I embedded the image of a space in each of last week's pieces on the barriers groups use to fence in their adherents while locking out their opponents. Since those images loaded from my own server I was able to learn something about the technology choices made by readers.
Here's what a typical log entry for the first time someone loads one of these looks like:
21X.5X.22X.2X - - [07/May/2007:02:24:34 -0600] "GET /images/sp_0507.jpg HTTP/1.0" 200 310 "http://blogs.zdnet.com/Murphy/?p=856" "Mozilla/4.0 (compatible; MSIE 6.0; Windows NT 5.1; SV1; .NET CLR 1.1.4322; .NET CLR 2.0.50727)"
As of right now (7:22 AM, Sunday May 13th) these logs show 13,574 unique IPs. The oddest thing? 65.4% overall show Firefox as the browser of choice - and the percentage among Windows and Linux users is exactly the same.
That's a change - I've embedded images in stuff before, and if I look at the record for all of 2006, the Firefox percentage is only 40.8.
Among last week's readers there were 671 Mac users who preferred Safari; 168 Linux users who opted for Konqueror; while only 20.8% of people using Windows stuck with IE; Thunderbird, Opera, and Netscape, have about 4% each; and, a left-overs group amounting to less than 0.5% split among browsers identifying as Alexa toolbar, K-Meleon, MultiZilla (SeaMonkey), BonEcho, and several others I'm equally unfamiliar with.
Overall:
The records show a whopping 36 Solaris users - 15 on x86 gear, 21 on SPARC. However two of those IPs trace to Sun internal network servers: meaning that an unknown number of their 32,000 or so internal Sun Ray users would be represented by those two addresses.
Among Windows records:
That low representation for NT 5.2 seems odd, but is broadly similar to the distribution for all of 2006 when something close to 2.6% of all hits identified as NT 5.2.
Three browsers self-identified as on Windows 95, 79 identified as Windows 98, and 22 others included Windows CE, a Safari user masquerading as IE on Windows 2000 (huh?) and a couple of palmsource users apparently masquerading as MSIE 6 on Windows.
Ubuntu dominates among Linux users with 51% of records and SUSE beats Red Hat on representation here by 135 to 32.
I don't know of a way to tell from the Apache server record whether a particular IP represents one reader or several, but 91% of IPs do not appear multiple times in intervals of ten minutes or less - suggesting that these represent one reader each time they appear.
Others recur in groups. For example:
6X.6X.16X.8X - - [10/May/2007:12:06:42 -0600] "GET /images/sp_0508.jpg HTTP/1.0" 200 310 "-" "-"
6X.6X.16X.8X - - [10/May/2007:12:12:51 -0600] "GET /images/sp_0507.jpg HTTP/1.0" 200 310 "-" "-"
6X.6X.16X.8X - - [10/May/2007:12:12:52 -0600] "GET /images/sp_0507.jpg HTTP/1.0" 200 310 "-" "-"
6X.6X.16X.8X - - [10/May/2007:12:12:54 -0600] "GET /images/sp_0507.jpg HTTP/1.0" 200 310 "-" "-"
6X.6X.16X.8X - - [10/May/2007:12:12:56 -0600] "GET /images/sp_0507.jpg HTTP/1.0" 200 310 "-" "-"
6X.6X.16X.8X - - [10/May/2007:12:12:57 -0600] "GET /images/sp_0507.jpg HTTP/1.0" 200 310 "-" "-"
6X.6X.16X.8X - - [10/May/2007:12:12:59 -0600] "GET /images/sp_0507.jpg HTTP/1.0" 200 310 "-" "-"
6X.6X.16X.8X - - [10/May/2007:12:13:01 -0600] "GET /images/sp_0507.jpg HTTP/1.0" 200 310 "-" "-"
6X.6X.16X.8X - - [10/May/2007:12:13:02 -0600] "GET /images/sp_0507.jpg HTTP/1.0" 200 310 "-" "-"
6X.6X.16X.8X - - [10/May/2007:12:13:04 -0600] "GET /images/sp_0507.jpg HTTP/1.0" 200 310 "-" "-"
6X.6X.16X.8X - - [10/May/2007:12:13:06 -0600] "GET /images/sp_0507.jpg HTTP/1.0" 200 310 "-" "-"
6X.6X.16X.8X - - [10/May/2007:12:13:07 -0600] "GET /images/sp_0507.jpg HTTP/1.0" 200 310 "-" "-"
31 IPs show 11 or more loads within 10 minute groupings, with the one excerpted above holding the record at a total of 80 in nine groups - and the next highest accounting for 27.
So, what can we conclude about ourselves? As a group we prefer Firefox, lean to Ubuntu Linux, and probably under-represent the Mac and Windows NT 5.2 communities while somewhat over-representing the Windows Vista population.