Ontolog Forum


Where the action and the benefit is, as I see it, is in having something that existing standards bodies (both industry standards bodies and ISO) can use. Many of the standards groups have been using XML, and then UML, and are starting to recognise that meaning is the problem not technology. So there are plenty of examples, both as industry consortia and ISO committees (and ISO has a mechanism for industry bodies as well as countries to be involved in standards setting).


If the financial industry is anything to go by (and I think it is), we have industry groups, all of whom are also engaged with the ISO process. Some standards have a message syntax but no business representation of what is in the messages, and don't see this as a problem. Others have a message or logical data representation but no business representation of semantics, and /do/ see it as a problem. Business data managers are starting to ask questions about what the techies have been doing in the standards world and why they haven't delivered something useful and maintainable in many cases. People are starting to recognise semantics as a vital component of business requirements.


The opportunity is not to (a) create some giant consortium to go out and map the world's semantics for them, or (b) get Congress (or Parliament or Duma or Knesset or Loyal Jirga or Althing) to vote billions of dollars for us all have fun doing it. Sorry to pop those particular bubbles.

The Way Forward

Where we can have fun and do something useful as I see it is:

1. Show some leadership in helping industry bodies to capture and model their business semantics in ways which are complete, logically consistent and can be owned, reviewed and updated by business SMEs (but:

1a. given that OWL and so on don't really meet that requirement,

figure out a useful answer before the inustry bodies get bored and move away - currently they think OWL is it because this is the hot new thing in buzzword heaven); 2. Deal with the common terms that form the simple semantic building blocks that those different industry groups might want to build semantics from, for example financial building blocks (use XBRL - no question), maths terms, country, currency, all that simple stuff. Also common things like risk, payment, liability etc. that form the basic terms that any business entity has to relate to. Like "business entity".

Granted that (2) reignites the "primitives" debate which isn't my intention here, but I do think that the SUMO ontology splits too early into industry verticals. As an example, to define financial instruments, one wants to use the basic concepts of contracts, contractual terms (which are a set of contractual clauses, and therefore a set), jurisdiction, equity (financial), debt (financial), cash flows, schedules (events + time), variable parameters (for interest rates, indices) etc. etc.

Desired Outcome

If we can work out how to help with these two simple tasks, and help develop and improve the standards to accommodate 1a (for example I've been using an OWL-based thing but I've added the concept of archetypes, renamed everything in English, and output it in in diagrams and spreadsheets - surely OWL can be extended in these ways), then we will have something useful to offer which industry bodies in the different industry sectors might be able to cobble together a few grand to do. This is what I have been doing in the financial industry - through a US-based global industry body called the Enterprise Data Management Council, who tentatively found just enough money to keep me alive and dry while I had a go at putting something together. Assuming it works and and be shown to add value, we can keep this up without ever having to ask Congress to scrape together a few billion for us to keep body and soul together  :-)