Re: link areas within images

cailliau@cernnext.cern.ch
From: cailliau@cernnext.cern.ch
Date: Mon, 5 Jul 93 11:16:21 +0200
Message-id: <9307050916.AA00794@www2.cern.ch>
To: Guido van Rossum <Guido.van.Rossum@cwi.nl>
Subject: Re: link areas within images 
Cc: www-talk@nxoc01.cern.ch
Status: RO
> It's not just character grid PCs, most bitmapped displays (e.g.
> X11, Mac) do it it this way, too.  My choice would depend on the  
kind
> of coordinate system to be used; if it were directly related to
> pixels, let Y point down, but if you normalize to e.g. [0.0, 1.0]
> then Y would best point up, in correspondence with standard
> graph conventions. (Just trying to settle the argument with
> facts instead of personal feelings...) 


Well, there are no personal feelings here: if the image is a photo,  
then Y should point up: I believe that when I see a map or a  
landscape, I measure things higer in higher coordinates.
If the image is a drawing, then all draftsmen would agree to have Y  
point up.
The ONLY case in which I can see a slight advantage to have Y point  
down is text, and that should be editable text, not pixels. A  
PostScript original image certainly has Y pointing up. This is also  
the only reason that was taken for pointing Y down on bitmapped  
displays: it was just an extension of the line-count on character  
grid terminals.
Just think of placing the sensitive rectangles using real tools on a  
screen: When I place a rectangle in any serious drawing program I  
know of (I do not include MacPaint of course!), the position of the  
mouse is given with Y pointing up.
This is what we have to think of, not what has already been done.
Robert.