Next
Assumptions, risks and known limitations
The greatest strength of the approach is also its greatest weakness: simpler postcoordination rules will usually enable different users to represent some clinical notions in more than one way. For example, suturing might be represented as either of:
|
Closing USING suture |
inserting ACTS_ON suture TO_ACHIEVE closing |
Defining templates might be one way to limit the number of alternative schemas. Read about templates in the tutorial.
However, provided all alternative schemas can be automatically detected and normalised to a preferred form, enabling freedom of expression within carefully controlled limits becomes a strength not a weakness: it makes it easier for users to say anything at all.
Whilst a very large number of alternative schemas exist that might appear plausible to some end users, not all of them will ever be used in practice. Further, few of them will be used very commonly. Therefore, it is sufficient to cope only with those that are actually encountered, ie to lazily enumerate them. This implies an iterative process of release-monitor-revise, such that the Authoring Terminology becomes progressively more mature and robust by virtue of its continuous tuning to actual usage.
However, an inevitable consequence of lazy enumeration is that, until an Authoring Terminology becomes mature, some alternative schema expressions will be authored before the centre has devised a transform for them. In turn, expanded Target Terminology expressions will be incorporated into information systems whilst not compliant with the preferred semantic schema. Semantic interoperability will therefore be compromised. This is considered to be acceptable firstly because it is a temporary transition state and secondly because, without an intermediate representation approach, it is very likely that this same situation would pertain indefinitely as people struggle directly with an over-complex Target Terminology.
Next