W3C home > Mailing lists > Public > www-svg@w3.org > March 2000

RE: What is wrong with SVG?

From: Don Park <donpark@docuverse.com>
Date: Wed, 8 Mar 2000 19:59:36 -0800
To: <martind@netfolder.com>, "'Jon Ferraiolo'" <jferraio@adobe.com>, "'Elliotte Rusty Harold'" <elharo@metalab.unc.edu>
Cc: <xml-dev@xml.org>, <www-svg@w3.org>
Message-ID: <001601bf897b$e36e42a0$d1940e18@c1033339-a.smateo1.sfba.home.com>
>I tried to convert the paths (used quite often in the figures) 
>into elements and I got an explosion of elements and a huge
>DOM. But with the actual format the number of nodes seems
>reasonable.

The node explosion you encountered is actually quality of
DOM implementation issue.  Most DOM implementations build
nodes as XML data is parsed, not when they are accessed.
SVG DOM's custom path data access interfaces has similar
effects.

One difference I notice is that SVG's non-node objects are
not part of the document tree.  Changing x and y values of
a path's SVGPoint will not affect the path.  This approach
prevents node explosion at the cost of having to instantiate
fresh copy of value for each access.  However, read-only
version of SVG DOM should have this problem.

Considering all this, I think that the decision made by SVG
WG not to use elements for path segments is close to the
80/20 point.  I still do not agree with the decision to use
attribute.

Best,

Don Park    -   mailto:donpark@docuverse.com
Docuverse   -   http://www.docuverse.com
Received on Wednesday, 8 March 2000 22:57:45 GMT

This archive was generated by hypermail 2.3.1 : Friday, 8 March 2013 15:54:18 GMT