Re: [cgmo-webcgm] ISSUE: remaining Delimited String sub-issues

[Lofton Henderson <lofton@rockynet.com>]

I'm fine with either solution.  Easy on implementors seems good here.  The 
case we would prohibit for 2.0 -- user data containing both quote and 
apostrophe in the data -- is marginal.  I think only 'title' of linkuri 
would be affected.

Proposed Resolution:  Alt.2, disallow that a single instance of user data 
(content of single substring) contains both QUOTE and APOSTROPHE.

Objections?

(Dave, add to resolutions, etc.)

-Lofton.

At 02:59 PM 5/25/2005 -0400, Benoit Bezaire wrote:
>Hi Lofton,
>
>Regarding the remaining open issue...
>
>I don't like my suggestion very much, but I don't like the alternative
>either... I'm tempted by Alternative 2:  no, define a rule, "Can't
>have both QUOTE and APOSTROPHE in your data."
>
>It's less work for implementors (but limiting for users). We could
>always introduce the escaping concept in a 2.1 update (if users ask
>for it).
>
>--
>  Benoit   mailto:benoit@itedo.com
>
>
>Wednesday, May 25, 2005, 12:24:18 PM, Lofton wrote:
>
>LH> WebCGM TC --
>
>LH> #3 is still open (#2 is pending LH to draft solution and put it in 
>text for
>LH> review)...
>
>LH> At 07:21 AM 5/25/2005 -0600, Lofton Henderson wrote:
>
> >>>About the "delimited string inconsistency" item:  "It was decided that
> >>>whitespace delimited will be used  for viewcontext (simple string), but
> >>>for region we will use single quote delimited sub strings as defined in
> >>>5.5  linkuri for multiple links or multiple regions. Do we need a special
> >>>case for single strings?"
> >>
> >>(To be clear, "whitespace-delimited" refers specifically to the 'wsp'
> >>production of 5.5.)  Correction:  for the Delimited String data type, the
> >>definition is as in 5.5:  multiple substrings are either quote-delimited
> >>and wsp separated, or apostrophe-delimited and wsp separated.
> >>
> >>Open question 1:  does single substring revert to Simple String
> >>encoding?  (Recommendation:  yes)
>
>LH> 2005-05-25:  Resolved, YES.
>
>
> >>Open question 2:  what is the definition of
> >>validChar?  (Recommendation:  Refer to 3.1.1.3 -- the definition will vary
> >>depending on whether it is a 'name', a 'linkuri', a 'title', etc)
>
>LH> 2005-05-25:  Lofton will make specific proposal (possibly in next Editor's
>LH> version text).
>
>
> >>Open question 3:  is problem-char escaping needed within the content of
> >>substrings?  (Recommendation:   yes.  Use "\", recognizing that JS will
> >>also consume some of the "\" characters.)
>
>LH> OPEN.  Here is an description of the problem, if we don't have some
>LH> escaping mechanism (ignore 1st part about "Valid...", and look at 2nd 
>part,
>LH> about "Escaping..."):
>LH> 
>http://www.oasis-open.org/apps/org/workgroup/cgmo-webcgm/email/archives/200505/msg00046.html
>
>LH> To summarize concisely, here are two alternatives and a recommendation.
>
>LH> Alternative 1:  yes, define escaping mechanism ("\").
>LH> Alternative 2:  no, define a rule, "Can't have both QUOTE and 
>APOSTROPHE in
>LH> your data."
>
>LH> Recommendation:  Alt.1.
>
>LH> Other views (esp. implementors)?
>
>LH> Regards,
>LH> -Lofton.