Eclipse Architecture discussion

Tooling Meeting

Meeting Information

HL7 Tooling Meeting Agenda/Minutes Location: Phone: +1 770-657-9270; Participant PassCode:586935# GoToMeeting URL: https://www1.gotomeeting.com/join/699884034	Date: 2012-05-02 Time: 13:00 PM EDT
Facilitator: Jane Curry	Note taker(s): Lynn Laakso
Attendee	Name, Affiliation
.	.
x	Woody Beeler, Beeler Consulting
x	Wilfred Bonney, HL7 HQ
x	Jane Curry, Co-Chair
x	Austin Kreisler
x	Lynn Laakso, HL7 HQ Tooling Support
x	Ed Merks
x	Brian Pech
x	Abdul-Malik Shakir (MAX)
x	Andy Stechishin, Co-chair
x	Darin Wright
Quorum Requirements Met (co-chair plus 3 counting staff): yes

Agenda

Agenda Topics

• Roll Call & Agenda Review
• Eclipse Architecture Discussion
  • discuss how HL7 has used the Eclipse platform to date,
  • what HL7 Tooling WG intended to achieve using the Eclipse platform
  • seek advice from a senior Eclipse architect.

Supporting Documents

Minutes

Minutes/Conclusions Reached:

• Roll Call & Agenda Review
• Eclipse Architecture Discussion
  • Jane refers to earlier document circa 2006 on why we recommend Eclipse as new platform for integrated set of tooling. Andy recalls open source and vendor neutrality were part of the rationale. Jane asks if those are still our objectives and is Eclipse still the right platform to meet those objectives. Woody suggests that "right" might be considered the "best opportunity" compared to other platforms.
  • Jane notes we have two experienced Eclipse architects on the call so the opportunity is now to learn what we need
  • Introductions made.
  • Jane describes the complexity of the environment in which we are trying to manage a tooling strategy. Static Model Designer replacement intended in Eclipse.
  • AMS asks how commercial products fit into the Tooling strategy? Interchange between formats like CSV, etc to take components of model content into other tools?
  • Woody describes the current environment and its interaction with Eclipse. EMF is used to demonstrate core models.
  • Registration of concept domains in our vocabulary data store, currently in MIF. Ongoing evolution of terminologies is also a significant challenge. Support of publishing is also an important commitment. Artifacts can have their own change management lifecycle and versioning is an important part of the function.
  • Woody asks about integration; Ed cautions against combinatorial complexity explosion using EMF with multiple input models. Modeling is happening more in eCore. Woody described work importing MIF ballot material into eCore EMF. Jane notes that we also interact with other tools from other sources and maintain semantic information, whether an XMI import or export. She mentions the OWL approach to ontology management and the vocabulary data maintenance and semantics management that is part of the larger picture.
  • Mapping and transformation tools in the eCore space? M to M project, ATL Atlas Transformation Language, OMG has QVT spec. XMI as standard from OMG is one possible serialization but not the only one. Woody receives questions to receive RIM in XMI…
  • Tools that provide XMI provide schema on which it's based, an emof file. Need a reference implementation to demonstrate consumability for users, unlike V2.
  • Jane notes we need to identify what the next step for what each of these tools is going to do, in addition to demonstrating that they do what they are intended to do. If we are in refactoring mode with existing tools, what will it take to do that? Also how to formalize a continuous change management process…
  • They asked what is our primary way of exchanging data - XML format? Woody says yes, we represent in MIF, then XML.
  • Eric adds that XML serialization is okay for interchange but not for persistence in a repository. See CDO from the EMF page at http://www.eclipse.org/cdo.
  • Jane notes that consulting opportunity is being pursued to flesh out the rational way to set the strategic vision and tactical plan. AMS wants to ensure there is a knowledge transfer component as well. Jane notes that full-time staffers do not have Eclipse knowledge now. We also have templates to manage, as formalized use-case specific constraints against the balloted models that need to be artifacts in themselves, governing rules for populating instances.
  • AMS volunteers to act as Tooling liaison to Education if there is formal education that will be developed as a part of this effort. Jane seconds the idea as well as building our toolsmiths. AMS also suggests we offer HL7 training to any consultants that come on board to make them familiar with our space.
  • Woody offers to get together with Andy to look at CDO while in Vancouver. Jane feels this discussion has validated the strategic direction to work in Eclipse from five years ago. Ed concurs that while EMF may not be the perfect choice it's the best choice available for what we're doing. Jane would like to get a list of Eclipse-respected expertise as a pool from whom we can obtain some training/consulting. Ed knows the people who work in the various areas and can help identify sources especially in some of the satellite areas. Jane will discuss with him offline and send him the set of documents he sent to Darin. Woody will also forward his exploratory discussion on translation in EMF.

Adjourned 2:49 PM EDT.

Return to Tooling

© 2012 Health Level Seven® International. All rights reserved.