> * Ted Zlatanov <gmm@...> [2011-04-21 13:21:42 -0500]:
>
> On Tue, 12 Apr 2011 08:32:37 -0500 Ted Zlatanov <tzz@...> wrote:
>
> TZ> On Mon, 11 Apr 2011 18:22:32 -0400 Sam Steingold <sds@...> wrote:
>>>> * Ted Zlatanov <gmm@...> [2011-04-11 15:52:06 -0500]:
>>>>
>>>> With Gnus at least, we could have a last-seen field that's updated
>>>> whenever you see a message from that person.
>
> Sam> Are you suggesting a last-seen field in addition to creation-date and
> Sam> timestamp? Or maybe a last-seen field for each e-mail address?
>
> TZ> I didn't think timestamp was updated every time you see a message from
> TZ> someone, just when the record is modified. This one would explicitly
> TZ> say "I saw a message from you on YYYY-MM-DD" and leave timestamp
> TZ> unchanged because it's not a true data change.
>
> Sam, are you in silent agreement or will this feature not help you?
a third timestamp field is fine by me, but it will not solve the problem
(which as I say below does not seem to have an automatic solution).
note however that this should be optional. I keep my .bbdb under
mercurial and I do not want a record to change whenever I see the person
in a newsgroup.
>>>> Then you can filter all the records with last-seen over 2 years ago,
>>>> for instance.
>
> Sam> This is no good.
> Sam> I might be seeing messages from the person in some newsgroup every day
> Sam> but still not want him in the diary or the printout.
> Sam> I think both diary and printout are too "personal" for a
> Sam> "catch-all" solution.
>
> TZ> You can apply more filters, this is just the "have I seen this person
> TZ> lately" filter.
>
> TZ> I think it may be nice to have an integer friend-level field that could,
> TZ> for instance, range from 10000 (has my DNA) to 0 (eh) and down to -10000
> TZ> (would kill on sight). This could be used to score discussion threads
> TZ> in Gnus, for instance, and to do the "I want to know about this person"
> TZ> filter you want.
>
> Same question.
I think this goes in the wrong direction.
While having this "friend-level" field might be useful and interesting
for some purposes, it is largely orthogonal to the problem I am talking
about.
I want to be able to _permanently_ mark some records as
- diary (i.e., the dates from them go to the emacs diary display)
- print (i.e., the records printed out so that I will carry them with me)
- fridge (i.e., a _very_ short list to be printed out and attached to
the fridge for the kids and/or babysitter call in an emergency)
- gmail (i.e., the records which I want to export as csv to be imported
into gmail contacts)
as you can see, one can hardly hope for an automatic catch-all solution,
and even your "friend-level" is not good enough because these sets are
not guaranteed to be linearly sorted by inclusion.
I think the right approach is to add a field called "export-class" with
values being a set of symbols and each "export operation" (e.g., print,
diary &c) should operate on the records of the specific class (`print',
`diary' &c) which should be the value of an option, e.g.,
`bbdb-print-class', `bbdb-diary-class', so that I can bind them around a
call to an export function.
--
Sam Steingold (http://sds.podval.org/) on CentOS release 5.5 (Final) X 11.0.60900031
http://jihadwatch.org http://openvotingconsortium.org
http://honestreporting.com http://www.PetitionOnline.com/tap12009/
Bus error -- please leave by the rear door.
|