After an evaluation, GNOME has moved from Bugzilla to GitLab. Learn more about GitLab.
No new issues can be reported in GNOME Bugzilla anymore.
To report an issue in a GNOME project, go to GNOME GitLab.
Do not go to GNOME Gitlab for: Bluefish, Doxygen, GnuCash, GStreamer, java-gnome, LDTP, NetworkManager, Tomboy.
Bug 745232 - scoring system of contacts should take frequency into account
scoring system of contacts should take frequency into account
Status: RESOLVED OBSOLETE
Product: geary
Classification: Other
Component: engine
master
Other Linux
: Normal normal
: ---
Assigned To: Geary Maintainers
Geary Maintainers
: 714381 (view as bug list)
Depends on:
Blocks:
 
 
Reported: 2015-02-26 17:34 UTC by Federico Bruni
Modified: 2019-03-17 04:53 UTC
See Also:
GNOME target: ---
GNOME version: ---



Description Federico Bruni 2015-02-26 17:34:08 UTC
Discussed on list:
https://mail.gnome.org/archives/geary-list/2015-February/msg00034.html

Geary does have a scoring system, but it's based on the role a given email address plays.  Highest ranked are those used in emails you've sent, then come those where you were a To: recipients, followed by those where you are a CC: or aren't listed.  The full ranking can be seen in src/engine/api/geary-contact-importance.vala.  If an email address appears in multiple roles, the highest one wins.  The contact completion uses these rankings when sorting.

However, the scoring results may not produce what a user expects, since an email address you wrote to once counts as much as an email address you write to every week. See link above for an example.
Comment 1 Robert Schroll 2015-02-26 17:48:55 UTC
I think this is a very reasonable idea, but there are a number of details we'd need to work out:

- Should this be based on the frequency the address shows up or the time since you last used it?  Or should it be a hybrid frequency measure that weights recent use highest?

- Should this metric include all appearances of the addresses, or only those in the current most important role?

- Should this be used only within a given importance, or should we come up with a metric that combines both importance and use?

- Do we want to bin the frequency measure, and sort alphabetically within the bins?  Otherwise, the order will appear random.
Comment 2 Jim Nelson 2015-02-26 19:59:12 UTC
An additional consideration would be to detect (via standard and well-known headers) if the message is from a bot or automated system and deprioritize its rating, or even discard it entirely.  Many of the emails Federico listed on the mailing list could reasonably be dropped from the contact table.  They are addresses you would reply to, but would almost never send a new email to.
Comment 3 Federico Bruni 2015-05-31 09:19:24 UTC
I don't know if this should be not recommended, however I decided to manually change the score of some email addresses that were bugging me.

I used sqlitebrowser (Browse data, order by email and optionally filter the highest_importance field).
Comment 4 Federico Bruni 2015-09-09 20:24:01 UTC
Well, above method doesn't work well with mailing lists, as <mailing-list-name>-bounces@ addresses quickly jump at top of the score after a while.
Comment 5 Federico Bruni 2018-01-12 17:38:46 UTC
*** Bug 714381 has been marked as a duplicate of this bug. ***
Comment 6 Michael Gratton 2019-03-17 04:53:50 UTC
Closing in favour of: https://gitlab.gnome.org/GNOME/geary/issues/318