I reread this original post.
The HTTP protocol calls for a header variable named "HTTP_ACCEPT". This variable defines which MIME-TYPES are accepted by the client's browser.
This variable should be constant for each client browser, that is to say, it doesn't change what it will accept based on the type of query it is doing.
So in general, if a client says it accepts text/html and */* it will accept everything, as demonstrated by the */*.
When a client browser says */* and other stuff, it's just additive. It already said that it supports */*.
The base point here is that you can't base the type of download being made based on the http-accept, http-accept purely announces to the server which MIME-TYPES the client browser can accept.
We have reviewed the logs from a transaction web site, which does pure HTTP POSTINGs. We have grepped out the client browser type and the http-accept variables and show some output below:
[HTTP_USER_AGENT] => Mozilla/4.0 (compatible; MSIE 6.0; Windows 98)
[HTTP_ACCEPT] => image/gif, image/x-xbitmap, image/jpeg, image/pjpeg, application/x-shockwave-flash, */*
[HTTP_USER_AGENT] => Mozilla/4.0 (compatible; MSIE 6.0; Windows 98; Win 9x 4.90)
[HTTP_ACCEPT] => image/gif, image/x-xbitmap, image/jpeg, image/pjpeg, application/vnd.ms-excel, application/vnd.ms-powerpoint, application/msword, application/x-shockwave-flash, */*
[HTTP_USER_AGENT] => Mozilla/4.0 (compatible; MSIE 6.0; Windows NT 5.0; iOpus-I-M)
[HTTP_ACCEPT] => image/gif, image/x-xbitmap, image/jpeg, image/pjpeg, application/x-shockwave-flash, */*
[HTTP_USER_AGENT] => Mozilla/4.0 (compatible; MSIE 6.0; Windows NT 5.1)
[HTTP_ACCEPT] => text/html, image/png, image/jpeg, image/gif, image/x-xbitmap, */*;q=0.1
[HTTP_USER_AGENT] => Mozilla/4.0 (compatible; MSIE 6.0; Windows NT 5.1) Opera 7.11 [en]
[HTTP_ACCEPT] => image/gif, image/x-xbitmap, image/jpeg, image/pjpeg, application/vnd.ms-excel, application/vnd.ms-powerpoint, application/msword, application/x-shockwave-flash, */*
[HTTP_USER_AGENT] => Mozilla/4.0 (compatible; MSIE 6.0; Windows NT 5.1; Alexa Toolbar)
[HTTP_ACCEPT] => image/gif, image/x-xbitmap, image/jpeg, image/pjpeg, application/x-shockwave-flash, */*
[HTTP_USER_AGENT] => Mozilla/4.0 (compatible; MSIE 6.0; Windows 98; Win 9x 4.90)
[HTTP_ACCEPT] => image/gif, image/x-xbitmap, image/jpeg, image/pjpeg, application/x-shockwave-flash, application/vnd.ms-excel, application/vnd.ms-powerpoint, application/msword, application/x-gsarcade-launch, */*
[HTTP_USER_AGENT] => Mozilla/4.0 (compatible; MSIE 6.0; Windows NT 5.1; .NET CLR 1.1.4322)
[HTTP_ACCEPT] => image/gif, image/x-xbitmap, image/jpeg, image/pjpeg, application/x-shockwave-flash, application/vnd.ms-excel, application/msword, */*
[HTTP_USER_AGENT] => Mozilla/4.0 (compatible; MSIE 6.0; Windows NT 5.1; .NET CLR 1.1.4322)
[HTTP_ACCEPT] => image/gif, image/x-xbitmap, image/jpeg, image/pjpeg, application/vnd.ms-powerpoint, application/vnd.ms-excel, application/msword, application/x-shockwave-flash, */*
[HTTP_USER_AGENT] => Mozilla/4.0 (compatible; MSIE 6.0; Windows 98)
So the point is, HTTP_ACCEPT says what the browser WILL accept, not what this particular request IS getting. Here's an example page supporting this:
http://www.billwood.com/ssivar.htm
If the test script considered all "*/*" versus all those that (lack */* and HAVE IMAGES) then it would show which client browsers were not accepting anything but images, and were not accepting HTML.
Simple fact is, we're not hit-botting anyone. It's awful hard to prove a negative.
But I have shown above that simply looking at whether or not HTTP_ACCEPT has an IMAGE in it is not a good test.
Quote:
Originally posted by cluck
Well a little while ago I wrote some php to test if I was getting img src hits. It logs the accept: headers into a text file for one. The other part is it increments 2 values in a mysql table. One is called through a php include(so it'll load even if it's img srced) and the other is called through an img tag so it will only load on a real hit.
So far it's about 40% img src traffic. It's only been a few hundred hits, let's see how well the rest of the 10k goes.
|