[Top][All Lists]

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [Pan-devel] You need to incorporate Pre-processing tool (filters) to

From: Rajib Bandopadhyay
Subject: Re: [Pan-devel] You need to incorporate Pre-processing tool (filters) to control downloading
Date: Sun, 7 Sep 2014 16:44:50 +0530

Dear Sir, 

I have used PAN's scoring rules
in the past. It was often noted that an
email-based discussion forum had, and
has, irrelevant posts, often by trolls,
and whatever memory I retain of PAN,
there is no method to mark individual
messages as irrelevant, i.e., by low
scores, among a plethora of
messages belonging to a particular
thread (header). I wanted a pre-filter
desparately, but I knew of none who
actually did it. Claws was not within
my awareness horizon!

In claws-mail I can specifically use
the pre-filters to mark a particular
message to not download it, or a group
of messages in future, if it, or they,
comes, or came, from say, Jon Dhow
(spelling changed intentionally).

I am certain, as a former user of PAN,
of its special utilities, but
pre-filter incorporation can take it to
the next level, like, ..., okay, time
for an example: Once the
computer was for the most erudite,
because it was expensive and it was
worked upon by only a handful of people.
Today, it is ubiquitous in the form of
mobile phones. The difference is
because many, like you, have worked
hard day and night, to simplify things
for non-erudite users.

Scoring messages in terms of
discrete numbers is not intuitive and
simple as it appears, and is perhaps
even nerdy. It is difficult for an
average individual to maintain
objectivity while scoring, or even more
distant, remembering the evaluation
methodologies developed and continue
maintaining that standard. We have a
limited memory, cognitive window and
attention span.

Yes, it would altogether be different
if there was a distinct algorithm to do
it for us, but assuming that there are
-9999 to +9999 scoring points,
theoretically, there needs to be 19,998
seperate rules (or algorithms) just to
mark messages. And even more
importantly, it would need an elaborate
planning from a dedicated planner to
form those rules. Again, falling into
the hands of algorithm makers or
masters, like Stallman says.

But setting up filters, as many as
possible, with AND, OR, and NOT
combinations, based on phrases
contained, or not contained, are the
real aces -- intuitive and simple. It
saves bandwidth as well.

So let us incorporate this technique
within PAN by its next release.

Rajib Bandopadhyay

reply via email to

[Prev in Thread] Current Thread [Next in Thread]