[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index] | [List Home]
Subject: logical flaw with new SPs initial value?
All -- As we have been discussing (on the TC list [0]) issues that Benoit raised about fill-related new SPs, I have noticed what might be a subtle logical flaw in the specification of the new 2.1 SPs. Have a look at the table [1]. Notice the "Initial value" column. Now go look at the inheritance model [2]. [1] http://www.w3.org/TR/2008/WD-webcgm21-20080917/WebCGM21-DOM.html#styleprop-table [2] http://www.w3.org/TR/2008/WD-webcgm21-20080917/WebCGM21-DOM.html#L2666 It appear to me, from [2], that there is always a valued defined for each SP. The value, after application of the inhertance model of [2], comes back to the Initial Value if no explicit DOM calls have been made. Notice the old (2.0) SPs -- their Initial Values are all such that the value of the associated CGM attribute is unaltered if there is no explicit SP setting via DOM calls. That is the intention of "100%" and "metafile", I believe. Now have a look at the new (2.1) SPs. Their initial values are particular values of the associated CGM attribute. Therefore, with no SP-setting DOM calls, there is some implication of an initial value of the SP that could be at odds with the contents of the metafile. The outcome of such a state is somewhat ambiguous. We apparently were careful to avoid that trap with 2.0, by appropriate choice of Initial Value. Am I missing something? Does the 2.1 spec say anywhere that, despite the establishment of definite values for everything with the inheritance model [2], there shall be no action to override attributes in an APS without explicit SP-setting calls? (I can't find anything like that.) Solution? One option would be to mimic the 2.0 Initial Values. For example, the I.V. of interior-style would be "metafile" instead of "1". Thoughts? -Lofton.
[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index] | [List Home]