Re: Why style sheets

James Aylett (sleeper@cryogen.com)
Fri, 17 Jan 1997 22:39:58 +0000 (GMT)


Date: Fri, 17 Jan 1997 22:39:58 +0000 (GMT)
From: James Aylett <sleeper@cryogen.com>
To: Eric Holstege <Eric_Holstege@broder.com>
cc: www-html@www10.w3.org
Subject: Re: Why style sheets
In-Reply-To: <00001A25.1407@broder.com>
Message-ID: <Pine.LNX.3.93.970117222402.16160C-100000@crystal.clare.cam.ac.uk>

On Fri, 17 Jan 1997, Eric Holstege wrote:

> Why should I pay the penalty of two HTTP accesses per page (one for the text, 
> another for the style sheet file) on *every page access*, when I can pay a 
> penalty for global search and replace only for every *style update*, which 
> happens much more infrequently. In the latter case I have to be a bit 
> intelligent about how I mark up my pages. In the former case I am in some sense 
> doubling the load on the network, my server, and my filesystem.

Not true. Once a stylesheet has been downloaded once it won't have to be
transferred again. The browser can simply ask the server, whenever
specified by its settings, as to whether the stylesheet has changed, and
if not, to continue using the old copy. You change one document, once, and
next time someone visits your site their copy is updated too. Obviously if
it disappears from their cache then they need to download it again, but
the efficiency of this can be increased by using proxies with
correspondingly larger caches. (As an aside, it would be nice if
stylesheets could reside in a separate cache, so that only other
stylesheets will displace them.)

> Also, why is it worse to add HTML tags or tag attributes than to add style 
> properties. There are well defined rules for ignoring unknown HTML tags and 
> attributes. Perhaps someone could explain.

Apart from the above reason - along with the fact that most complicated
documents will typically get smaller with the use of stylesheets - there
are several other points.
I know I'm only reiterating to many people on this list something which
has appeared many times in the past, but it is better in general to tag
something by what it means ("a paragraph", "a piece of Italian", "part of
a speech by JFK") than what it looks like ("12 point Times Italic in Key
black with pure white strikeout, or 14 point Chicago bold in Ochre on
Saffron if Times isn't available", "any italic serifed font", "bright blue
in 40 point sans serif with a sky blue wavy border cornered with an
'IMPORTANT' logo"). I could go on - what about a presentational definition
which is six pages long, depending on exactly what output methods - plain
text, graphical, sound by .wav file, sound by speech synthesis, output as
braille ... - are available? If you wanted to use that twice in two
separate documents which are only a page long each, in what way has adding
all the relevant tags and attributes to the HTML helped you?

James

-- 
/-----------------------------------------------------------------------------\
  James Aylett - Crystal Services (crystal.clare.cam.ac.uk): BBS, Ftp and Web
   Clare College, Cambridge, CB2 1TL -- sleeper@cryogen.com -- (0976) 212023