RE: Re[2]: [cgmo-webcgm] Re: The highlight() method

["Galt, Stuart A" <stuart.a.galt@boeing.com>]

Benoit,

First I apologize for not using a mail program sophisticated enough to
reply inline... Please forgive my cut/paste errors.

>> I think that the ability to apply a highlight to a single item is an 
>> important feature to have.
>Ok. Enough to hold the spec a bit longer and make more modifications to
fix
> the problem? (just asking)

I am not sure.  I think that all issues are being deferred until 
we address all of the issues found in the last call process.

>> So my thoughts turned to something along the lines of option 4. Can I

>> either create a list on the fly or can I dereference an existing
list?
>Dereferencing the list I don't think is technically possible. You would

>instead need a 'for' loop and call highlight with each item, but them 
>you run into timing issues. Create a list on the fly makes more sense 
>in my opinion.

From what I have seen in javascript you can have a variable number of
arguments in a function and get at them via the arguments method.  My
understanding of ecmascript 262 is fuzzy, but I cannot find anything
in my version of the spec about the arguments method.  But I have seen
javascript functions that use the arguments property as an array of the
parameters.  But like at that point you would need to iterate through
them (just like you would with a nodeList).  I think the better of the
two would be to create the list on the fly.  As a side note are there
any times when we would want to turn a node into a nodeList or will
we only see it associated with setting highlights?

>What do you think of the option of having two highlight method (ex:
>highlightOne (takes a single APS), highlightMany (takes a list of APS)?

I guess my personal preference is to have a single interface if
possible.
But it is a personal preference and if people feel strongly the other
way
I can learn to make two different calls.

Summary:
- I think we need to do something that will be included when we process
last call
- I personally prefer makeNodeList() more than
highlight()/highlightMany()
  but if the group feels the other way I can live with that too.

Stuart.
--
Stuart Galt
SGML Resource Group
stuart.a.galt@boeing.com
(206) 544-3656  <--- new number! 


-----Original Message-----
From: Benoit Bezaire [mailto:benoit@itedo.com] 
Sent: Friday, December 09, 2005 12:40 PM
To: cgmo-webcgm@lists.oasis-open.org
Subject: Re[2]: [cgmo-webcgm] Re: The highlight() method

Hi Stuart,

Please see inline...

Friday, December 9, 2005, 1:20:23 PM, you wrote:

> Hello,

> I think that the ability to apply a highlight to a single item is an 
> important feature to have.
Ok. Enough to hold the spec a bit longer and make more modifications to
fix the problem? (just asking)

> My first thought was to overload the argument list, but that is not 
> allowed by ecmascript (or javascript).
Same here, but as you said, that's not an option (too bad).

> So my thoughts turned to something along the lines of option 4. Can I 
> either create a list on the fly or can I dereference an existing list?
Dereferencing the list I don't think is technically possible. You would
instead need a 'for' loop and call highlight with each item, but them
you run into timing issues. Create a list on the fly makes more sense in
my opinion.

> creation on the fly would require something like:
> var aps = pic.getAppStructureById("apsToHighlight");
> pic.highlight(aps.makeNodeList(), true);
Exactly.

> and dereferencing the list would look something like:
> var list = pic.getAppStructuresByName("listToHighlight");
> pic.highlight( list.items(), true );
As I said above, I don't think it's doable without a 'for' loop.

> I think that the dereference option would work because you can have a 
> variable number of arguments in a function.
No, see above (or at least, I don't know how to do that).

> Either one of these would require changes to support.
That's correct.

> I think that the first has the advantage that all the existing scripts

> would continue to work...
True.

> I am not sure that I understand the problem of the live node view of 
> the DOM.
Forget I said that for now... that's an implementation hurdle.
Although they are important, let's look at it from a user perspective
for now.

> A kludge workaround would be to be get the object by id and then get a

> list of objects that share the same name.  You then iterate through 
> the list removing items that do not have the id that you want... (Ugly

> but can be done)
:-) Ugly as you said. That's what users would have to do if we leave the
spec like it is.

> I cannot speak to which version of the spec highlighting should be 
> fixed.
What do you think of the option of having two highlight method (ex:
highlightOne (takes a single APS), highlightMany (takes a list of APS)?

--
Regards,
 Benoit   mailto:benoit@itedo.com


> --
> Stuart Galt
> SGML Resource Group
> stuart.a.galt@boeing.com
> (206) 544-3656  <--- new number! 


> -----Original Message-----
> From: Benoit Bezaire [mailto:benoit@itedo.com] 
> Sent: Friday, December 09, 2005 7:42 AM
> To: cgmo-webcgm@lists.oasis-open.org
> Subject: [cgmo-webcgm] Re: The highlight() method

> I haven't heard from anyone regarding this one either? Your comments
> would be appreciated.