[OT] dynamic resolution web sites
Jason Dewayne Clinton
me at jasonclinton.com
Sat Nov 5 16:05:35 CST 2005
On Saturday 05 November 2005 09:57 am, you wrote:
> This is very interesting. However, playing devil's advocate, I see
> two serious problems with svg for web pages.
>
> 1. Complex SVG files are huge. Those node and fill definitions get
> massive in a hurry.
This is true. However, because the information is text, it is highly
compressible (compression ratios of around 0.2). And that's why svgz is
part of the standard. And of course, no one is advocating using SVG all
the time; just when it's appropriate (user interfaces, text and
clipart).
> 2. Complex SVG files are mega-cpu intensive. Everytime you try to
> scroll the page, it shuts down your computer for x seconds to redraw.
This is true inside of an SVG editor and perhaps also in SVG animations.
However, on static SVG images in web pages, the SVG is rasterized and
then blitted to the scroll buffer for hardware accelerated scrolling
(this is true for Firefox, Konqueror and Opera). So there should be no
difference once the picture has been rasterized. There might be a
problem on really low video memory systems, though. (Like less than 8MB
of VRAM.)
> But, I definitely do like the concept. I've talked about this
> before, but try http://www.zeni.net/trf/mgspec/157.php for scaleable
> graphics. I generate a libraray of jpgs with different scaling for
> the most common screen resolutions, although the user can plug any
> "zoom" value that he wishes into the location bar and the server will
> choose the closest sized image and scale it for you on the fly. Php
> is your friend.
Yes I think this is very interesting approach. I considered doing it
using the Batik [1] server-size SVG rasterizer which can do almost
exactly what you suggest but in the end the unresolved printing issue
turned me off. But I still think there's some very useful applications
for this approach.
[1] http://xml.apache.org/batik/
> Yes, these catalog page background images start out as vector but
> they are mega smaller when scaled and converted to jpg raster.
Yes that's a huge improvement. It would print poorly but its ideal for
screen display. This is also how the DejaVu file format works. It
obtains huge compression ratios versus PDF but it isn't well suited for
printing.
> On the other hand, experts have been telling me for years that the
> whole world will be using broadband by Tuesday.
We're at about 40-50% market saturation in the US aren't we? Maybe
not...
--
I use digital signatures and encryption. My key is stored at pgp.mit.edu
0x8DB3BF09 FP: F628 D9D3 E57A C281 5EFE - 7DF7 B52A A393 8DB3 BF09
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 189 bytes
Desc: not available
Url : http://kclug.org/pipermail/kclug/attachments/20051105/bb66a099/attachment.pgp
More information about the Kclug
mailing list