Tracking case files across organizations

Petter Reinholdtsen pere at hungry.com
Tue May 16 12:44:07 CEST 2017


It would be great if it was possible to track a case in the goverment
across organisational units, to get a more complete picture of the
documents involved.  This would also make it possible to discover holes
in the mail journal, ie documents that are recorded as sent or received
by one organisation, but not recorded on the other end.  I believe such
tool would be valuable both for the decision makes in government,
journalists and the public at large.

I've been thinking a bit about how to do this, and would like to sketch
a proposal to see what the rest of you think.

This mostly involve what the Noark 5 specification call 'journalpost',
which is a entry in the public mail journal of a goverment office.  A
letter or email sent can have many recipients.  But there will be only
one sender.  If all entries have an unique ID, we can track posts across
organizations by adding a 'sender document ID' to each recipient
journalpost.  This allow us to both find all other recipients of a
document, and the sender, and to look up all cases where the message is
filed.  With such 'sender reference' in place, it is possible to start
with a given case file in one organisation, look up all entires in that
file, look up their 'sender reference' in other organisations and find
all case files where the same message was filed.  This will give all
case files where the message was filed, across all organisations.

The hard task is of course to figure out what the 'sender reference'
should be.  The general idea for this is to look at sender and recipient
name for a given message, and try to map the name to other organisations
where we got 'journalpost' entries.  When we have such candidates, we
can look at entries with matching document/message dates, to see if
there are good candaites with similar or matching document/message
titles.  The entries will typically not have identical titles, and the
document date can be mistyped, so there is no guarantee for a perfect
match.  I am unsure how hard this will be, and how many document matches
will be missing with this approach.  But I am sure we will be able to
find at least some matches and thus track some cases across
organisations this way.

A key requirement for this is that all journalpost entries are given
unique IDs across all organisations.  Not quite sure how to do this, but
at least for the entries in oep.no, one can use the oep ID (number).  If
we import entries into nikita first, we can use the systemID value
instead.

Thoughts?

-- 
Happy hacking
Petter Reinholdtsen


More information about the nikita-noark mailing list