From graphics-request at octave dot org Mon Feb 27 03:20:47 2006
Subject: Re: Object Graphics v0.2
From: Bill Denney <denney at seas dot upenn dot edu>
To: Shai Ayal <shaiay at gmail dot com>
cc: graphics at octave dot org
Date: Mon, 27 Feb 2006 04:21:17 -0500 (EST)

On Mon, 27 Feb 2006, Shai Ayal wrote:

> In matlab you have the "userdata" property which is a matrix and the 
> "tag" property which is a string. both are left to the user to do with 
> as he pleases.

Tag seems to be what I was thinking of.  Userdata seems to be for UI 
creation (please correct me if I'm wrong), and I'm thinking that at least 
for the first version of this standard UIs are outside the scope of the 
document.

> Well, the RFC should define the minimal set of properties, and any
> backend can add specific properties. This would of course break
> compatibility, so I suggest that these specific properties be limited
> to the root and/or figure objects. I.e. in octplot I have the
> root.fontdir property so I can find the fonts. Since it's in the root,
> and should only be set at startup, it doesn't really break
> compatibility.

OK, that sounds good.  We will just need to be sure that when we're done 
with this that when the coding is done there are general tests done by 
octave for the minimal required types and it also allows the backend to do 
tests for their optinal types.  One thing that I've been thinking about 
that is along these lines is if we're going to store color values in rgba, 
is it just up to the backend to strip the alpha channel and warn if it's 
~= 1 if the backend doesn't support alpha?

Bill

-- 
"I am not one of those who believe that a great army is the means of
maintaining peace, because if you build up a great profession those who
form parts of it want to exercise their profession."
   -- Woodrow Wilson