Re: <g x="10" y="20"> was: does the systemLanguage attribute apply to tspan? alternatively does the switch element take x & y coordinates?

Doug,

thanks for your explanation**

perhaps put another way and in a more general sense,

Why is it that x and y are not css style attributes as in:
<style type="text/css" id="css"><![CDATA[
text.tt { x: 130; y: 65}
]]></style>
...
<text class="tt">games too</text>

in html positioning is possible using css....
whether absolute or relative.

I do wonder about my thought processes, when I get stuck, like this.
Am I completely missing the point? did I forget something obvious?
I simply cannot remember and have searched and tried without success.

cheers

Jonathan Chetwynd

**probably I am misusing the term inherit.

however it's not clear to me why
<g transform="translate(130,65)">
is preferable to
<g x="130" y="65">

the former is relative, the latter absolute, but there are use cases  
for each.
the latter may be conceptually easier to understand




On 14 Mar 2007, at 14:04, Doug Schepers wrote:

Hi, Jonathan-

~:'' ありがとうございました。 wrote:
 >
 > <g x="10" y="20">
 >
 > What is the rationale or reason for the WG decision that x and y
 > attributes unlike many others, are not attributable to a group, and
 > therefore cannot be inherited**
 > in my example I specifically limited myself to these two attributes.


I wasn't there at the time, but I may be able to answer this (as me,  
not for the SVG WG).

First, x & y never inherit.  They are geometric properties of a shape.

Second, a group does not have intrinsic geometry, so it does not have  
an x or y.

Third, the idea was to allow a larger range of consistent and  
additive transformations of the coordinate space (thus changing the  
appearance of the geometry).

I think it was in the same spirit as some aspects of CSS (akin to  
presentation vs. content).  I admit that it's not always as  
convenient as I'd like, since for drag operations and such, you have  
to parse them out of a string, or use the CTM.  But neither do I  
think it particularly hampers the author, as 'translate(x,y)' is  
effectively the same thing for most purposes as x & y attributes  
(except that it does inherit, which is the desired behavior).

Am I missing your use case?


Regards-
-Doug

Research and Standards Engineer
6th Sense Analytics
www.6thsenseanalytics.com
mobile: 919.824.5482

Received on Wednesday, 14 March 2007 16:25:50 UTC