Ian Hickson wrote:
On Fri, 21 Oct 2005, Matthew Thomas wrote:
...
I'll leave it in until someone comes up with a better idea, so that we
have a placeholder (and so that people who wish to experiment with the
idea can do so -- there seems to be at least some interest in it).
...
But there is already a better idea: redirects. As dolphinling said,
redirects will work while ping= doesn't. And the script you provided to
get around that not only adds even more complexity, it also won't work
for the 10 percent of visitors who don't have JavaScript turned on,
while redirects still work in that case too.
The script can be written backwards too, if that is a concern.
This still doesn't "force" it to work. As a user-tracking-implementer
doing it for money, I want to make absolutely sure I count properly.
That means forcing people to hit the counter _before_ even telling them
where they're going, so they can't get around it. There's no way to do
this with ping=.
But given the small proportion of authors who would use ping=
I think you underestimate the potential number of sites that would use
this. This kind of tracking happens a _lot_ and people are always trying
to find ways of making that experience better. There have already been
people on this list saying they want something like this just within the
last few hours.
The audience of people who would use tracking is huge. The audience of
people who would use ping= is, for the reasons I said before, much much
smaller.
Like I said before, I like the semantics of ping=. But it doesn't fit
into the usage model that advertisers and other trackers want.
Semantically, I want notification and linking to be separate. In usage,
they want them to be linked. They seem to me to be mutually exclusive.
--
dolphinling
<http://dolphinling.net/>
