Re: [Proposal] CSS gradients

Charles A. Landemaine wrote:
> On Dec 8, 2007 8:05 PM, fantasai <> wrote:

> visual effects) only with a style steet. Think about these large
> portals that have more than a hundred image files (ie: CNN, Yahoo...).

Sounds as though the page is too busy.

> Each three-way TCP/IP connection to download an image file make the
> page rendering slower. If most graphics were created with CSS, and

With any modern browser and server, you only need one connection, 
although I think all will open more than one.

Basically, any modern browser and server should use HTTP pipelining and
request multiple images on a single connection and without waiting for 
each one to return.

Even if you didn't pipeline, multiple connections would mean the TCP 
three way handshakes were overlapped.

As also mentioned, you can use data URLs, which are good for small 
images, where the printable encoding takes less space than the 
additional HTTP headers.

> disn't need an external graphics file, it would load a lot faster, it
> would load less the server and it would use less space or files on the

Providing you don't do something silly (although common), like trying to 
force a cookie on each image, the server should see hardly any of the 
requests, because images are generally static, and should be served with 
long dated cache timeouts and made publicly cacheable.

I'd certainly question any proposal to make CSS do things that can 
already be done (by a fullly conformant product) using existing W3C 

David Woolley
Emails are not formal business letters, whatever businesses may want.
RFC1855 says there should be an address here, but, in a world of spam,
that is no longer good advice, as archive address hiding may not work.

Received on Saturday, 8 December 2007 20:31:44 UTC