When reported trends don't match observed trends

Posted by Jim Jagielski on Monday, August 6. 2007 in ASF

No doubt, there will be many posts regarding the latest Netcraft survey, some good, some bad and some just plain FUD. Even ignoring other surveys that show vastly different results, the Netcraft report is confusing for one real reason which has nothing to do with Apache at all. For the survey to match reality, it means that web hosting providers are selecting Windows over Linux, or Solaris, or BSD or OS X or, well, basically, any other OS. Well, of course that's not the case. Is anyone seeing that? At anything close to the numbers that would be required? No, of course not. It simply shows that once one knows the secret sauce on how rankings are made, one can make a very directed effort to shift those to reflect the version of reality you want to present. We saw it with web search rankings and we're seeing it now. It's almost as if someone said "We are ranking big business using the metric of how many phone lines they have; the rationale being more phone lines means more employees which means a bigger business." Once that's done, it's easy for a small business, who has a vested interest in appearing much larger than they are, to simply order more phone lines, increasing their ranking. With all that being said, it is quite nice seeing lighttpd, which is an Apache-"based" variant, making increases. It shows just how useful FastCGI still is and how neat mod_proxy_fgci will be, once we finish it, that is.

The author does not allow comments to this entry


Search for an entry in IMO:

Did not find what you were looking for? Post a comment for an entry or contact us via email!