|
Boost : |
From: Stefan Seefeld (seefeld_at_[hidden])
Date: 2005-10-04 15:07:11
Dave Harris wrote:
> In-Reply-To: <4342AC3C.1060100_at_[hidden]>
> seefeld_at_[hidden] (Stefan Seefeld) wrote (abridged):
>
>>Exactly. To be clear: moving from a 15" to a 17" monitor would, in my
>>ideal world, have *no* impact on the physical size of the displayed
>>graphics, but rather, would add ~ 1.4" in both dimensions to the useful
>>screen area. How many pixels your display has then only affects the
>>quality of the rendering, exactly as what you were looking for.
>
>
> Isn't that quite hard?
>
> Specifically, it sounds rather like the problem of drawing text nicely at
> various resolutions. Consider TrueType, or Type 1 Postscript fonts. The
> basic character shape is defined by vectors, by Bezier curves. To render
> it prettily, so that, for example, all the vertical strokes on an 'm' are
> the same thickness regardless of how the letter falls onto the pixel grid,
> you need to add what Adobe call hints and Microsoft call grid-fitting.
Indeed, small objects such as text are a problem, as they either have to
have a size that depends on the sub-pixel alignment or they will look bad.
I'm not sure what the solution is here. It seems that for small-size objects
we have to relax the device-independence a bit.
I'm still confident this can be done, since others have done it before
(DPS, NeWS, and nowadays MacOS X, I believe).
> Are you intending this system to be usable only on high-res desktop
> machines? Or are you including hand-held devices, phones and wotnot, that
> have relatively slow CPUs and where 640x320 is ambitiously high
> resolution?
I'm not sure. It's definitely easier to do on high-res devices.
>>Of course, this requires a completely vector-graphics-based GUI, which
>>is still the exception.
>
>
> Maybe there's a reason for that. I'm not saying don't do it, but I'd
> recommend doing some experiments that prove it's actually possible before
> investing too much.
Definitely, there are reasons. But as with the Y2K bug, it is mostly a
question of how long into the future you plan. Imagine new > 200 dpi
devices being ready for production but not usable as no GUIs are available
that can deal with that high a resolution.
Regards,
Stefan
Boost list run by bdawes at acm.org, gregod at cs.rpi.edu, cpdaniel at pacbell.net, john at johnmaddock.co.uk