[thelist] Site redirect check : old browser

aardvark roselli at earthlink.net
Thu Jun 7 00:27:31 CDT 2001


> From: "Mark Cheng" <mark.cheng at ranger.com.au>
[...] 
> 1)  Some background - our research indicated (this was based upon a
> simple google search of browser stats) that ie5 was the predominant

no offense, but that ain't research... i'm talking about polling your 
target audience... that means asking your current customers what 
they use, and polling prospective customers as well... you can do 
this simply with a few phone calls and lunches... and the results 
are actually useful to you, since they describe your users... the 
reports you read describe users to other sites, sites that aren't 
yours...

> browser in use. I've gone back today and updated for april 01 from
> these sites: websnapshot.mycomputer.com, ews.uimc.edu, w3schools.com. 
> Based upon those numbers, the ie5+ percentages were : 76.2%(not
> including AOL versions), 72%, 79% respectively.  No info was available
> on NS6 because Netscape was bundled v4.5+, so I'm not counting
> anything for that.

they wrapped N6 into their NN4.5+ numbers?  bad data, throw it 
out... not only do you have no N6 numbers, but your NN4.5 
numbers are bloated...

> We (myself and the Executive Directors) discussed that here and agreed
> that it was ok to design for ie5+, NS6 equivalent, but that we did not
> want anyone with less than that to hit the main site - because we
> didn't know what would happen.

so, you made your decision on flawed data from an incorrect pool 
of respondents...

and because you aren't prepared to download and test on 
alternative and older browsers, you decided to restrict access to 
the site because you didn't know what would happen...  that's a 
decision based on no research...

> As most requests we receive are for the latest annual, quarterly or
> press announcement (or employment!) and the number of hits(which is v
> low ) increases whenever we put something like that on the site, we
> decided to put a text only site together containing just those three
> documents, to help the immediate needs of those users who did not have
> a getElementById browser.

hell, i'd put those links right on your 'bugger off' page, and make 
them as plain vanilla HTML as possible, with the purpose of 
printability.... but that's just me...

> Therefore, we hope (being the operative word) that we will meet >95%
> of the needs of the users (even if they can't see the whizz bang
> site).  If the stats of the new site come up with something different,
> then we'll obviously reconsider.

but you'll never know without polling... you'll be satisfied when your 
numbers stay the same... and without vocal users, you'll assume 
that if the phone ain't ringing and site ain't down, that the 
clients/users are happy...

i can guarantee that you aren't meeting >95% of the needs of your 
users... i can't back that up, but experience has been good to me, 
and i've been down this road, and seen many others go down it as 
well...

> 2)  Using the definition of style which includes markup, i agree that
> I have not separated content from style. Full separation doesn't sound
> as hard to do as I thought. Thanks to .jeff for the ssi links - time
> to continue my learning :).  I think that I have separated style from
> markup, which was the intention, as at the time we didn't plan to
> serve this site up to anything but web users, so i didn't need to take
> into account WAP, Palm etc (I still don't know what RSS is?)

RSS allows you to provide XML feeds of your content so that other 
sites can ostensibly post little headline blurbs on their site showing 
the content on your site... they never need to change it, and you 
never need to change it... it's driven from the content source...

now, you may not need to serve it up to anyone but web users, but 
if you do, at least you know some of the issues you'll face now...

[...]
> that why I view them as object markers, not as style - but your point
> is true.

yeah, i'm not gonna touch that... since client-side script is so often 
low on my priority list, and since your site is your site, i can't 
speak to the value of the JS, other than to say that i see none, and 
the way you've coded the site, yes, your approach seems 
reasonable as far as accessing events... i don't agree with it, but 
that's just a subset of the overall issue, IMO, and it would take too 
long (and be too redundant) to address it...





More information about the thelist mailing list