Read the linked page. It has its own disclaimer about stats.
It does indeed and it illustrates many of the points already made in
this thread. To quote it:-
|<URL:
http://www.w3schools.com/browsers/browsers_stats.asp >
| Statistics Are Often Misleading
|
| You cannot - as a web developer - rely only on statistics.
| Statistics can often be misleading.
The absolute statement that statistics cannot be relied upon if offset
slightly by stating that they can be misleading. That implies that it is
failure to properly interpret on the part of the individual mislead that
is the problem instead of the impossibility of gathering meaningful
statistics. A more hones disclaimer would just read "There statistics
are inaccurate", which would remove the potential for them being
misleading as nobody would even consider drawing conclusion from them.
| Global averages may not always be relevant to your web site.
Not that they are claiming that these are global averages.
| Different sites attract different audiences.
| Some web sites attract professional developers using professional
| hardware, other sites attract hobbyists using older low spec
| computers.
| Also be aware that many stats may have an incomplete or faulty
| browser detection.
To say "may have an incomplete or faulty browser detection" is a bit of
an understatement as we know that server-side UA header information is
next to useless for browser discrimination and that client-side browser
detecting is almost as inaccurate. We can also be certain that whenever
client-side scripting is disabled the only browser detecting strategy
available is the server-side UA header reading.
| It is quite common by many web stats report programs,
| not to detect new browsers like Opera and Netscape 6
| or 7 from the web log.
And it is almost certain that IceBrowser, Web Browser 2 and any other
browsers that put some effort into spoofing IE will be lumped in with
the IE statistics.
| (The statistics above are extracted from W3Schools' log-files,
All else being equal, given the disclaimer above about different sites
attracting different audiences, admitting that there statistics are
derived directly form the logs of one site would suggest that, even if
accurate, they should only be of significance to the organisers of that
site.
| but we are also monitoring other sources around the Internet
| to assure the quality of these figures)
Meaning that if our statistics resemble other statistics they must be of
similar quality. Which doesn't make them good quality statistics, more
it demonstrates that the factors that render any individual statistics
gathering endeavour meaningless apply equally and similarly to them all.
Which is the main thrust of the entire thread.
As with all published web statistics, there is a distinct lack of detail
about how the statistics have been gathered and analysed. Browser
detection figures, for example, really should be accompanied by the code
used for testing else it is not possible to tell how the more difficult
to identify and spoofing browsers will be reported. Questions like
whether the logging is done from one HTML page or many. Or are images
used in the logging (with the obvious implications for the reporting of
browses incapable of showing images, or with image display disabled).
And, if images, then client-side Image objects or HTML IMG elements.
Without the details of what is being measured and how that is being
done, even without the problems inherent to the network, the results can
be nothing more than labelled numbers. Any meaning that could be
attached to them could only exist in the mind of the reader, and as a
result they can be nothing but misleading.
Possibly you are missing the distinction between a Usenet discussion
group and a help desk.
It is an important distinction because if the group allowed itself to
descend into becoming a vending machine for quick fix code we would be
no better than the many appalling copy-and-paste script collections that
exist. Instead of working to expand an appreciation of the skill-set
needed for effective cross-browser script design and implementation we
would be actively contributing towards making the Internet worse for
everyone.
If you consider that c.l.j should be churning out ill-considered quick
fixes in response to posted questions without even mentioning the issues
let alone addressing them, then nobody can stop you. But such postings
will (should) attract comment.
A couple of months ago, after commenting on such a posing, I received
the reply "My proposal is not a solution for the actual problem.". How
was the OP going to appreciate that if they were unable to solve the
problem without assistance? Should his post have gone uncommented and
the issues unmentioned? The real problem was with server-side code that
hadn't been designed to work with its interface over HTTP, and there
were no client-side solutions. Any javascript proposal would have been
bad and, even if reliable, would have been solving the wrong problem.
If you ask questions of honest, responsible people with no vested
interest they are likely to tell you what they think you need to know,
that is not necessarily always what you want to hear.
I also thought that the OP was looking for something beyond the rather
trivial question asked. My impression was that it was an attempt to
justify or refute some sort of design decision. Insofar as providing
references to "data" goes, beyond the truth that a public web site with
any significant number of visitors will get visited by javascript
incapable/disabled browsers, there is no "data", just bogus statistics.
So it makes more sense to explain why there is no factual information on
the subject and that design decision really should not be cloaked in
"information" that is known to be inaccurate, they should be made to
stand (or fall) on their own.
The OP asked for data, not stats
Then why did you post the URL of a page with nothing more than labelled
numbers?
Hmmm. Get personal? Or do something to actually improve the abysmal
noise to signal ratio? Screw the people who come to this newsgroup
for a little guidance: let's get personal instead!!
If you are going to derogate the entire group (and therefor anyone who
participates, especially regularly), your position in that one to many
relationship will mean that any response you induce will be directed at
you personally. (Granted comparisons with George Hester cannot be
intended to do other than offend but implying responsibility for a
"sorry state" or an "abysmal noise to signal ratio" should be expected
(maybe, was intended) to antagonise.)
Richard.
