Lower all threads started by a particular tro... er, person?

Lower all threads started by a particular tro... er, person?

[Dave Abrahams]

Hi All,

There are some groups I read where one person reliably starts
unproductive threads that I'd like to ignore entirely.  Is there any way
to use scoring to make that happen?

Thanks,
Dave

Re: Lower all threads started by a particular tro... er, person?

[Dave Marquardt]

Dave Abrahams <dave@boostpro.com> writes:

> There are some groups I read where one person reliably starts
> unproductive threads that I'd like to ignore entirely.  Is there any way
> to use scoring to make that happen?

I believe "L S" lowers the score of an article. You could lower based on
author. If you lower it far enough, I think there are other settings
that would make the thread not appear at all, I think the keyword to
search for in the Info pages is "expunge".

-Dave

Re: Lower all threads started by a particular tro... er, person?

[Steinar Bang]

>>>>> Dave Abrahams <dave@boostpro.com>:

> There are some groups I read where one person reliably starts
> unproductive threads that I'd like to ignore entirely.  Is there any way
> to use scoring to make that happen?

`L a' permanently lowers the score on the contents of the "from" field.
You can then interactively decide whether to use a regexp or just a
substring match.  The default is the from field of the article you do
the command on, but the string is editable.

Whether this ends up in the group score file, or in the global score
file for that server, or the global overall score file, depends on which
score file you visited last (ie. default is the group score file, but if
you wish to do a global down-score, just visit the global score file
first).

Re: Lower all threads started by a particular tro... er, person?

[Adam Sjøgren]

Steinar writes:

>>>>>> Dave Abrahams <dave@boostpro.com>:
>
>> There are some groups I read where one person reliably starts
>> unproductive threads that I'd like to ignore entirely.  Is there any way
>> to use scoring to make that happen?
>
> `L a' permanently lowers the score on the contents of the "from" field.

Does this hide the entire _thread_?


  Best regards,

    Adam

-- 
 "I wonder if you can refuse to inherit the world."           Adam Sjøgren
 "I think if you're born, it's too late."                asjo@koldfront.dk

Re: Lower all threads started by a particular tro... er, person?

[Steinar Bang]

>>>>> asjo@koldfront.dk (Adam Sjøgren):

> Does this hide the entire _thread_?

No, just the troll in question.

There is a possibility to score down on both the subject, and on that
particular thread (using references (at least I _think_ there is...)),
but I don't know how to do this without looking it up.

Re: Lower all threads started by a particular tro... er, person?

[Dave Abrahams]

on Sat Jan 09 2016, Steinar Bang <sb-AT-dod.no> wrote:

>>>>>> asjo@koldfront.dk (Adam Sjøgren):
>
>> Does this hide the entire _thread_?
>
> No, just the troll in question.

Yeah, the problem here is that other people jump on the thread and waste
my bandwidth dealing with the unproductive input and I really want to
down-score the whole thing.

> There is a possibility to score down on both the subject, 

Using the subject can lead to down-scoring too many articles.
In principle there should be no reason to do this based on subject,
since thread references are present in the message headers.

> and on that particular thread (using references (at least I _think_
> there is...)), but I don't know how to do this without looking it up.

I guess that would be `L r'.  But I want to find messages from a
certain author and then programmatically down-score everything that
refers to that message.  I suppose
[[info:gnus#Advanced%20Scoring][info:gnus#Advanced Scoring]] could be
used to write a sufficiently ugly rule, but I worry that it might take
too long to process.  I wonder how hard it would be to extend that stuff
to allow for arbitrary ancestry searching, e.g. with ^*

-Dave

Re: Lower all threads started by a particular tro... er, person?

[Lars Ingebrigtsen]

Dave Abrahams <dave@boostpro.com> writes:

> There are some groups I read where one person reliably starts
> unproductive threads that I'd like to ignore entirely.  Is there any way
> to use scoring to make that happen?

Yes, `L f' will do that, I think?

-- 
(domestic pets only, the antidote for overdose, milk.)
   bloggy blog: http://lars.ingebrigtsen.no

Re: Lower all threads started by a particular tro... er, person?

[Dave Abrahams]

on Fri Feb 05 2016, Lars Ingebrigtsen <larsi-AT-gnus.org> wrote:

> Dave Abrahams <dave@boostpro.com> writes:
>
>> There are some groups I read where one person reliably starts
>> unproductive threads that I'd like to ignore entirely.  Is there any way
>> to use scoring to make that happen?
>
> Yes, `L f' will do that, I think?

I don't *think* so.  I don't just want to ignore the sender; I want to
entirely ignore any *thread* that particular sender might start.  Does
`L f' do that?

-- 
-Dave

Re: Lower all threads started by a particular tro... er, person?

[Teemu Likonen]

[-- Attachment #1: Type: text/plain, Size: 1116 bytes --]

Dave Abrahams [2016-02-07 07:58:31-08] wrote:

> on Fri Feb 05 2016, Lars Ingebrigtsen <larsi-AT-gnus.org> wrote:
>> Yes, `L f' will do that, I think?
>
> I don't *think* so.  I don't just want to ignore the sender; I want to
> entirely ignore any *thread* that particular sender might start.  Does
> `L f' do that?

Yes. It scores all followups to an author. In practice it adds this kind
of entry to your score file:

    ("followup"
      ("author" -1000 nil s))

That, in turn, will cause Gnus to automatically create adaptive scoring
rule whenever that "author" sends a message. Those adaptive rules will
look like this:

    ("references"
      ("<message-id@author>" -1000 736000 s)
      ...)
  
The above message-id is from "author's" post and since it's a
"references" rule it scores all messages which include
<message-id@author> in their References field.

So. "L a" scores a person and "L f" scores all followups for that
person's messages.

-- 
/// Teemu Likonen   - .-..   <https://github.com/tlikonen> //
// PGP: 4E10 55DC 84E9 DFF6 13D7 8557 719D 69D3 2453 9450 ///

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 818 bytes --]

Re: Lower all threads started by a particular tro... er, person?

[Greg Troxel]

[-- Attachment #1: Type: text/plain, Size: 1253 bytes --]


Teemu Likonen <tlikonen@iki.fi> writes:

> Dave Abrahams [2016-02-07 07:58:31-08] wrote:
>
>> on Fri Feb 05 2016, Lars Ingebrigtsen <larsi-AT-gnus.org> wrote:
>>> Yes, `L f' will do that, I think?
>>
>> I don't *think* so.  I don't just want to ignore the sender; I want to
>> entirely ignore any *thread* that particular sender might start.  Does
>> `L f' do that?
>
> Yes. It scores all followups to an author. In practice it adds this kind
> of entry to your score file:
>
>     ("followup"
>       ("author" -1000 nil s))
>
> That, in turn, will cause Gnus to automatically create adaptive scoring
> rule whenever that "author" sends a message. Those adaptive rules will
> look like this:
>
>     ("references"
>       ("<message-id@author>" -1000 736000 s)
>       ...)
>   
> The above message-id is from "author's" post and since it's a
> "references" rule it scores all messages which include
> <message-id@author> in their References field.
>
> So. "L a" scores a person and "L f" scores all followups for that
> person's messages.

Wow - I should have read that before commenting.  Thanks for the clear
explanation.

But in general, it would seem useful to have more programmability
surrounding scoring.

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 180 bytes --]

Re: Lower all threads started by a particular tro... er, person?

[Greg Troxel]

[-- Attachment #1: Type: text/plain, Size: 867 bytes --]


Sort of related, I find the docs hard to follow, and I think this might
be because of the score lowering function reads characters rather than
being a command prefix.  I can see why it does this; it seems like they
are parameters, but it means the command docs don't explain the subcases:

  gnus-summary-lower-score is an interactive autoloaded compiled Lisp
  function in `gnus-score.el'.

  (gnus-summary-lower-score &optional SCORE SYMP)

  Make a score entry based on the current article.
  The user will be prompted for header to score on, match type,
  permanence, and the string to be used.  The numerical prefix will
  be used as SCORE.  A symbolic prefix of `a' (the SYMP parameter)
  says to use the `all.SCORE' file for the command instead of the
  current score file.

  [back]


Even reading the source, the behavior of followup is not entirely clear.

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 180 bytes --]

Re: Lower all threads started by a particular tro... er, person?

[Lars Ingebrigtsen]

Greg Troxel <gdt@lexort.com> writes:

> Even reading the source, the behavior of followup is not entirely clear.

But I think this is explained more in depth in the manual...

-- 
(domestic pets only, the antidote for overdose, milk.)
   bloggy blog: http://lars.ingebrigtsen.no

Re: Lower all threads started by a particular tro... er, person?

[Dave Abrahams]

on Sun Feb 07 2016, Teemu Likonen <tlikonen-AT-iki.fi> wrote:

> Dave Abrahams [2016-02-07 07:58:31-08] wrote:
>
>> on Fri Feb 05 2016, Lars Ingebrigtsen <larsi-AT-gnus.org> wrote:
>>> Yes, `L f' will do that, I think?
>>
>> I don't *think* so.  I don't just want to ignore the sender; I want to
>> entirely ignore any *thread* that particular sender might start.  Does
>> `L f' do that?
>
> Yes. It scores all followups to an author. In practice it adds this kind
> of entry to your score file:
>
>     ("followup"
>       ("author" -1000 nil s))
>
> That, in turn, will cause Gnus to automatically create adaptive scoring
> rule whenever that "author" sends a message. Those adaptive rules will
> look like this:
>
>     ("references"
>       ("<message-id@author>" -1000 736000 s)
>       ...)
>
> The above message-id is from "author's" post and since it's a
> "references" rule it scores all messages which include
> <message-id@author> in their References field.
>
> So. "L a" scores a person and "L f" scores all followups for that
> person's messages.

Okay; I have to turn on adaptive scoring for this to work, I suppose?
I'm a little wary of missing messages because of rules it infers from
previous actions.

-- 
-Dave

Re: Lower all threads started by a particular tro... er, person?

[Lars Ingebrigtsen]

Dave Abrahams <dave@boostpro.com> writes:

> Okay; I have to turn on adaptive scoring for this to work, I suppose?

No, `L f' uses the adaptive scoring machinery, but adaptive scoring
itself is not switched on.

-- 
(domestic pets only, the antidote for overdose, milk.)
   bloggy blog: http://lars.ingebrigtsen.no

Re: Lower all threads started by a particular tro... er, person?

[Greg Troxel]

[-- Attachment #1: Type: text/plain, Size: 1090 bytes --]


Dave Abrahams <dave@boostpro.com> writes:

> on Fri Feb 05 2016, Lars Ingebrigtsen <larsi-AT-gnus.org> wrote:
>
>> Dave Abrahams <dave@boostpro.com> writes:
>>
>>> There are some groups I read where one person reliably starts
>>> unproductive threads that I'd like to ignore entirely.  Is there any way
>>> to use scoring to make that happen?
>>
>> Yes, `L f' will do that, I think?
>
> I don't *think* so.  I don't just want to ignore the sender; I want to
> entirely ignore any *thread* that particular sender might start.  Does
> `L f' do that?

You are correct; L f should lower all messages with a matching followup
header.

I use Lsft, bound to 'k' to add a scoring rule for a subject (that will
expire).  But that's also now what you want.

I think to do what you want will require code.  I wonder about having a
new kind of scoring action, which is able to run code instead of just
assigning a value, and then to match on author, and if there is no
followup header add a temp rule to match that followup.  Or maybe
there's some easier way to just run a procedure on each message.


[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 180 bytes --]