The Vision (potential)
A sketch of the vision based on the objectives in the Project Poster
Aspects to consider when defining the vision include:
Technical
SNOMED CT content - How should SNOMED CT content be organized. For example:
What module structure is required to support the vision?
Does the current URI for the SNOMED CT International Edition need to stay the same?
What should the history mechanism look like? What happens if the history grows much more rapidly with a potential increase in content changes? What does a 'Delta' look like?
How should SNOMED CT reference sets (including subsets, maps etc) be created and distributed?
Authoring - Terminology authoring tools and functionality, including for distributed authoring and promotion of content
QA - QA tools and mechansims for quality improvements (both at time of authoring and post-authoring)
Distribution - Distribution mechanism, release format and frequency of releases
Governance & Process
Responsibility for editorial rules, authoring, quality assurance (manual and technical) and education
Who authors what content and when?
Who processes which requests and when?
Who quality assures what content and when?
Who distributes what content and when? (including how frequently)
What automated processes could be introduced to each stage in the production of a release
What skills are required by terminology authors and reviewers, what accreditation should be required, and how should this accreditation be achieved?
Perspectives - SNOMED International, NRCs, vendors, implementers and clinicians
Comments
While I agree with this .... I was trying to avoid making specific assumptions about the vision (as we discussed earlier in the week), but rather list the aspects that need to be considered when discussing this with the group.
However, (in reference to the "RF2" with the red line through it in your diagram) I think we should distinguish between the RF2 logical model and the RF2 distribution files. I am assuming that we would retain the logical model of RF2 (e.g. with modules, effectiveTime, the RF2 properties of each component etc), while (potentially) moving away from the flat distribution files (as they are today).
absolutely... we conveniently skipped making assumptions to just getting into it
And yes, the RF2 logical model is key to SCT and that is unlikely to change, but the distribution may well do. How systems ultimately consume it may change though once you move into a services-based delivery mechanism.
I love (and share) your vision.
I'd be happy to collaborate as our NCR strategy is alligned with this vision.
Let's get to work!
@Former user (Deleted) in response to one of your points...
Part of the vision would be a move away from file distribution completely (or at least almost...) with a distribution mechanism via services, to a point where the need for releases should be discussed. If current content can be turned around immediately, there is an argument to make it avaialble as soon as marked as complete, obviously with an effective date to identify which version. It changes how we think about what SNOMED CT is, instead of a single block, it becomes a living terminology which evolves rapidly.
All make for interesting discussions!