adaptive expiry? feature request?

adaptive expiry? feature request?

[Eric S Fraga]

Hello,

I used to use splitting to move all of my incoming email to various
groups.  Each group would then have different expiration policies (and
targets).  However, I hated having to switch between groups so I moved
to using scoring instead and having many of my emails (non-list ones,
say) sit in one group.  Scoring (even non-adaptive) is brilliant [1] and
I would have a hard time not using it any more!

However, this means that all of these emails are expired using the same
policy.

What I would like is to have an expiry policy that is adaptive, in the
same way as scoring is.  In fact, I would love to have an expiry time
that is a function of the score (meaning that active threads, for
instance, would not expire for quite a long time) and I guess other
message attributes.  In other words, a more general version of the
nnmail-expiry-wait-function and nnmail-fancy-expiry-targets functions
would be quite nice to have!

Is such a thing possible?  I've searched gnus variables and done a web
search of sorts but found nothing.  I think it should be possible (maybe
even straightforward?) to implement but my elisp is not good enough to
even think about doing this.

Thanks,
eric


Footnotes: 
[1]  does any other MUA have scoring?  Not that I intend to change :-)
Just curious.

-- 
: Eric S Fraga, GnuPG: 0xC89193D8FFFCF67D
: in Emacs 24.3.50.1 + Ma Gnus v0.6 + evil 1.0-dev
: BBDB version 3.02 ($Date: 2013/02/16 14:37:17 $)

Re: adaptive expiry? feature request?

[Ted Zlatanov]

On Tue, 26 Feb 2013 18:41:48 +0000 Eric S Fraga <e.fraga@ucl.ac.uk> wrote: 

ESF> What I would like is to have an expiry policy that is adaptive, in the
ESF> same way as scoring is.  In fact, I would love to have an expiry time
ESF> that is a function of the score (meaning that active threads, for
ESF> instance, would not expire for quite a long time) and I guess other
ESF> message attributes.  In other words, a more general version of the
ESF> nnmail-expiry-wait-function and nnmail-fancy-expiry-targets functions
ESF> would be quite nice to have!

ESF> Is such a thing possible?  I've searched gnus variables and done a web
ESF> search of sorts but found nothing.  I think it should be possible (maybe
ESF> even straightforward?) to implement but my elisp is not good enough to
ESF> even think about doing this.

I think atomic expiry it's a pretty hard assumption to break in Gnus,
but Lars or others may know better.

Ted

Re: adaptive expiry? feature request?

[Lars Magne Ingebrigtsen]

Eric S Fraga <e.fraga@ucl.ac.uk> writes:

> What I would like is to have an expiry policy that is adaptive, in the
> same way as scoring is.  In fact, I would love to have an expiry time
> that is a function of the score (meaning that active threads, for
> instance, would not expire for quite a long time) and I guess other
> message attributes.  In other words, a more general version of the
> nnmail-expiry-wait-function and nnmail-fancy-expiry-targets functions
> would be quite nice to have!
>
> Is such a thing possible?  I've searched gnus variables and done a web
> search of sorts but found nothing.  I think it should be possible (maybe
> even straightforward?) to implement but my elisp is not good enough to
> even think about doing this.

I don't think there's anything in Gnus for doing that fine-grained
expiry, unfortunately.

If you look at (for instance) `nnml-request-expire-articles' (or
whatever mail backend you use), `nnmail-expired-article-p' is called to
determine whether an article is "old".  You'd have to wrap that call
with something that sets nnmail-expiry-wait with something that depends
on the score of the article.

> Footnotes: 
> [1]  does any other MUA have scoring?  Not that I intend to change :-)
> Just curious.

At the time Gnus introduced scoring (1993?) there were definitely a lot
of discussion about how to deal with the amazingly huge information
overload from Usenet.  >"?  So I don't think the scoring idea was
original to Gnus.  I think we developed the adaptive scoring thing at
Gnus, though.

-- 
(domestic pets only, the antidote for overdose, milk.)
  No Gnus T-Shirt for sale: http://ingebrigtsen.no/no.php

Re: adaptive expiry? feature request?

[Eric S Fraga]

Lars Magne Ingebrigtsen <larsi@gnus.org> writes:

[...]

> I don't think there's anything in Gnus for doing that fine-grained
> expiry, unfortunately.

Thanks for looking at this.  I have changed my approach so this is less
necessary now.  I'm using adaptive scoring and more ruthless expiry of
uninteresting emails instead.

Thanks again,
eric

-- 
: Eric S Fraga, GnuPG: 0xFFFCF67D
: in Emacs 24.3.50.1 + Ma Gnus v0.8 + evil 1.0.6
: BBDB version 3.02 ($Date: 2013/07/07 07:16:09 $)