Thursday, August 25, 2005

Blogger and Explorer

I am a bit annoyed with Blogger lately. It suddenly stopped letting me log on with Explorer, insisting that I had to enable cookies and java (both of which are actually set correctly). When I went to the blogger help files, the info on Explorer settings only applied to Window's version or older versions of Mac, so that was no help. When I emailed support, they emailed back the exact same text from their help files, even though I had specifically pointed out in my email question that the help files did not apply to the version of Explorer I was running and that in any event, my software settings were the ones they said they needed. So their only suggestion was to dump Explorer in favour of Safari.

Yeah, all well and good, I can do that, but what about my 300 students, most of whom only have access to whatever software is available in the computer lab? I've been specifying for my blog assignments so that we can all be using the same software, but I may have to rethink that decision. It was clear that whoever answered my queries at Blogger could care less that I was using blogger as a class assignment and that this one glitch could affect hundreds of users. I was a bit disappointed, to say the least.

The other annoyance really isn't their fault, so I'm just whining about my luck, not really complaining. Having launched a fairly large scale study of blogs using the "Next Blog" button as the basis of my sampling, damned if they don't go and change the software to include a new "Flag this" button. I understand the need for such a button, which allows people who are cruising the "Next Blog" network to complain about obscene or offensive material -- not everyone appreciates coming across nude pictures or porn fiction or hate literature as they browse at random. Once flagged, Blogger staff look at the post, and if it is potentially offensive, they delist it from the "Next Blog" button (though leaving it on the web -- people are free to go there through search engines or through knowing the author or through direct links, but won't now stumble across it accidentally when "Nexting".) Okay, well and good, this seems to me an appropriate compromise to ensure that the average blogger/reader has a good experience, but it does screw up my sampling by introducing the bias that 'potentially offensive' blogs will be under-represented in my sample. Since the whole point of the study was to identify characteristics of typical blog, this is kind of a problem!

I also wonder what qualifies for the "flag this" compliant department. What offends me the most is not the occasional porn site or hate literature (all gist for the sociologist's mill) but the endless fake and commercial blogs. Some days as much as 75% of the "Next Blog" sample I collect consists of blogs with nonsense text (random words or letters) between search terms and links to a commercial site. I'm guessing that the purpose of these sites is not to be read, but to be indexed by google bots -- a 'black hat' trick unscrupulous webmasters use to raise the link ratings for the commercial site to which the fake blog links, since sites move higher in the search engines the more other sites link to it, and the more recently updated those links are. Creating free blogger sites is so easy, some of these guys can knock off and/or update fake blogs almost as fast as I can hit the next blog button (well, it takes me a couple of minutes to record the info into my study database), so that I often get three or four of these sites in a row, all obviously by the same guy (Blogger user name 'Phone1', next blog, it's 'Phone2', next blog it's 'Phone3', etc.) linking to the same mobile phone sales site, or whatever. I'd like to flag these bad boys so the Blogger staff could take them down (blogger is owned by Google, and Google naturally does not want the 'black hat' tricksters using its own free site to defeat it's ratings software.)

Of course, in terms of my study, the existence of these blogs is a significant and heretofore overlooked factor in estimating blog usage. All the studies of blog usage so far have looked at how many blogs are started on various hosting services (like, and maybe look at how many are abandoned vs frequently updated – but these fake blogs may be significantly inflating the numbers, since a single 'black hatter' might create dozens of fake blogs in support of one commercial site; and they would show up as 'frequently updated', as the evil genius goes in once a day and drops another chunk of nonsense boilerplate into each of his fake blogs and moves on to the next one as fast as his browser can rotate windows -- as I say, I often find ten or twelve of these fake sites in a row, the result of their being the most recently updated blogs when I happen to be doing my browsing.

Which brings up another frustration with Blogger. I can't actually find the analog for the "Next Blog" function described anywhere on the Blogger site. Most people assume that it is completely "random" (using that term in its popular, rather than technical sense) or that it represents the most recently updated blogs, but I can't seem to find any specific information, which I really need if I am going to be using this as the basis of my study sample. And I can't find an email address to which I can send my queries, though I'm sure those in the know would be happy to tell me what I need for my study, if not giving away any industry secrets. But given my disappointing previous contact with support, I won't be contacting them for answers. Well, I'll continue to poke around until I find someone suitable to ask.

No comments: