Re: [voldemort] why not add repositories to Volemort?

I couldn't agree more Tom. The reason for no repository in this first VDM pass is administrative: any data store would have to be chosen and validated for use inside the VA. In order to get on with (system) data-analysis, it was decided to leave the ins and out of such plumbing decisions to later, just grab data, cache it on disk and process from there and the code is designed so that a data store can be slotted in later. I like this order - focus on content, see how far it gets you and then upgrade the infrastructure. On Fri, Nov 2, 2012 at 10:53 AM, Tom Munnecke wrote: > Last January, I proposed an architectural component I called a Foundation > that would serve as a repository for everything required to operate an > instance of a VistA or VistA-like installation. I proposed that this > repository be available as a SPARQL endpoint, allowing standardized access > into one instance, or comparison of multiple versions; such as comparing a > VistA and CHCS instance. The repository would also serve as an archive for > future activities, to study the evolution of the instance over time, > showing upgrades, changes to coding systems (e.g. transition from ICD9 to > ICD10), etc. It would also be valuable for reconstruction of longitudinal > clinical studies - "what was the coding structures used 20 years ago" when > doing future "big data" searches. > > Tracking the flow and changes of medical semantics over time - what > linguists call the "diachronic" perspective, is a much different > perspective than the semantics at any given time - what is called the > "synchronic" perspective. Unfortunately, health IT is awash with > synchronic structures - "Let's everyone upgrade to ICD10 in a couple of > years so we can talk about "W59.83XD subsequent > enounters of being crushed by non-venomous reptiles" > > Having an ongoing repository gives us a diachronic perspective of VistA - > how it changes over time, where it was in the past, and perhaps, an > indication of where it is going in the future (think of an undefined wiki > word in a wiki shaping the future growth of the wiki). > > I think it is also necessary for coping with the avalanche of changes > coming at us from genomics and life science discoveries in general. > Understanding the flow of the repository over time (not unlike the "recent > changes" log in a wiki) gives us a much more flexible platform for dealing > with a rapidly changing semantic environment. > > And finally, I think that a huge untapped issue in medial informatics is > the notion of temporal semantics - and having a formal basis for naming and > managing the flow of meaning in a clinical system (as well as a patient's > record) has great value. > > It just seems to me that having a permanent repository of the foundation > is a really powerful tool to add. It seems to me that growing Voldemort > from a "snapshot" tool to a contributor to a permanent repository would be > a good thing to do. > > > -- > Full post: > Manage my subscriptions: > > Stop emails for this post: > >