About eight years ago, I toured the country with a number of seminars conducted by the Delphi Group on corporate portals. Portals were quickly becoming an option for corporations (including law firms) and the market was raging with a number of portal vendors. Most of these are not around today as consolidation and evolution of the concept saw mergers in the industry, and many early entrants didn’t make it through the first few laps. But I was impressed then, and still am today, with the basic concept Delphi had developed and with their framework for thinking about and building portals.
Delphi defined corporate portals as … “a single point of access for the pooling, interaction and distribution of organizational knowledge.” I always liked this definition, as it placed emphasis on ‘knowledge’ as the key goal.
Delphi proposed a model / framework for corporate portals built largely on developments with the public portals of the day (think AOL, Yahoo, and Netscape). The seven basic components of the framework / model included:
- a Taxonomy Engine:- to provide organization for knowledge artefacts and pointers for source materials. Indeed, early efforts were pointing to the possibility of sophisticated automated taxonomy engines to complement manual efforts.
- a Metadata Engine:- using either manual or inferred tag or markup data (think SGML/XML). Alternatively, they envisioned DBMS links to unstructured data.
- a Search Engine:- with full text search, linguistics, heuristics and semantics. And with links to the Taxonomy engine.
- an Integration engine:- they envisioned a tool (engine) to link the metadata and search engines.
- a Publishing engine:- that would help organize information from various sources and present it in various sources (web pages, as PDF documents, etc.)
- a Presentation engine:- dealing with cognitive presentation and allowing for interaction and learning from user experience.
- a Learning engine:- that would allow for system self-adjustments to fit apparent work needs of users and to provide a feedback loop between user interaction and system content and processes.
In addition, they also talked about desirable components and features such as process support.
Looking back, I have some observations and some questions. I think a lot of these concepts still hold today; but I do not think many firms have realized this full vision. How many have work flow as an integral part of their engines today? Tags have moved beyond the domain of SGML and XML to the world of folksonomies — but how many firms have leveraged this in their portals? Have we really figured out the taxonomy issues (or, as I prefer, the knowledge organization frameworks)? Should integrations engines be more like the components we see in other industries — linking applications together (see healthcare and Health Level 7 or HL7).
There are a number of components that have emerged over the years that were not envisioned at the time. Take for example the impact of RSS on the web today. Content management had not yet reared its head as a key application — but it is a significantly important component today. The concept of community, although well entrenched in the early web and the market at the time, did not figure prominently in the architecture; whereas, today, driven by our understanding and the success of Web 2.0 applications, this is a crucial component for any good portal.
Search engines have come a long way in these last few years. And although adoption has started to take hold with law firms, I doubt the full vision has been realized, as I said earlier. So I would be curious to know which firms are closest to this ideal portal today and what your thoughts are on how this model could be refreshed given today’s environment and array of tools.