From mboxrd@z Thu Jan  1 00:00:00 1970
Message-ID: <73e92e797fe4409d9224bfc51969f0c3@caldo.demon.co.uk>
To: 9fans@cse.psu.edu
Subject: Re: [9fans] sub-pixel
From: Charles Forsyth <forsyth@caldo.demon.co.uk>
In-Reply-To: <20031026174839.GA21901@ionkov.net>
MIME-Version: 1.0
Content-Type: text/plain; charset="US-ASCII"
Content-Transfer-Encoding: 7bit
Date: Sun, 26 Oct 2003 21:06:27 +0000
Topicbox-Message-UUID: 799b58e0-eacc-11e9-9e20-41e7f4b1d025

i'm sorry that earlier message was a bit abrupt because i was in a hurry,
and it sounded a little grumpy although i wasn't.

i've since had time to look up a cheerfully enthusiastic
summary on a web page http://grc.com/ctwhat.htm,
and reduced my ignorance a bit, assuming the page is accurate.
unfortunately, now that i've read it, i'm feeling uneasy about
building so many assumptions about a particular implementation
of the display device into the drawing primitives, especially
a technique with so many restrictions.   you're probably
right that it's an easy way to get the effect you seek,
and indeed i can't suggest a sensibly constructive alternative
if that's what you aim to do, but is it the right thing to do
to the primitives?

it's possibly not as bad as the decision to build a particular
gamma correction into all image values, as is sometimes
done elsewhere, but is it really that different?  for instance,
with `sub-pixel' rendering (as i understand it) the resulting
image if stored and redisplayed won't look good
as you change the type of display.  the `only landscape
not portrait' aspect is also unsettling (it won't work on
my iPAQ, again, as i currently understand it).  unfortunately unlike
gamma correction it seems hard to shift the transformation
to the display, although that's what i'd have preferred
to do in this case.