[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index] | [List Home]
Subject: RE: [cgmo-webcgm] Problem with 'attributes'
Hi Benoit, see inline >-----Original Message----- >From: Benoit Bezaire [mailto:benoit@itedo.com] >Sent: Tuesday, July 13, 2004 9:32 PM >To: CGM Open WebCGM TC >Subject: [cgmo-webcgm] Problem with 'attributes' > > >Hi, > > I'd love to fix this problem in the draft and would like the group's > opinion. Here's the problem, let us first consider the following > CGM file: > > BEGAPS 'myId' 'grobject' STLIST; > APSATTR 'name' "14 1 'myName'"; > APSATTR 'screentip' "14 1 'This is a tip'"; > APSATTR 'viewcontext' "16 4 126.6082 97.4197 471.8895 318.0822"; > APSATTR 'linkuri' "14 3 'http://www.w3.org' 'W3C' ''"; > BEGAPSBODY; > ... > > and the following XML companion file: > <companion> > <grobject apsId="myId" wire:has-wire="true"> > <wire:data type="22" color="green"/> > </grobject> > </companion> > > So let's assume that the XML companion file is applied, if I > remember correctly, we agreed that the memory model would look like > this after loading and applying the companion file: > > BEGAPS 'myId' 'grobject' STLIST; > APSATTR 'screentip' "14 1 'This is a tip'"; > APSATTR 'viewcontext' "16 4 126.6082 97.4197 471.8895 318.0822"; > APSATTR 'linkuri' "14 3 'http://www.w3.org' 'W3C' ''"; > APSATTR 'name' "14 1 'myName'"; > APSATTR 'wire:has-wire' "14 1 'true'"; > BEGAPSBODY; > BEGAPS 'madeUpId' 'wire:data' STLIST; > APSATTR 'type' "14 1 '22'"; > APSATTR 'color' "14 1 'green'"; > ... This is what I would expect. > > Given the current spec, a user can get access at: > -screentip and viewcontext by calling getAppStructureAttr() on myId; > -linkuri by calling getLinkuris() on myId; > -name by calling getNames() on myId; > -wire:has-wire by calling getAttributeNS() on myId; > > Correct? Yes. > > My question is... what should 'attributes' return when called on > myId? Returning all five attributes (screentip, viewcontext, > linkuri, name, wire:has-wire) as Attr nodes currently can't be done > because a linkuri can't be expressed as a Attr Node; linkuri have > their own interface; Hyperlink which doesn't derive from Node :-( Too bad > > Do people agree that it is unacceptable to return less than all 5 > attributes? Yes. I would like to have a list of all attributes here. > > One possible approach for fixing the problem would be to make the > 'value' attribute of the Attr interface a DOMStringList instead of a > DOMString. That way we could remove the Hyperlink interface and > state that a linkuri has to be expressed as a DOMStringList > containing 3 strings. That seems to be an acceptable way. The only alternative would be to derive the Hyperlink interface from the Node interface, which would be nonsense in my eyes. I prefer the strings. Regards, Dieter > > Thoughts? Or any other suggestions? > > Regards, > >-- > Benoit mailto:benoit@itedo.com > >
[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index] | [List Home]