Re: FAQ Posting problem



VK wrote:
Dr John Stockton wrote:
Better to post it to a standards-respecting system. One
cannot rely on such as Google to receive and post an article
unmodified, even if they do so at present. Also, it's getting
tempting to kill-rule anything posted /via/ Google.

That would be a highly unwise decision because say
comp.lang.javascript currently has 3460 active Gougle Groups
subscribers,

What evidence do you have that these subscribers are 'active'. Google do
not appear to timeout their subscriptions so all you could learn from a
total number of 'subscribers' is that that is the total number of people
who have subscribed and have never unsubscribed. Any individual
subscriber may have only ever used google groups once in order to appear
in those statistics (assuming that they used the subscribe to group link
during that visit). And that is assuming that you trust Google to add
the numbers up correctly, which doesn't seem to be a good idea given the
evidence that they are not capable of handling statistics effectively.

so this way you would cut off approx. 30% of the monthly
newsgroup traffic; and the next year (if nothing awful
happens with Google and the trend speed remains the same)
it will be 70%-90%.

If this were an issue you would expect to see total traffic on the
newsgroup rising, while in reality it has been pretty steady for the
last 5 years or so.

A bit of too high price to pay to not see VK anymore IMHO :-)

The problem with kill-filing people is that it then becomes difficult to
mitigate the harm that they may do by giving bad advice and making false
statements.


<http://groups.google.com/group/microsoft.public.scripting.jscript/about

and btw:
<http://groups.google.com/group/comp.lang.javascript/about>
<snip>

Those pages act to undermine Google's credibility as a source of
statistics. For example, there is a lit of "All time top posters" on the
'about comp.lang.javascript' page, supposedly stating the number of
posts made by the identified individuals. However, if you click on any
of the links in that list you get sent to the 'profile' for that
individual, also listing the number of posts that they have supposedly
made to the group. These two numbers are purported to be the same
statistic, but they do not correspond, and they differ in different
directions for different individuals.

When a system generates two different numbers for the same statistic you
can be certain that at lest one of those numbers is in error, and have
to suspect that both of them are. You certainly should not trust a
source of statistics that is not even consistent in its output.

(And that is without even observing that Google's 'profile' for me
asserts a total number of posts that is about 500 items short of my
newsreader's record of my posts to the group (which does not include any
posts I have made while away from home) and that their quoted total of
3664 should put me towards the middle of the "All time top posters"
list, in which I do not feature at all.)

Richard.


.



Relevant Pages

  • John Uebersax resorts to unsubstantiated claims and LIBEL
    ... John Uebersax PhD wrote, in a response addressed to Herman Rubin, ... JU> history of posts, has the opposite of constructive motives. ... Mr. Uebersax had posted were incorrect, as well as what Google ... that contain "Bayesian statistics" in it. ...
    (sci.stat.math)
  • Editorial: Guiness World Record for the shortest lived moderated group, "scistatmath"?
    ... 2006 believsky wrote about the Google group ... "scistatmath" he rashlyand impulsively created, ... the number of subscribers in sci.stat.math had INCREASED ... Statistics from a strictly mathematical ...
    (sci.stat.math)
  • Re: ! Where are the RST people?
    ... Google reports 636 subscribers. ... usenet servers (like I used to until I realized that I was getting ... wimbledon *spoilers* posts during the second week of the USO... ...
    (rec.sport.tennis)
  • Re: ! Where are the RST people?
    ... Google reports 636 subscribers. ... "subscribed" via Google Groups. ... unsubscribed posts, but I don't really have a clue. ...
    (rec.sport.tennis)
  • Re: The Ballad of Sir Patrick Spences Idiot Son
    ... Google Groups claims we have 401 subscribers. ... I don't know how they measure that - if that is only those subscribed through Google or if somehow they estimate those subscribed through newservers. ... I just told my news reader to read scs and if I post while reading scs my news reader posts to scs by default. ...
    (soc.culture.scottish)