[Wylug-help] Browser Detection (and HTML loops)

[Dave Fisher]

On Mon, Nov 22, 2004 at 06:31:55PM +0000, Smylers wrote:
> Dave Fisher writes:
> > I think I am aware of most of the difficulties, including the often
> > well-intentioned, but misguided practice of user agent spoofing.
>
> I'm not sure how misguided it is, especially if done well.  For example
> a site that only works in 'IE' usually checks for the string "MSIE" in
> there.  So 'Opera' in 'IE'-spoofing mode includes that string -- but it
> still leaves "Opera" in there too.  So by checking for 'Opera', gecko,
> and so on before 'IE' you avoid most of the spoofing problems.

Ah, when I said 'misguided' I was thinking about the problem in terms of
its socio-political effects on common human behaviour rather than on the
technical problem of browser detection, i.e. that because browser
detection is often practised in a very unsophisticated way, it causes
non-IE browsers to be undercounted and, hence, reduces pressure on
IE-only development teams to mend their ways.

In fact, that comment was really an aside to the main point that I was
trying to make, which was that the presence of spoofing doesn't
necessarilly undermine the plausibility of browser detection, since it can
dealt with in precisely the way Smylers suggests.

In short, the existence of bad browser detection is not a decisive
argument against browser detection per se  ... and not much of an
argument at all if there are no better alternatives available.  Anyone
heard of one?

Dave