[WEB4LIB] Re: testing web pages- a question

[Drew, Bill]

 I like Chris' choice of musicians to illustrate the points being made.

My plan is now to support browsers versions as follows:

1) IE4 and hgher (priority of IE5 and higher)
2) Opera 6.0 and higher
3) Mozilla 1.0 and higher
4) Netscape 4.7x and higher with support to disappear for versions lower
than 5 as time goes on.
5) Lynx, latest version
6) K-meleon (?) as long as it works.


I have looked at our server logs over the last year at different times and
90% of our traffic is for IE5 and higher.  This is because 90% of our
traffic is from on our campus where our standard was IE5 and is now IE6.

As many of you know, we are an IBM Thinkpad University campus.  Over two
thirds of our students will have IBM ThinkPad laptops this fall running
Windows XP with the latest version of IE and all of the latest plugins.
Those that don't have laptops have access to computer labs all around campus
to the latest versions of Windows and IE.


Bill Drew
drewwe at morrisville.edu


-----Original Message-----
From: Chris Gray
To: Multiple recipients of list
Sent: 8/12/02 3:26 PM
Subject: [WEB4LIB] Re: testing web pages- a question

Juggle a number of factors to suit taste and circumstances.

The general principles are:

0) KISS principle.  Somebody needs to rewrite Strunk and White for the
Web.  If you stick to the HTML code that even Lynx can understand and
work
on giving your pages good content, clear structure, and robust code,
you'll find it hard to go wrong.  These are values that can't be erased
by
technological advancement.  ("Website designers can help improve user
confidence by prioritizing quality and robustness over features and the
latest technology." --Jakob Nielsen)

1) Head toward standards.  Standards are aimed at getting us out of
having
to worry about these questions.  Validate and forget it.  _If_ you can
manage it, write to strict W3C standards with graceful degradation in
mind.  ("You can't always do what you're supposed to do" --Arlo Guthrie)

The more we gently but firmly insist on standards the better the
situation
will become.  (The extreme form of this is A List Apart's To Hell With
Bad
Browsers attitude.)

2) The needs of _your_ users.  If you're serving a population that's
still
using Cello and NCSA Mosaic, you'd better test with those.  If you are
obliged to serve people who need adaptive technology, you'd better test
with that.

3) The 20/80 rule, or "ars longa, vita brevis".   Surely, there are
better
uses for our time than testing Web pages for every browser under the
sun.
("There's not much more that you can do.  Go on and get some rest."
--Paul
Simon)

Chris Gray
Systems Analyst
University of Waterloo Library

On Mon, 12 Aug 2002, Drew, Bill wrote:

> I want to thank everyone for their help in testing my web page.  I
have a
> general question aimed at when a page has been tested enough.
> 
> I have gotten the page to work in Netscape4.77 and higher, IE5 and
higher,
> Mozilla 1.0, Opera 6.04, Kmeleon, and also Lynx.  I am getting input
from
> people using really buggy browser versions such as Netscape 4.03 and
older.
> 
> Where do I stop and say that is it?  What browser version do I cut off
at?
> 
> Wilfred (Bill) Drew
> Associate Librarian, Systems and Reference
> SUNY Morrisville College Library
> E-mail: mailto:drewwe at morrisville.edu
> AOL Instant Messenger:BillDrew4
> BillDrew.Net: http://billdrew.net/
> Wireless Librarian: http://people.morrisville.edu/~drewwe/wireless/
> Not Just Cows:http://people.morrisville.edu/~drewwe/njc/
> Library: http://library.morrisville.edu/
> SUNYConnect: http://www.sunyconnect.suny.edu/
> SUNY Morrisville College: America's Most Wired 2 Year College - 2001,
2000
> Books just wanna be FREE! See what I mean at:
> http://www.bookcrossing.com/referral/BillDrew 
>