Hi,
It was Martin Klapetek who said at the right time 20.07.2011 17:02
the following words:
On Wed, Jul 20, 2011 at 15:25, Leo Sauermann <[email protected]> wrote:
Hi Martin,
I focus on the second part...below
It was Martin Klapetek who said at the right time
20.07.2011 14:59 the following words:
On Wed, Jul 20, 2011 at 14:13, Sebastian Trueg <[email protected]> wrote:
Hi Martin,
On 07/19/2011 07:46 PM, Martin Klapetek wrote:
> 2) How to pick "the right data"?...
...
The way it was designed is that you end up with one
pimo:Person to represent claudia:DirkHagemann - "The one
and only Dirk Hagemann which I know and whom I contact
using various ways".
Then there is the "grounding occurrence" for Dirk
Hagemann, which usually should map to the address book
entry you would naturally open when clicking on Dirk -
think of it as the default data object to open when the
user wants to do something with Dirk.
Then there are many occurrences of Dirk all over the
place.
Yes, I understand that. But you can have more than one
address book entry for the Person - for example you have "home
postal address" and "work postal address", similarly a
"personal email address" and a "work email address". These can
be all stored as a one nco:personContact per email/postal
address or all in one single nco:personcontact. But that's not
really the point. When you click on Dirk, you'd like to see
his home postal address or his personal email address (this
could also be activity triggerable), in other words, you want
to set some email address as default. My question is, how can
I tag a particular email/postal/phone/whatever as a
"preferred/default", so when you want to quickly write an
email, you just select Dirk and email him to his default
address.
Hope it's clearer now :)
For home/work distinguishing you would have one Contact
(PersonContact) and connect multiple ContactMedia entries to the
single PersonContact. The Role is a superclass of PersonContact, so
all ContactMedia apply.
Dirk->hasPostalAddress->dp1.
Dirk->hasEmailAddress->de1.
Dirk->hasEmailAddress->de2.
you would then comment on de1 and de2 if they are primary or not.
For this, there is no property yet, we wanted to use
nco:contactMediumComment and some text, but the standardization
ended at that point.
If you need that, please create a ticket on oscaf.sourceforge.net as
I am not on top of that discussion
This is all NOT on the level of PIMO or grounding/occurrence.
In the PIMO description [1] (btw, thanks for using it,
this helps a lot and I need not to repeat), on page 10, in
Section 6.5, there is an example explaining what it means
in practice.
I would not go so far to add pimo:occurrences to every
email that Claudia and Dirk sent between each other. But
adding multiple address book entries is fine.
Sticking with pimo:occurrence and pimo:groundingOccurrence
should be fine, I doubt that real-world systems will
interpret pimo:referencingOccurrence soon (a Facebook URI
would be something like this).
So what should I use pimo:occurence and
pimo:groundOccurence for? I'm currently using groundOccurence
for all nco:personContacts.
use one pimo:groundingOccurrence to point to the main address book
entry (PersonContact). Use pimo:occurrence for additional address
book entries from other address books or secondary web accounts
(facebook, etc)
hth
Leo
Btw. when reading the PIMO guide [3], it says
on page 31 to copy all identifiers. Do I
understand it correctly that it should copy all
identifiers from new NCO:PersonContact into
PIMO:Person (and should I do that too)? This seems
like an unnecessary data duplication.
identifiers:
Copying the email addresses and other identifiers to the
pimo:Person will help you identifying arbitrary data.
Given you have some new data coming into the system, you
may want to check if its already known.
I have currently some checks in place, but I'll add those
from the guide as well.
("is this "Dirk" guy
someone we know?). This is the "Check identifiers" step In
Section 11.1 on page 30, where you check if you find a
Pimo:Person that represents the resource you are working
with.
This is why it would be good to copy all identifiers over
to the pimo:Person, to speed up lookups later.
Ok, I'll add that as well.
its data duplication, but it doesn't matter. The key point
here is: you want to be veeery quick when identifying if
any new data does mean something to you. And optimizing
this read-accesses is always good. Say you read 100 emails
a day, roughly 10.000 a year. Quickly checking them
against the 1000 contacts you have is nicer if the
identifying information is all in one place (that would be
the pimo:Persons).
Sounds reasonable :)
At least, thats the theory.
We'll see how it will stand in practice :D
Thanks,
Marty K.
--
Leo Sauermann, Dr.
CEO and Founder
mail: [email protected]
mobile: +43 6991 gnowsis
Where your things come together,
Join: http://www.gnowsis.com/about/content/newsletter
Follow: http://twitter.com/Refinder
Like: http://www.facebook.com/Refinder
Learn: http://www.gnowsis.com/about/blog
Try: http://www.getrefinder.com/accounts/register/
____________________________________________________
|
