Re: Another Hell Mutt Discussion

On Sat, 18 Jul 2009 18:43:09 +0100, in rec.bicycles.tech "Clive
George" <clive@xxxxxxxxxxxxxxxxxx> wrote:

You argue like you are, and you seem to be ignoring the important bits of
what Frank's saying. You're not alone in that behaviour.

Well, I don't doubt that. Most Usenet discussions turn into a battle
of the lists. I take position A, so I go out and find studies that
say A. You say !A, so your studies all say !A. Thus, if your list is
longer than mine, you "win".

It's probably obvious by now that bicycle helmet laws are not my
primary field. I'm willing to take a survey of the existing
literature. A careful analysis of any significant study would take me
several hours... and I can pick *any* study apart if I'm willing to
invest the time; I'm not. Thus, I tend to use the abstracts and go
for what I term "emergent research" where I'm looking for a general,
albeit blurred, picture... I want to get the shape of the elephant,
not the color of its toenails.

As I progress, a few studies will stand out... not necessarily by
their quality, but by the number of times they're cited. Thompson,
et. al. is clearly one of these; the "Hurt Report" is another. I like
to get half a dozen of these... these I read.

My field is operating system security. There are literally thousands
of operating systems that have come and gone over the years. If some
obscure system surfaces and disappears in a month or so, then it will
be seen as secure because nobody bothered taking a hard look at it.
OTOH, if a system is heavily used, then people will find its holes and
exploit them... hackers don't exploit systems nobody uses. Thus,
you're in a Catch-22... if nobody ever uses it, it's secure.

The same is true of studies. When any study is heavily cited, it will
be the subject of heavy criticism.. When you have a "whole
population" study from Lower Slobovia that only applies there and has
been cited exactly six times in twelve years... well, if I haven't
made my point by now, then it won't be made.

Jones

.

Relevant Pages

  • Re: A Nice Richard Tucker Story
    ... nobody did until years after he passed away. ... Nobody wants to read your absurd posts, ... and checked the various lists and groups. ... I knew that Charlie had started up. ...
    (rec.music.opera)
  • Re: Building a new app in python and need some architecture advice.
    ... > I dont know exactly what i mean:) It doesnt matter to me as long as ... Well you have to define what "secure" is for you. ... Pyro's goal is to make the network as transparent as possible ... access remote objects as lists etc. ...
    (comp.lang.python)
  • Re: Background Intelligence Transfer Service
    ... Could not start the Background Intelligence Transfer Service service ... This service is running and lists BITS as a dependent service. ... operating system is Windows XP Home Edition SP2, ...
    (microsoft.public.windowsxp.general)
  • Re: A Nice Richard Tucker Story
    ... nobody did until years after he passed away. ... Nobody wants to read your absurd posts, ... and checked the various lists and groups. ... I knew that Charlie had started up. ...
    (rec.music.opera)
  • Re: set like table
    ... > What I do is use a hash table with each hash table key being the note ... - _merge_ several sorted lists of events. ... Nobody can fix the economy. ...
    (comp.lang.lisp)