rday wrote:
>I've read this thread with interest. One of the people who I provide
>stats for - not in the least inclined towards astrology, but who has to
>market the product he's working on - has requested features like
>"determining stickiness".
If he requested a dollar for every entry in the log file, would you tell
him he was being unreasonable?
Ask him to define "stickiness". And then tell us why you think a program whose
sole purpose is to tell you what's in your log files is supposed to tell you
something that isn't recorded in those log files.
>Maybe it's just me, but I cannot agree with the "ethical" stances taken
>dismissively by most of the thread. I provide a service: I look after data and
>allow people to question that data. I see no reason to limit the scope of their
>enquiry, even though the queries are fuzzy, and the results fuzzier. The people
>who ask these questions should not be limited by the constraints I place on
>them - and if I do put constraints on them, I regard this as a failure on my
>part, just as much as I would regard it as a failure if I asked a question but
the answer was "Oh no, you can't ask that - it's the wrong sort of question".
It's not the "wrong sort of question" - it's a question for which there
is no accurate answer. Nobody is saying you can't ask it, you just can't
demand an accurate answer, because it's impossible to give an accurate
answer. You're not putting constraints on anyone, the constraints that
do exist are because "that's the way the web was designed to work".
>I find the attitude implicit in such an answer narrow and disappointing.
I find the willingness to delude customers disappointing (but not exactly
unusual in this day and age). And the fact that some service providers
aren't deliberately deluding their customers, because they aren't
technically competent to understand what they're doing is at least as
worrying.
I don't have any problem with using Analog to look for trends, or unusual
activity, and then using this information, in conjunction with your own
knowledge about the site, to make inferences about what's happening
("hey, it looks as though that link on slashdot really drove up our
numbers").
>I have to satisfy the queries of my customers - if not, they will go
>elsewhere. In this case, I have to get another package to satisfy their
>queries hence I get to maintain another package (joy!), probably
>windows (joy!!), both sides get to feel superior (joy!!!).
Why go to all that trouble if you don't actually care about accurate
answers - just create a small random number generator. Tell the customer
that they can have this information for free, or they can pay for more
elaborate (not necessarily more accurate) information if they want.
> All of which is a shame: I always thought answering the fuzzy
>questions was part of what this programming vocation is all about.
???
>You could read this as a defense of marketeers and maybe it is. Then again it's
>also a defense of tolerance and pragmatism.
Nuts. Defending ignorance isn't tolerant and pragmatic.
> It's also a vote for folding "stickiness algorithms" into analog.
So write the stickiness algorithms, and see if the world beats a path to your
door.
Aengus
------------------------------------------------------------------------
This is the analog-help mailing list. To unsubscribe from this
mailing list, send mail to [EMAIL PROTECTED]
with "unsubscribe" in the main BODY OF THE MESSAGE.
List archived at http://www.mail-archive.com/[email protected]/
------------------------------------------------------------------------
