On Fri, 15 May 1998, Chris Dilworth wrote:
> >Yes, but it works with all browsers except AOL (we finally banned them
> >and their evil proxy servers from accessing the site).
>
> Would you elaborate on this for me?
This particular site uses cookies to track visitors from page to page.
In the application at issue, they go through a sequence of about 46 pages
and complete a bunch of fields. When they submit, the data is kicked
into an Oracle table, tracked by username.
The AOL proxies have no respect for any of the NO-Cache tag options.
What we found happening is that our pages would get into their monolithic
proxy cash and sit there for days. however, the accompanying cookies
expired in 3 days. So, the first person to reqeust a page from the
application at 72 hours and 1 second after a previous visitor retrieved
it would get it back with the other person's data--the proxy server would
send it by default rather than retrieving a new original.
You get really, really nasty calls to tech support when people's credit
card numbers get mixed up.
After repeated calls to AOL to ask them to simply turn off caching for
our domain (no useful results), and a weak attempt to keep up with the
changes in the darn-near-daily browser updates, we resorted to their
document collection and help files. They suck. If you read carefully,
they claim their entire cache system is proprietary, so they can't tell
you how it works. They do tell you they use special standards, which is
why you can't prevent caching. They also, in some instances, strip off
everything after the ? in URL encoded strings, so we couldn't use those
to genereate random numbers on each page (we thought maybe by doing that,
at 20000 visitors a day with 46 pages each, our 86K individual pages in
their cache daily might eventually annoy them . . .).
In short, AOL makes it darn hard to effectively deliver any type of
dynamic, time-sensitive data to their users. They're basically caching
every page that comes in on disk and saving it, and adding some apparent
shortcuts to cut down on what they perceive as "unnecessary" traffic to
the Internet.
Similarly, we do a lot of updating of files where we keep the original
file name. The updates are data some customers require, and they know,
for instance, that some data goes in Monday morning. So on Monday
morning they're there looking for it. Because we use the same filename
(so they can bookmark it), we also encounter a bunch of misconfigured
proxies from AOL that ignore the last modified info and feed them an old
copy--no matter how many times they reload. Get them off an AOL
connection, the data is there for them.
Thus, since they wouldn't tell us how to get around this, we found them too
much of a pain to serve, and most of the
people using our system could get to another non-AOL connection, we just
looked up the IP numbers of every AOL proxy we could find, filtered the
whole Class B block, then screened them for AOL proxy server names--and
if they matched, they get a page saying "sorry" and giving them
instructions on how to use the application from a real internet connection.
Enough generic elaboration, or was there something specific in there you
wanted more on? Basically, AOL is only a fraction less evil than MS in
my book :P
B
____________________________________________________________________
--------------------------------------------------------------------
Join The Web Consultants Association : Register on our web site Now
Web Consultants Web Site : http://just4u.com/webconsultants
If you lose the instructions All subscription/unsubscribing can be done
directly from our website for all our lists.
---------------------------------------------------------------------
