 |
 |
At a former research university employer, I talked to a new high-level research data officer type person, whose team had spent months just trying to make a list of all (or even most) of the academic/research organizational units currently existing, and their hiearchical relationships. Before even getting to change management for the future. No such list or org chart even existed. On Thu, Oct 12, 2017 at 1:44 PM, Kyle Banerjee <[log in to unmask]> wrote: > On Wed, Oct 11, 2017 at 2:34 PM, Stephen Hearn <[log in to unmask]> wrote: > > > > > ... The complexity arises from the way organizations > > constantly alter their internal and external organizational structures > and > > relationships. That's not something that a short retrospective horizon > can > > dispense with, unless the intent is to describe only a snapshot of > > organizations. If the intent is to build a registry of lasting value, > then > > policies for coping with the changes organizations undergo and the > complex > > relationships those changes entail will need to be developed by the > > openPIIR initiative... > > > I totally get the motivation -- we struggle with this exact problem since > we continually need to demonstrate the impact of departments, institutes, > specific grants, etc. > > The trick is that enormous percentage of the data must be > provided/maintained manually and the staggering amount of missing/filthy > info is a problem that won't go away with time. > > Even in an ideal world where all the individuals, entities, and > relationships are tracked perfectly, there is still the issue of exposing > these identifiers in the already huge and growing number of forms of > significant activity. I believe that precision is a pipe dream and would > hope that modeling is simple enough to allow more accurate estimation of > error factors based on well-understood and transparent weaknesses. > Otherwise, the likely result is what I like to call, "Measure with > micrometer, mark with chalk, cut with axe...." > > kyle >