Some brief comments on semantic web technology and synthetic intelligence networks:
Management Problems:
Advanced technology problems are always management problems. Very few US managers have the technical, operations and financial experience to manage advanced technology.
Semantic Web technology is not different from 1973-75 Oracle database technology, 1982-85 Intel/Microsoft PC, 1996-98 ERP SAP/BAAN revolution or any other fundamental information technical revolution. It is the front line of the information technology revolution. This is the same; only faster; and more dangerous for those who lag.
Concerns:
DOD and IC partners/clients don't have the technical background or management experience to understand Semantic Web Technology and Synthetic Intelligence Networks; or understand the consequences of new technology.Best bet is a small “with it” shop like SOJICC where new ideas can be tested outside the beltway, and the proof is in the results that cannot be denied.
Too Much Money Can Be Negative:
The DOD and IC are so well funded, so isolated from commercial technology, and so politically protected; only a spectacular failure like 9/11 will force any changes. And, while the 9/11 commission report touches on information architecture; is misses the core problems and solutions. UBL may replace Bill Gates and Larry Ellison as the poster boy for IC technology change.
Painful Motivation:
At some point it will become painfully clear to the public that current static library architecture (Oracle) and simplistic statistical tools (Goggle), and armies of amateur translators do not solve their problems. The tidal wave of unstructured text data (particularly non-English language) is completely overwhelming the DOD and IC. And more 400% marked up GS-9 /96-Bravo bodies make the problem worse.
Window Dressing:
That fact still does not prevent expensive window dressing exercises like Trilogy program; $500M in patch code that does not address the core problem. The fundamental information architecture does not support the mission; and does not deal with the organizational problem: floods of unstructured text data in isolated stovepipes being addresses by unrelated tools operated by manual mouse click.
Technology Leadership:
In the meantime, some intelligent low level managers will be able to quietly deploy and test these systems, and take a huge leadership position. ( The first word processor, digital database, on-line access and simple search tool in State Department/EA in 1979-80; and it took only 20 years for these to become standards.)PREDICTION: If SOJICC gets semantic web right, it will still take rest of government 20 years to catch up, absent major Congressional interest.
Congress and GAO as Catalysts for Change:
Fortunately; there are enough folks in business community and Congress now that have seen these technologies and understand the current failures. The next round of Congressional inquiry may dig deeper, identify the real issues; and force the management changes required to make any progress. GAO inquiries may become the incentive to reform. But many heads will roll before any changes are made.
Briefing points on semantic web technilogy:
1. History: Semantic Web technology has been in development for 30 years; the challenge was the processing power to make it feasible and the years of hard work to polish the code. Both problems are solved; and powerful new Intel HT (hyperthread) microprocessor technology with Microsoft NT and XPPro OS operating systems now puts the CPU power required in small servers, desktops, and even laptops.
=> We can provide synthetic intelligence laptops that "think" in multiple languages and 100GB of data.
2. Technology: The core technology for Semantic Web architecture and applications is based on using the content of data to create the context; and use the context to understand the content. In effect; the more data you add; the more detail you have. The net result is a huge, throbbing network of inter-linked concepts; like a schematic logic. The key point is simply this: When you add new data; does it automatically improve your knowledge of old data; and automatically notify you? Or is it just another isolated book on the static shelf.
=> The technical demo system can illustrate this very clearly to nontechnical senior executives.
3. Applications: Semantic Web products have been developed and deployed commercially for 5 years. IBM has been at this longest; but has not yet gone commercial. (The issues are IBM politics; not technical). The Folks on Rt32 were the pioneers in federal space; and have developed, deployed and patented their semantic web systems. These programs have also led to several spin off development efforts. And there is a growing tribe of start up firms launching programs now.
=> We are testing a system for Homeland Security first responders funded by Justice. More than 300 commercial clients use our technology over three years.
4. Standards: The Semantic Web industry has a technical standards body led by Tim Berners-Lee, W3C Director and inventor of the World Wide Web. ( The technical issues and common standards are in place now. This is critical; because US corporations and federal agencies are extremely slow and timid in adopting new technology. The problem is that there are now no federal RFI/RPF common text in place; so that even if federal managers were willing to take a leadership role; they would have no idea how to ask for help or qualify vendors or technologies.
=> We wrote the FRI/RFI/Contract terms, qualification process, and recommended an "Industry Expert" for the Justice-funded program.
5. Sources: There are no federal qualified vendors of GAO approved Semantic Web products now. SAIC and MITRE have a huge advantage in taking over this market because they understand the technology; and directly advise their client decisions. (MITRE has several "academic" experts; but are wound in circles in raising questions to justify future research budgets. SAIC has the most knowledgeable applications group; but is forced to support the present SAIC static library architecture; and has no technology ownership position to support a semantic web program.) STEELE: get SRA and General Dynamics in to talk about this with MITRE and SOJICC, on 7 October?
=>Federal manager need to follow the 1981-84 IT model; when Intel/Microsoft PC-AT systems with simple spread sheets and word processors penetrated the federal market under the radar of the mainframe IT management and OMB bean counters. "Just Do It…" [MITRE could actually be part of the solution!!]
6. Next Steps: The problem is that the information architecture technology is now very far ahead of organization ability to understand and digest it. And, the DOD and IC are falling further behind the technology edge. Increased budgets only produce more Trilogy programs; legacy patch code; not new architecture. IT organizations are still struggling to get their email and security systems more reliable. This is not a budget or IT people issue; There are simply not enough senior management talent to deal with core issues of migration to the information architecture.
=>Your clients do not need new products; They need an experienced technology management team that understands how to handle the migration to the next generation of information architecture.