Re: Resolutions and combinations

At 18:49 12/03/02 +1100, John Aitchison wrote:
>There are so many possible combinations of client machine
>configurations that it is mind boggling. I just wrote down some of the
>factors that could go to make up the combinations..
<snip/>

You left out two very important ones, at least for windows machines. The 
display font size (which essentially changes the number of pixels per inch 
on the screen) and the setting that controls the width of window borders. 
I've seen people with both of these changed, and changed for good reason too.

There are also significant differences between the various Netscape 4.x 
browsers. Some are worse than others.

Cross browser testing is important, and it will remain so for some time to 
come. Very few websites have a sufficient budget to test in all the 
browsers and platforms that the site may be viewed in, and human nature 
being what it is if you have a complex time consuming testing process the 
chances are it will rarely be done. IN any case you will never be able to 
test in all browser and resolution combinations.

I recommend the following:

         Have a small range of browsers that is small enough so you can 
test as you develop, but wide enough to give a good impression of the 
different capabilities.

         Have a small range of resolutions that you routinely check on as well.

         Validate your HTML and CSS.

My testing setup consists of a dual monitor[1] PC with one monitor at 800 
by 600 and the other at 1024 by 768. I test all my sites on both monitors, 
fullscreen on the 800 by 600 and windowed on the 1024 by 768. This means I 
see the site on the smallest screen that is in common use, and in a window 
of variable size. Together these catch the vast majority of resolution 
problems.

For most sites, I use IE 5.5, Netscape 4.78, and the latest Mozilla. 
Together these browsers display most of the different characteristics you 
need to test with, but there are still only three of them. I used to test 
with multiple versions of Netscape 4, but I found it was so time consuming 
I never did it.

Most hideous browser problems are actually cased by invalid HTML or CSS. By 
validating these you avoid problems in, for example, browsers that haven't 
been released.

Of course, if you know a significant number of users will be using a 
different set up you should test on that as well.

Andrew

--
http://aamcf.co.uk/

[1] If your OS supports it, I really recommend getting a second monitor for 
your testing and development machine. It makes life so easy when it comes 
to testing at multiple resoultions. The cost is surprisingly low, given 
that the second monitor only needs to be 15" running at 800 by 600.

HTML: hwg-basics mailing list archives, maintained by Webmasters @ IWA