Re: Adaptive Scoring

Re: Adaptive Scoring

[Lars Magne Ingebrigtsen]

lal@max.miel.mot.com (Lal Samuel Varghese) writes:

> adaptive scoring seems to enhance score for articles
> with null subject lines. isn't this a bug?

Well... I don't know.  Is it?

-- 
(domestic pets only, the antidote for overdose, milk.)
  larsi@ifi.uio.no * Lars Ingebrigtsen

Adaptive scoring

[Dmitry Yaitskov]

Hi,

I did not turn adaptive scoring on (that is, not knowingly :), and
sure enough the `gnus-use-adaptive-scoring' is nil. Still, it does
take place - some messages are highlighted, and there are *.ADAPT
files in my News directory. What am I missing? Thanks.

(Using pgnus 0.97 under NT)

-- 
Cheers,
-Dima.

Re: Adaptive scoring

[Kai Großjohann]

Dmitry Yaitskov <dimas@home.com> writes:

> I did not turn adaptive scoring on (that is, not knowingly :), and
> sure enough the `gnus-use-adaptive-scoring' is nil. Still, it does
> take place - some messages are highlighted, and there are *.ADAPT
> files in my News directory. What am I missing? Thanks.

You can type `V C' to customize the scoring and turn on adaptive
scoring there.  Dunno if that's what you did.

kai
-- 
Life is hard and then you die.

Re: Adaptive scoring

[Yair Friedman (Jerusalem)]

Dmitry Yaitskov <dimas@home.com> writes:

> Hi,
> 
> I did not turn adaptive scoring on (that is, not knowingly :), and
> sure enough the `gnus-use-adaptive-scoring' is nil. Still, it does
> take place - some messages are highlighted, and there are *.ADAPT
> files in my News directory. What am I missing? Thanks.

Scoring by followup also create *.ADAPT files.  In this case those files
will only contain references clauses.

--
Yair Friedman.

Re: Adaptive scoring

[Dmitry Yaitskov]

"Yair Friedman (Jerusalem)" <YAIRFR@Amdocs.com> wrote:

> Dmitry Yaitskov <dimas@home.com> writes:
> 
> > Hi,
> > 
> > I did not turn adaptive scoring on (that is, not knowingly :), and
> > sure enough the `gnus-use-adaptive-scoring' is nil. Still, it does
> > take place - some messages are highlighted, and there are *.ADAPT
> > files in my News directory. What am I missing? Thanks.
> 
> Scoring by followup also create *.ADAPT files.  In this case those files
> will only contain references clauses.

I am not sure what scoring by followup is, so if it's something that
is turned on explicitly - I didn't do it. And (regarding Kai's
suggestion) I'm sure I didn't turn adaptive scoring on a per group
basis by mistake, since *all* groups that I read/post have *.ADAPT
files... Thanks for your suggestions guys but... any other ideas
please?

-- 
Cheers,
-Dima.

Re: Adaptive scoring

[Yair Friedman (Jerusalem)]

Dmitry Yaitskov <dimas@home.com> writes:

> "Yair Friedman (Jerusalem)" <YAIRFR@Amdocs.com> wrote:
> 
> > Dmitry Yaitskov <dimas@home.com> writes:
> > 
> > > Hi,
> > > 
> > > I did not turn adaptive scoring on (that is, not knowingly :), and
> > > sure enough the `gnus-use-adaptive-scoring' is nil. Still, it does
> > > take place - some messages are highlighted, and there are *.ADAPT
> > > files in my News directory. What am I missing? Thanks.
> > 
> > Scoring by followup also create *.ADAPT files.  In this case those files
> > will only contain references clauses.
> 
> I am not sure what scoring by followup is, so if it's something that
> is turned on explicitly - I didn't do it.

It happens if you Lf* ot If*.  The *.SCORE files show followup clauses as
expexted but also *.ADAPT are created with references clauses.  What do
you have in you *.ADAPT files?
--
Yair Friedman.

Re: Adaptive scoring

[Dmitry Yaitskov]

"Yair Friedman (Jerusalem)" <YAIRFR@Amdocs.com> wrote:

> Dmitry Yaitskov <dimas@home.com> writes:
> 
> > "Yair Friedman (Jerusalem)" <YAIRFR@Amdocs.com> wrote:
> > 
> > > Dmitry Yaitskov <dimas@home.com> writes:
> > > 
> > > > Hi,
> > > > 
> > > > I did not turn adaptive scoring on (that is, not knowingly :), and
> > > > sure enough the `gnus-use-adaptive-scoring' is nil. Still, it does
> > > > take place - some messages are highlighted, and there are *.ADAPT
> > > > files in my News directory. What am I missing? Thanks.
> > > 
> > > Scoring by followup also create *.ADAPT files.  In this case those files
> > > will only contain references clauses.
> > 
> > I am not sure what scoring by followup is, so if it's something that
> > is turned on explicitly - I didn't do it.
> 
> It happens if you Lf* ot If*.  The *.SCORE files show followup clauses as
> expexted but also *.ADAPT are created with references clauses.  What do
> you have in you *.ADAPT files?

Here is the first bytes of one:

(("thread"
("<Pine.SOL.3.96.990504121948.7026D-100000@ux8.cso.uiuc.edu>" nil
729878 s) ("<372f278f.14345258@news.megsinet.net>" nil 729878 s)
("<Pine.SOL.3.96.990503214434.15321A-100000@ux11.cso.uiuc.edu>" nil
729878 s)

... and so on (they do not have any line breaks inside).

I have no idea what either Lf* or If* is or does so I suppose I don't
:)

-- 
Cheers,
-Dima.

Re: Adaptive scoring

[Yair Friedman (Jerusalem)]

Dmitry Yaitskov <dimas@home.com> writes:

> "Yair Friedman (Jerusalem)" <YAIRFR@Amdocs.com> wrote:
> 
> > 
> > It happens if you Lf* ot If*.  The *.SCORE files show followup clauses as
> > expexted but also *.ADAPT are created with references clauses.  What do
> > you have in you *.ADAPT files?
> 
> Here is the first bytes of one:
> 
> (("thread"
> ("<Pine.SOL.3.96.990504121948.7026D-100000@ux8.cso.uiuc.edu>" nil
> 729878 s) ("<372f278f.14345258@news.megsinet.net>" nil 729878 s)
> ("<Pine.SOL.3.96.990503214434.15321A-100000@ux11.cso.uiuc.edu>" nil
> 729878 s)
> 
> ... and so on (they do not have any line breaks inside).
> 
> I have no idea what either Lf* or If* is or does so I suppose I don't
> :)

It seems that you score by thread Pressing `L'
(gnus-summary-lower-score) or 'I' (gnus-summary-increase-score) and then
`t' to permanently score down a thread.  Scoring down (or up) a thread
as well as scoring down by followup (L and then f) uses the adaptive
scoring machanizm.  Maybe some other scoring options uses the adaptice
machanism to do this.

All of this isn't documented in the manual so it seems as a
documentation bug.
--
Yair Friedman.

Re: Adaptive scoring

[Dmitry Yaitskov]

"Yair Friedman (Jerusalem)" <YAIRFR@Amdocs.com> wrote:

> Dmitry Yaitskov <dimas@home.com> writes:
> 
> > "Yair Friedman (Jerusalem)" <YAIRFR@Amdocs.com> wrote:
> > 
> > > 
> > > It happens if you Lf* ot If*.  The *.SCORE files show followup clauses as
> > > expexted but also *.ADAPT are created with references clauses.  What do
> > > you have in you *.ADAPT files?
> > 
> > Here is the first bytes of one:
> > 
> > (("thread"
> > ("<Pine.SOL.3.96.990504121948.7026D-100000@ux8.cso.uiuc.edu>" nil
> > 729878 s) ("<372f278f.14345258@news.megsinet.net>" nil 729878 s)
> > ("<Pine.SOL.3.96.990503214434.15321A-100000@ux11.cso.uiuc.edu>" nil
> > 729878 s)
> > 
> > ... and so on (they do not have any line breaks inside).
> > 
> > I have no idea what either Lf* or If* is or does so I suppose I don't
> > :)
> 
> It seems that you score by thread Pressing `L'
> (gnus-summary-lower-score) or 'I' (gnus-summary-increase-score) and then
> `t' to permanently score down a thread.  Scoring down (or up) a thread
> as well as scoring down by followup (L and then f) uses the adaptive
> scoring machanizm.  Maybe some other scoring options uses the adaptice
> machanism to do this.

Thanks. This must be it. Although I do not score threads, I *do* score
down certain posters and subjects. I never realized it was ending up
in *.ADAPT files, although I never actually gave it much thought. What
I do not understand though is - some messages in summary buffers are
shown in bold, with '+' sign to the left. I could never figure out on
what principle those messages were selected. I assumed that that was
the result of that adaptive scoring which I didn't know why was going
on. If it is - why? I never raised any scores :) Or am I missing
something obvious again?

> All of this isn't documented in the manual so it seems as a
> documentation bug.

-- 
Cheers,
-Dima.

Re: Adaptive scoring

[Lars Magne Ingebrigtsen]

"Yair Friedman (Jerusalem)" <YAIRFR@Amdocs.com> writes:

> It seems that you score by thread Pressing `L'
> (gnus-summary-lower-score) or 'I' (gnus-summary-increase-score) and then
> `t' to permanently score down a thread.  Scoring down (or up) a thread
> as well as scoring down by followup (L and then f) uses the adaptive
> scoring machanizm.  Maybe some other scoring options uses the adaptice
> machanism to do this.
> 
> All of this isn't documented in the manual so it seems as a
> documentation bug.

It's documented in the "Score File Format" section...

-- 
(domestic pets only, the antidote for overdose, milk.)
   larsi@gnus.org * Lars Magne Ingebrigtsen