James McGovern is an industry thought leader whose focus is on the human aspects of technology around open source, SOA, software security, enterprise architecture and agile software development.
Thursday, September 27, 2007
Records Management, ECM and an Enterprise Architecture Perspective
Laurence Hart provides wonderful insight into the RM world and is an insightful read for others...
RM capabilities should be built into any ECM platform... final disposition needs to be a built-in part of the platform. I believe all, and I am open to correction here, ECM vendors charge separately for retention and RM functionality, if they even have it split into two products.
Any thoughts on whether disposition when it is more about destruction could leverage external encryption services? For example, if Documentum were deeply integrated with Active Directory and could leverage its certificate services, documents could be destroyed simply by destroying the keys required to decrypt them. Of course this begs two questions of the first being coming up with better integration between ECM and PKI and more importantly, the various ECM vendors thinking about integration with products other than what originates from their own company. Maybe they need to first understand that SOA is not the golden hammer for integration.
The other thought that I had is this feels like an opportunity for Alfresco and Nuxeo to take marketshare away from Documentum and Stellent. The separate SKU sales pattern feels fugly when it should be a feature of an existing product and not something distinct as it cannot standalone on its own.
Many people will still go with RM as tracks those paper records that James mentions and people still need CYA. However, it can lead an organization down a rabbit hole if they look at two products, RPS and RM, and pick the term Gartner uses to define the problem.
I wonder if Laurence believes the rat hole is in the products or in listening to Gartner? I wonder why the likes of Nick Patience of the 451 Group, Alan Pelz-Sharpe or other industry analysts haven't researched this aspect deeper?
He defined RM with the 20 year old definition focused around paper.
Actually, it wasn't my definition but one I learned from another architect. It did include Microfilm and COLD devices as well. The hard part of this conversation is that I understand how RM and ECM are complimentary, but I still have no sense as to what the gaps are? What features does Alchemy have that Documentum doesn't and vice versa? Jesse Wilkins describes retention policies which feel like they are more strongly typed metadata in ECM. Does this beg for an industry taxonomy of metadata? The issue at hand is that if retention is solely driven off the ECM system being authorative then you could get it wrong.
Imagine a scenario where you have a BPM system and an ECM system integrated where documents are in the ECM and processes are in the BPM. Shouldn't you be able to via some industry standard that is implemented in all products be able to say that all documents are disposed of when the business process reaches a certain state without fugly syncronization?