

-Lofton.