On Wed, 07 Dec 2011 00:03:19 +0200, Tomi Ollila <tomi.ollila@iki.fi> wrote:
> On Tue, 06 Dec 2011 08:45:49 +1100, Bart Bunting <bart@ursys.com.au> wrote:
> > Hi,

... 

> 
> > The only other idea I have, which is only half formed, is that it would
> > be nice to prioritize emails that are more important than others.
> > 
> > I'm not sure exactly how this would work but something like:
> > - Addresses I have actualy sent email to rather than jjust received from
> > get a high priority in the result.
> > - Addresses that have sent me email directly rather than just to a list
> > get next priority.
> > - All other emails after that.
> > 
> > Also some sort of weighting within the groups to do with frequency of
> > emails sent or something.
> > 
> > Does that make sense?  Would it be hard to implement?
> 
> Neat idea. To implement the feature into program is not too hard. 
> Somehow determine weight for each email address, store that into 
> hash (key address, value weight). 
> Then, when writing addresses file, sort first by weight and then 
> ascii order. The hash is also stored to disk (using tie()) and used 
> when address file updated.
> 
> Now, how to determine good (generic) rules for weights (and maybe
> some user-defined rules as well) is the question.

The latest version (1.3) now stores addresses in 'newest-first'
order to the "address file" -- the fgrep printing matches will
therefore also output the lines in this order.

For me this seems to work pretty nicely.

This version (1.3) relies notmuch show ... to output emails in
this order (which it seems to do now). If this changes I'll 
have to check how to "fix" my implementation...

The program is located at the same address:

http://www.iki.fi/too/nottoomuch/nottoomuch-addresses.sh

sha1sum is f9289ed4bf9034d99e390aaf3d5fcac3d815ea20 and doc page is at
 
http://www.iki.fi/too/nottoomuch/nottoomuch-addresses/


> > Cheers
> > 
> > Bart
> 
> Thanks,
> 
> Tomi

Tomi