Re: webpage hit counter in PHP

"Animesh K" <animesh1978@gmail.com> wrote in message
news:fno272$11ee$1@agate.berkeley.edu...
> Steve wrote:
>> "Animesh K" <animesh1978@gmail.com> wrote in message
>> christ almighty! 'relevancy' has to do with the *materials* ms has that
>> is most *appropriate* to what i'm searching for. not only is this
>> personalize and very specific to me and my keywords/phrases, the results
>> can be *ranked* as to whether or not the result(s) were *relevant* and to
>> what extent. NONE of that machinery has ANY counter mechanism to it...AT
>> ALL.
>>
>
> Well someone asked why are counters useful?
>
> I told they can tell you commonly sought articles/videos/ or commonly
> tread paths.
>
> You told you want to take the unpopular path in an effort to argue that it
> will be the right path.
>
> WTH has counter to do with the above statement of yours? You wanted to
> take unpopular paths, which excludes the most popular websites -- which
> you do use. Writing an essay or two will not make your past argument
> correct.

yet if you read *any* of the 'essay' (or understood english), you'd not
restate your stupidity.

counters *don't* help you find meaningful content (daft meter registered)
counters *don't* indicate popularity (daft meter elevated)
popularity *doesn't* help you find meaningful content (daft meter elevated)
NONE of your examples use counters. (daft meter too high to register)

"Animesh K" <animesh1978@gmail.com> wrote in message
news:fno20d$11cv$1@agate.berkeley.edu...
> Steve wrote:
>
>>
>> yes, please do. you'll soon learn in the process that counting will not
>> 'add up' to what is afforded in web page analysis packages. counting is
>> NOT the primary means of producing any of the analysed information.
>
> Read about indicator functions, if you haven't so far. That should lead
> you to the formula. Unfortunately, I cannot write math formula here.

i doubt you can write much of anything...anywhere.

..oO(Animesh K)

>Michael Fesser wrote:
>
>> That's a lot more than just counting. I wouldn't call the server's
>> logfile a "counter".
>
>It's your choice then. I will still treat the ip-address, time of hit,
>number of hit as a "counter vector" which is more sophisticated than
>just 1, 2, 3, ...

A logfile "counts" hits, that's it. The interpretation of the data is
left to the logfile analyzer.

>>> It would be better if you (both) sit and think what a (general) counter
>>> would mean before telling us what google analytics is, or what it is not.
>>
>> 1, 2, 3, 4 ... this is counting. And that's what a hit counter does,
>> nothing more. But GA _does_ a lot more than just counting.
>
>hits per day, hits based on location, hits in the past 1 hour, trend of
>hits, hits per month, hits to unavailable pages, .. all these are
>(conditional) counters. Should I write a formula now?

Most of these informations are not stored in logfiles, but extrapolated
from them, using different algorithms. Different analyzers will lead to
different results.

Micha

Michael Fesser wrote:
> .oO(Animesh K)
>
>> Michael Fesser wrote:
>>
>>> That's a lot more than just counting. I wouldn't call the server's
>>> logfile a "counter".
>> It's your choice then. I will still treat the ip-address, time of hit,
>> number of hit as a "counter vector" which is more sophisticated than
>> just 1, 2, 3, ...
>
> A logfile "counts" hits, that's it. The interpretation of the data is
> left to the logfile analyzer.
>
>>>> It would be better if you (both) sit and think what a (general) counter
>>>> would mean before telling us what google analytics is, or what it is not.
>>> 1, 2, 3, 4 ... this is counting. And that's what a hit counter does,
>>> nothing more. But GA _does_ a lot more than just counting.
>> hits per day, hits based on location, hits in the past 1 hour, trend of
>> hits, hits per month, hits to unavailable pages, .. all these are
>> (conditional) counters. Should I write a formula now?
>
> Most of these informations are not stored in logfiles, but extrapolated
> from them, using different algorithms. Different analyzers will lead to
> different results.


Agreed. But you can also think of many popular analyzers as a
conditional counter. I wonder if you have heard of the indicator functions.

To me a counter is just a summation of individual terms. The individual
terms *can be* a function of the raw data. The raw data need not be
displayed as it is.

Steve wrote:

> counters *don't* help you find meaningful content

agreed.

> counters *don't* indicate popularity

ok.

> popularity *doesn't* help you find meaningful content

sure.

You omitted an application that I think is useful: Counters may help to
keep track of the *relative* popularity of items within a defined set.

I have an application that features a growing set of images, currently
about 1,000. There are moderately small versions of these images as well
as blow-ups. I count accesses to the blow-ups and weigh them inversely
by their age.

As I add new images, and as my server space is limited, I need to remove
blow-ups from the bottom of the "charts". What's wrong with that?

--
Gregor mit dem Motorrad auf Reisen
http://hothaus.de/greg-tour/

"Greg N." <yodel_dodel@yahoo.com> wrote in message
news:fnpn6b$ufv$1@online.de...
> Steve wrote:
>
>> counters *don't* help you find meaningful content
>
> agreed.
>
>> counters *don't* indicate popularity
>
> ok.
>
>> popularity *doesn't* help you find meaningful content
>
> sure.
>
> You omitted an application that I think is useful: Counters may help to
> keep track of the *relative* popularity of items within a defined set.

if by 'popularity' you mean 'frequency', ok.

> I have an application that features a growing set of images, currently
> about 1,000. There are moderately small versions of these images as well
> as blow-ups. I count accesses to the blow-ups and weigh them inversely by
> their age.
>
> As I add new images, and as my server space is limited, I need to remove
> blow-ups from the bottom of the "charts". What's wrong with that?

nothing is wrong with that. that's not what's being argued.