RE: DR 17-0019 — SML: Order of Elements in schemeClr

[Francis Cave]

Murata-san

 

I have some sympathy with Charlie regarding this DR.

 

You are of course correct that the schema allows all possible permutations of the elements included in EG_ColorTransform, and it is certainly not appropriate for the standard to try to specify the semantics (i.e. visual effect) of every possible permutation, even if this were achievable, which I’m sure it is not. The standard is correct to allow any permutation.

 

However, the prose is somewhat disingenuous in almost completely failing to mention that the elements in EG_ColorTransform can be used in combination. All the examples given in §20.1.2.3, with the exception of the example in §20.1.2.3.16 hueOff, show only a single element being used. Only by a careful reading of the schema is it clear that any permutation is possible. Apart from the example in §20.1.2.3.16, the only reference to the use of multiple transforms is in §20.1.2.3.29 schemeClr, where it states:

 

“… As with all elements which define a color, it is possible to apply a list of color transforms to the base color defined.”

 

There are in fact six ways of defining a base color value, and I have been wondering whether it would be helpful to enumerate these in the introduction to §20.1.2.3. They are specified in §20.1.2.3.13 hslClr, §20.1.2.3.22 prstClr, §20.1.2.3.29 schemeClr, §20.1.2.3.30 scrgbClr, §20.1.2.3.32 srgbClr and §20.1.2.3.33 sysClr. In all cases it is mandatory to specify a color value, although the interpretation of the values in the cases of prstClr and sysClr are application-dependent.

 

On balance, I think that a minimal modification of the current prose is probably best. My proposal would be to add new sentences and a note at the end of the first paragraph in §20.1.2.3:

 

A color is specified by a base color value, which may be modified by one or more transforms. The effect of specifying color using a combination of several transforms is application-dependent. 

 

[Note: The order in which the transforms are specified is likely to be significant in terms of how the color is rendered by the application. end note]

 

Kind regards,

 

Francis

 

 

 

From: eb2mmrt at gmail.com [mailto:eb2mmrt at gmail.com] On Behalf Of MURATA Makoto
Sent: 09 November 2017 00:43
To: SC34 <e-SC34-WG4 at ecma-international.org>
Subject: DR 17-0019 — SML: Order of Elements in schemeClr

 

Dear colleagues,

 

I propose to close this DR without action.

 

This DR points out that an element order is significant 

and then argue that "The schema is quite clear 
that order is not important and, indeed, 
*Off elements precede *Mod elements:"

 

The schema merely allows any order.  It does not 

(and cannot) indicate significance of orders.

 

Regards,

Makoto

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.vse.cz/pipermail/sc34wg4/attachments/20171109/b8c7e502/attachment.html>