Code Convergence Update

Conor and Nancy; thanks for contributing the comparison.  I think this latest post is a good opportunity to discuss the progress surrounding this process, thanks to WorldVistA for getting started on this front.  We held an initial kick-off on Friday; as a result, I have sent out emails to identify points of contact within each key organization identified at the VistA Community Meeting, including DSS, Medsphere, IHS, and VA.  I will keep you posted on the progress as they are identified, having representatives from each organization will be invaluable as we reconcile differences into the future.

In the interim, there are a couple of initial questions I would pose to the community.  What tools would best benefit this process, and how should we document this information going forward?

Foremost, Conor’s tool provides a good high level overview of system variations, and performing diffs through git can contrast the routine variations.  I understand we still need to be able to contrast the data dictionaries, and in some cases the data must be compared as well.  How capable is Conor’s tool of approaching this, and are there any other tools in the community we would like to explore to support us overall?

I’d appreciate any feedback, and I will keep you posted on what I get back.

Thanks;

Matt

like0

Comments

Code Convergence Update

conor dowling's picture

Matt,

thanks to Nancy now we have WorldVistA and FOIA access too - hence some of
the report upgrades. I'm going to be putting an upgraded report on OSEHRA
this evening that gathers all this information together.

On dictionary diffs: yes, you can diff dictionaries this way. Any aspect of
any File in FileMan can be queried.

1) Counts
The filesAudit (see:
http://www.caregraf.org/fmql/reports/filesFOIAvsOVAudit.html ) begins the
process. See the counts. Those say how many instances there are in
files/dictionaries in each VistA's FileMan. Example: VA Product (Drug file)
has only 17989 in OpenVistA but 22754 in FOIA. FOIA has a much more up to
date standard dictionary.

2) Differences in Schemas:
Here's the Patient File (2) in OpenVistA (
http://vista.caregraf.org/schema#!2 ) and FOIA (
http://vistaewd.net/schema#2 ). At a glance, OpenVistA has these extra
fields ...
21400enterprise_patient_identifierMSCD;14MSC21400.033display_age621400.98
divisional_mrns621400.99id621405.1admitting_diagnosis621405.8
admitting_physician629320.8account_numbers6

3) Rows/Entries in System and Concept Files
You can get more specific too: here's a list of the first 100 entries in
the lab test definition file (60) of an OpenVistA (
http://vista.caregraf.org/query?fmql=SELECT%2060%20LIMIT%20100%20OFFSET%...)
and here's full definitions of 10 entries in that file (
http://vista.caregraf.org/query?fmql=DESCRIBE%2060%20LIMIT%2010%20OFFSET...)
Basically, diff'ing comes down to diff'ing JSON-definitions, spotting
equivalent items in systems even if identifiers vary.

We just have to define what's significant. The ability to query is there.
The question is - what are the questions?
Conor

On Tue, Jan 24, 2012 at 1:28 PM, mmccall <mccallm@osehra.org> wrote:

> Conor and Nancy; thanks for contributing the comparison. I think this
> latest post is a good opportunity to discuss the progress surrounding this
> process, thanks to WorldVistA for getting started on this front. We held
> an initial kick-off on Friday; as a result, I have sent out emails to
> identify points of contact within each key organization identified at the
> VistA Community Meeting, including DSS, Medsphere, IHS, and VA. I will
> keep you posted on the progress as they are identified, having
> representatives from each organization will be invaluable as we reconcile
> differences into the future.
>
> In the interim, there are a couple of initial questions I would pose to
> the community. What tools would best benefit this process, and how should
> we document this information going forward?
>
> Foremost, Conor’s tool provides a good high level overview of system
> variations, and performing diffs through git can contrast the routine
> variations. I understand we still need to be able to contrast the data
> dictionaries, and in some cases the data must be compared as well. How
> capable is Conor’s tool of approaching this, and are there any other tools
> in the community we would like to explore to support us overall?
>
> I’d appreciate any feedback, and I will keep you posted on what I get back.
>
> Thanks;
>
> Matt
> --
> Full post: http://www.osehra.org/discussion/code-convergence-update
> Manage my subscriptions:
> http://www.osehra.org/og_mailinglist/subscriptions
> Stop emails for this post:
> http://www.osehra.org/og_mailinglist/unsubscribe/477
>

like0