2007-03-08 Developers Conference Call

<html><head><title></title></head><body>h1. Date
8 March 2007

In Attendance

  • Ben Wolfe
  • Martin Were
  • Burke Mamlin
  • Justin Miranda
  • Darius Jazayeri
  • Christian Allen

Others present

  • Viki Prescott


Licensing Agreement

Darius had a call with Paul, their COO and their lawyer a couple of days ago. We finally got their revisions at Midnight last night. Everything should be good to go by the time we have our call next week. We needed to add to the assignment document. There is a paragraph at the end saying if OpenMRS LLC does not release the stuff within 30 days as OpenSource under the OpenMRS license then the assignment becomes null and void or something like that.

Burke asked Darius to send him an email about some of the attributes he needs for the token things to do Darius’ filtering. It can be just a list of what we need to know (i.e. what the data type is).


Burke asked Darius about an email he sent out about migration and that you could get a sample HL7 message from that tool. Little discussion was made on this subject.

Data Model Update

Darius asked about the data model updates. Burke said that the primary one is the person. Ben said he has been held up because of other administrative things, so he hasn’t been getting that much work done but that he hasn’t hit any major snags yet and he is at about 85%. Burke said that he met with Andy Kantor and he and a friend and Andreas. Where we ended up was with a little bit of fine tuning to the concept map table and stuff, the big change would be to level the playing field between synonyms and names.


Some discussion concerning tokens. Darius asked if all tokens have to actually exist in the database. If I create a new concept are you going to create a token for that concept? Burke-at this stage, the idea of tokens is that whatever you are going to look up through tokens would be registered. I haven’t thought about or planned, at this stage, to persist all of that. It would get loaded as the system got loaded and then as concepts were added they would be registered. The idea would be that any time, whether it’s a concept or some rule or anything you are going to want to pull in and want to be able to result out through that service would be registered with a stream and so you would just zip through the existing concepts, register them and then any time a new concept was added, that would add a token and once deleted, it would remove the token. Darius-it seems like it would be nice to delegate that to concepts. You register the concept token delegate. From a perspective of someone who is using the token service you should, like if I ask the token for CD4 count. The way I would expect it to behave is, unless there’s a rule or something else already registered, like some special case for CD4 count, it’ll just got look up concepts and bring back a concept and pretend to behave dynamically. Burke-what you are going to get back is token but you’ll be able to get back to the concept through the token. Where I have kind of expanded it is that tokens and results from tokens will carry along this object. The idea of the token service and the logic service, in general, is basically to turn things into these date value pairs, arrays, that can happily come play with each other, whether they are concept or an API call or anything else. I kind of heard from you and agree that there are times when you want to be able to say, â€ÂÂdo this calculation and get the results for me†but I know that it is all based off of programs and so, I want to be able to take any of that value in the list and ask it to give me the actual object behind this. I can kind of carry that through the logic service and not thinking of coding it with type safety and stuff.

One other thing Burke wanted to get out was one of the other benefits with stuff with Andy Kantor and consolidating names was an idea about being able to say, when you can qualify names is either preferred or being for reporting or like a short name or something. When that becomes just a type. One of them you can say, â€ÂÂthis is kind of the patient preferred name.†So, if you’re going to be printing out a report for patient education materials, you can always go look for a patient version of that.

The question of where we are with tokens came up. Burke responded that they are between design and release phase. Burke is coding right now it is a moving target but there are a lot of things depending on it, so he is trying to get it out quickly.


Burke spoke a little about Mike McCoy coming to Regenstrief and what things he will be working on and stressed that the movement for Regenstrief is more toward OpenSource.

Brief discussion on cpoe.

Meeting Adjourned