Tooling WGM 2012Jan19

Tooling Minutes 2012Jan WGM 2012-01-19

Tooling Meeting Thursday Q1

Tuesday Q1 Meeting Information

HL7 Tooling Meeting Agenda/Minutes Location: 2012Jan WGM, Mesquite, San Antonio TX	Date: 2012-01-19 Time: 9:00 am CST
Facilitator: Jane Curry	Note taker(s): Lynn Laakso
Attendee	Name, Affiliation
x	Jane Curry, Co-Chair
regrets	Tim Ireland, co-chair
x	Lynn Laakso, HL7 HQ Tooling Support
x	Lloyd McKenzie
x	Andy Stechishin, Co-chair
Scanned Attendance‎
Quorum Requirements Met (co-chair plus 3 counting staff): yes

Agenda

Agenda Topics

• Roll Call & Agenda Review
• Migration Planning (V3 Generator, SMD), ES alignment update
• HL7 - Tooling & Electronic Services alignment

Supporting Documents

Minutes

Minutes/Conclusions Reached:
Call to order at 9:08 AM CST

• Roll Call & Agenda Review
• Migration Planning (V3 Generator, SMD), ES alignment update
  • migration strategies for existing products. We needed to get an update from Woody on succession planning for PubDb. Andy reports. Jane wants to get to the level of explicit plans which include iterations. MAX project has not been able to make progress; we may have to seek funding - perhaps a discussion for Q2.
  • Of the picture we already have, what is still in play. The existing picture is in the tutorial from last May. July F0B 2011 is the version Jane finds. Yellow items are not done. Woody has a proof of concept on line A WYSIWYG editor for MIF annotations, not working in all environments. There is a non WYSIWYG form in XML Spy. Second mechanism must run IIS on local machine to serve browser functions for JavaScript for editing. You don't have to enter tags; it has controls for formatting. Word introduces unprintable characters with its proprietary format if it is used in editing. Last entry on line A PubDb functionality to Eclipse: Andy could just dump it out of Access and get it to work, or he could work towards something that manages artifacts correctly. Jane notes that what we have right now works; we don't know what the new artifact generator and behavioral framework will look like. She suggests we wait and put our energy into something else for now. Andy cautions that V3 won't go away that quickly. Is it worth putting time into something that is ready just before the next thing comes out. Jane cautions that commercial tools are working on this BF artifact set description. Despite known problems, it may be "good enough". Andy notes that what Woody was doing in Eclipse may be easier. That direction was tied to progress we were seeing on the SMD.
  • Line C WYSIWYG XML for V3 MIF Annotations is closely aligned to the task in line A. How much of that solution can be used for C? Andy thinks that the solution for line C is about 40-50% of the solution you would need for line A. You might do it in Eclipse or something similar. D thread is pretty much done.
  • V3 Generator in line E is not stalled notes Lloyd. It may be reasonable to expect something before the next WGM. Lloyd says the dependency stuff is done, two elements generation of default build to infer named dependencies where it has to remains. Validation process to use dependencies in validation is also outstanding. Andy thinks Lloyd will be busier with FHIR than the last cycle, with education sessions to plan and put together. Grahame may be the one presenting, notes Lloyd; but Andy suggests Lloyd will have significant involvement.
    • One task, Lloyd notes, will require some 8 contiguous hours to sink his head into.
    • Another non-contiguous hours task estimate is 20 hours. He may be able to delegate that to Jean once he details that out. There is still some cleanup moving XSL messages into proper validation and addressing issues, as well as test cases. This is work that Jean or Andy could do right now.
  • So we hope it is releasable by next cycle, to be rolled out (best case) by next January cycle, following Plenary WGM. In between we have a cycle to beat up on it. May cycle is also doing the Normative edition so introducing the separate publishing streams for Woody and Don is troublesome. We need to take the V3 Normative edition through the Canadian V3 Generator to see what the issues are. Jane notes Woody asserts the V3 Publishing process is right now currently pretty smooth; Andy cautions that it is smooth until they found a failure which took three months to shake out last year. Jane suggests we state integration testing of tool path during the summer, and Andy adds if it goes well we'll run a shadow build on the ballot as well as the Normative Edition (if it's out) using the new tool. Andy notes that with ballot publishing, there are new controls in publishing the ballot, such as spelling errors which will no longer be fixed.
  • Line F needed tool components - taking a Canadian tool for Static Model Refinement; also the Instance Generator on OHT in release 1.1 which is also behind in its MIF representation. No elements of this row have been reviewed since this plan was drafted. Row F may become nice-to-haves; if we cannot see the business need to complete Row A. Jane notes that PubDb is behavioral framework stuff. ACTION ITEM: Andy to discuss with Woody overall use of publishing artifacts in eclipse versus integration of BF elements in the PubDB and thereby in Eclipse.
  • Andy recounts that Lloyd asserted that there would be no more RMIMs, V3 artifacts in ballots in two years. Those would likely migrate to FHIR. CDA R2.1 may still do so notes Lloyd. CDA R2.1 wireformat backwards compatible to v2; CDA R3 may be FHIR. They will take Clinical Statement RMIM, expose attributes not previously exposed, add a participation and other codes. The structure in XML must be maintained. Schematron will have a problem. Microformats or microdata has been discussed in FHIR, but it will not likely be addressed in R2. Discussions exist on how microdata will be supported in FHIR. Open EHR is looking for FHIR datatypes to become Open EHR archetypes.
  • The artifacts that we're currently balloting are approaching end of life. DMIMs and RMIMs will likely be referenced by no extensive effort will likely go into the creation of them but instead into FHIR Resources.
  • Still need Vocabulary Maintenance resources in G stream but the F stream may be on hold. The tooling required for FHIR needs to be able to do Excel and an XML editor; instance generation is part of the approach so instance example generator is not an issue with FHIR. Lloyd asserts the foundation tool to bring in CSV files and build files would be in Java. Jane asks if it could be built in Eclipse; Andy agrees but the inclination is to require a technical person in the development of these resources. Need a "Russ" there to deal with the XML, for Vocabulary as an example. The core of this is the XML representation. Jane notes the architectural implications of our forward looking tooling path we need to address - what's currently on the books is converging on the Eclipse platform. Andy doesn't feel we can address the change in the strategic direction until FHIR is formally released. The IP question is at issue.
  • Need to schedule a tooling call to have Grahame present what tooling elements are required and discuss the tooling architecture and strategy for FHIR; should not be dependent on Eclipse. The most sophisticated tool to be likely required is the profile conformance tool, but still not as complex as the Eclipse tool. Jane asks if we can still assert we will have a strategy of open source, MIF based representation. The FHIR artifacts are completely redesigned and will not be computationally transformed. Lloyd asserts if you take an existing RMIM: it will take 20% of the elements of the RMIM exposed in the formal syntax - everything else will be representable in extensions, chosen by the committee for relevance to them. It retains the knowledge, but structures the interchange differently says Andy. They will take the requirements for 80% of the community for core. Committees will publish extensions for their 100%. Not all of the elements will be in the defined attributes. Jane observes that a committee will deliberately evaluate what they think the 80% of requirements will be. Lloyd adds that everything will be an extension point. Any extensions that HL7 registers should be mappable to the RIM, including abstract data types.
  • Jane notes that we still need the bottom layer; Andy adds we still need G as well. He's asking Vocab for requirements and from which we can work on tools to satisfy the requirements, rather than throw a chunk of money at them. In Q2 Jane expects to communicate that our requirements will formally go through the HEART project until otherwise specified.
  • Migration strategy is understood, notes Jane - still dependent on V3 Generator and Desktop XML. Andy cites that Woody notes that work with publishing. Implementation guide representation for CDA IGs is waiting for some work from external organizations. Artifacts in component form rather than one big document is desirable but we can live with it for CDA templates. Other areas are in greater need of attention. CDA IG integration into publishing suite is not critical; they are producing high quality documents in PDF. Woody feels that's not something to be chasing at this point. Where we must align our specifications with other SDO specifications we may have to incorporate DITA objects. The first document in that category may be the new TermInfo. Our publishing is more robust than DITA, but only for the things for which it was built. Some people just insist on doing things in Word.
  • Jane feels we have some idea of how to go about building this Tools Tactical Steps diagram. We may find in 2012 we have an entirely new strategy on tooling requirements for stuff done in existing tools, stuff in emerging areas. People may work on RMIMs in the future. How much are we prepared to invest in the SMD, with input from an experienced Eclipse architect on updated estimates other than B2i. Lloyd thinks zero. Andy asserts if they had followed Woody's path their estimates could have been halved. The bridge between what they built and what we need is very expensive to cross now. Jane asserts we have a tool that implements NHS methodology and we wanted it to morph into a tool that would work for us. Visio updates for new versions would not be insurmountable. There is upheaval in the architecture stuff, line B especially. There is an architectural review potential for us that would result in an estimate for a full code review. Andy thinks it would be good information to have no matter what. Jane feels a teleconference would include a minimum of cochairs plus Woody and Lloyd to review requirements against the SMD architecture and get an estimate. This iteration of budget request would be an architectural assessment for feedback on what degree of flexibility this approach has to address the future. We're at a pause state of investment in tooling for this reason. Andy hopes that there is a new person working with Vocab to document business requirements to drive their effort forward. He'd like to ask them to identify a strategic vision of the vocabulary management tooling required.
  • Jane also asks to what degree Core Principles is still valid in FHIR. Lloyd notes that vocabulary is still valid. Jane adds we cannot meet our core principles for vocabulary management in our current tooling. With FHIR the CoreMIF files and Harmonization will still continue says Andy - Lloyd notes it will even accelerate for a while. Jane asks if the Tooling vision picture is still accurate- Open source MIF based objective. However the current tooling architecture may not remain. Is Grahame prepared to be a member of OHT and adhere to their rules with the FHIR. Discussion ensued with the general feeling that he was participating with them in his new status. Jane asks if we still design static models in SMD? Lloyd suggests we design resources, as static artifacts linked to behavior, with operations that can be performed on a resource in REST are documented. Jane notes that resources have exposed elements or extension points more object oriented than SMD oriented. Lloyd notes resources are not design by constraint and not a static model in the typical way HL7 has used the term. Andy notes that vocabulary representation in resources is still being defined. Static relationships between classes or elements of a resource are mapped. One objective of the methodology is to provide xml in EA format. Jane's concern is to what degree we need to change our Tooling vision. By the end of 2012 we'll have a new tooling strategy with a corresponding new tooling vision.
  • HL7 - Tooling & Electronic Services alignment

Recessed at 10:34 AM CST

Meeting Outcomes

Actions .

Next Meeting/Preliminary Agenda Items see Tooling_WGM_2012_Jan for Q2.

Tooling Meeting Thursday Q2

Tuesday Q2 Meeting Information

HL7 Tooling Meeting Agenda/Minutes Location: 2012Jan WGM, Mesquite Room, San Antonio TX	Date: 2012-01-19 Time: 11:00 am CST
Facilitator: Jane Curry	Note taker(s): Lynn Laakso
Attendee	Name, Affiliation
x	Rob Snelick, CGIT
x	Rene Spronk, RIMBAA & Marketing
x	Attila Farkas, CHI
x	Jane Curry, Co-Chair
x	Lynn Laakso, HL7 HQ Tooling Support
x	Frank Oemig, ITS
	John Quinn, CTO
x	Corey Spears, EHR liaison, McKesson
x	Andy Stechishin, Co-chair
x	Michael Van der Zel, Patient care liaison
regrets	Tim Ireland, co-chair
Scanned Attendance‎
Quorum Requirements Met (co-chair plus 3 counting staff): yes

Agenda

Agenda Topics

• Roll Call & Agenda Review
• Tactical Plan update
• Liaison update.

Supporting Documents

Minutes

Minutes/Conclusions Reached:
Call to order 11:10 AM CST

Rene reports that Michael will be 10 minutes late.

• Roll call and agenda review - Jane would like ot allow the liaisons to go first.
• Liaison update:
  • Rene reviews the wiki page for the Tools for RIM based software development with focus on standards implementation tooling rather than standards development tooling. It used to be called Implementation Tools. The page reviews tools known for using V3 and RIM as basis for software development. They are working on a tool evaluation process and hope to show something similar to Wikipedia's "Comparison of" pages. Jane notes that this was promoted in September and has been agreed to by the CTO. Discussion ensued on MIF based class code generation and MDHT. Jane notes we need a list of the tools that are available and the Tooling page would link to this page so we can give people the opportunity to find these tools. Rene suggests that they've recommended V3 Publishing include an excerpt in the Ballot or Normative edition for recommendations on implementation / software development see the CGIT group, or see this set of tools. Andy notes that a note for implementers would be good on the normative edition, but not for the ballot as the ballot versions are not recommended for implementation. ACTION ITEM: Rene will send Andy an email describing the request. ACTION ITEM: Lynn will add a link to this page from the Tooling main page. Rob asks what the scope of the Tooling wiki in terms of linking to all the tools that Hl7 supports. Jane describes the scope as those tools that HL7 supports or maintains within GForge. This will be more detailed in the Tooling Communication plan and this information would make that communication more effective. Rob would like to see a directory of all the tools. He'd also like to see a list that included those tools that were for IHE, Meaningful Use, etc. Jane notes that another part of the communication would be the part of the interoperability space that HL7 had, such as standards development.

    Rene comments further that the Tooling Wiki information is very long and may not be reviewed "below the fold". Andy notes that there has been guidance offered by the ES group for the format of Work Group wiki pages. Jane asks Rene to contribute discussion on the discussion tab and link to a page that he'd recommend as an example. She further asks that Marketing may make recommendations for how Tooling should communicate to the outside world.

    Michael was trying to create a project scope statement for the EHR Template tool. He found the format a bit awkward and would like to see some changes for projects related to tooling. Jane suggests such recommendations must go to the Project Services WG; he could send his suggestions to the listserv and they can raise them further.

  • Andy starts with a Structured Docs liaison. Not much has happened here but he may try the approach he's using with Vocabulary, which is to ask for a strategic tooling plan to look 2 years out on what they think they'll need. Tuesday Q3 joint meeting with SWG, ITS, Publishing was focused on CDA R3. Jane asks if they're aware that no further action will occur on a CDA template effort until the requirements shake down further. Andy thinks that the coordination is occurring on the Toolsmith level; it has not been expressed that there is funded support required from Tooling. Jane notes that Board funding will now be requested each cycle on an iterative basis rather than annually. Projects in the hopper must have scope confirmed as well as an estimate.

    Andy noted that Tooling cochairs serve two year terms. Jane is Board appointed cochair for next two years. Andy should look into changing his cycle to May as Tim is up for re-election in September

  • Vocab proposals from NLM may streamline Vocab harmonization process. Andy noted they nearly addressed business requirements this week, which was heartening. Jane asks what the need is with participation with IHTSDO in terms of their work bench or some other tools. Andy clarified that the relationship is between Vocab and IHTSDO but not between Tooling and IHTSDO, that they must address this from the perspective of what works with Vocabulary. He thinks they are coming back to the IHTSDO WB. The NLM process/tool is a different issue. The IHTSDO is how you manage vocabularies; the NLM is how you manage vocab change requests. Wish to use web form system to capture vocabulary harmonization request. Workflow prior to and workflow after a request is submitted/accepted was discussed. Jane looks forward to seeing that written down.
  • Patient care; Michael explained what the liaison role is in the group. He asked for Tooling requirements; none were forthcoming. He showed them a wiki page where they can capture requirements as they emerge. He sought to create a list of tools which Patient Care uses; remarks were that the tools were too hard to use. Jane suggested to the PC list that they be specific about what they'd like to see.
  • Jane asks Frank about getting the MWB updated. She notes the original toolsmith wishes to retain the source; platform change or open source update is not in Pete's plan. It's not an orphan tool if the volunteer will continue to maintain it.
  • Rob notes that there are other needs right now; Jane suggests that groups that need tools need to take responsibility for the requirements documentation and coordination. Andy notes that CGIT has some specific requirements. Update to 2.7 is one. Rob notes that recent work with LRI using 2.8 and 2.9 revealed need for MWB to accommodate those versions. But the versions are not yet approved and updated. Jane notes there may be more than one toolsmith for the MWB, and the enhancements need to be rolled back into the users. Andy notes the risk with single source developer (Peter) he cannot accommodate backup support should the need arise. Right now the organization is at risk. Jane asks for a PSS to put the MWB on a more formal maintenance basis. Rob would do so but fears he lacks the resources. Certainly they can start on building requirements for 2.7.1. Andy suggests that if the changes become critical we need to know if Peter would prioritize if funding were available. Jane notes that CGIT is driving the process, and it looks like it's not going to be available by May. ACTION ITEM: Andy will discuss this with Karen. Rob notes that in addition to 2.7.1 requirements, to get user documentation and even source code, in escrow if needed, to mitigate our risk. Jane notes that those elements might be a prerequisite for funding.
  • Woody joins at 11:58 AM.
  • Corey represents EHR, where there is a flurry of both unofficial and official activity. R1 was all done in Word. It's translated to Excel and to an XML schema. We have ability to generate html and PDF with formatted objects. ISO ballot moving forward, with ballot response created with lockdown version of the XML. He created an amalgamation macro to import those and HL7 ballots into one spreadsheet. It will split out the different sections. Different teams can work on different sections in different tabs and merge them at the end. Woody asks him to share his macro, as there is applicability to other Work Groups. Corey says he will. EHR working on conversion from CCD to VA Blue Button; has gone through Steering Division and will be entertained by TSC on Jan 30th. It has an xslt, which may have tooling implications. Jane asks if Corey would like that to show up on a consolidated list of tools, allowing other members to see it. Corey will refer that question to Lenel James who is working on it. For a toolsmith ribbon we need to know what the tool is, how it's used by the Work Group, and make it available.

    Corey also notes Keith had started something where in Word you could take Track Changes and pull it into a ballot comment response. It makes commenting a lot easier but doesn't work with PDF or html. Woody reminds him that we're trying not to publish in Word. You can open html in Word but there's a lot of header information that gets added. Woody thinks it's interesting to explore the idea of pulling changes from html updated on the clipboard. Even the PDF versions he has asked for the Word version just to generate the comments. Michael points out that PDF allows you to make comments.

    Corey adds the EHR WG brainstormed on future efforts and prioritized development goals for the EHR Functional Profiling Tool, started draft project plan. Jane asks, from a prioritized list, does the WG have development estimates? Corey says no. Jane adds that she cannot request funding without both the prioritized requirements and development estimates. Corey would have to get those updates before the next WGM cycle. Jane advises the contracts are written in weeks so the estimates should also be in weeks. Corey notes they have some technical direction points they still must decide on. Scope for iteration 1, to be delivered before the WGM, with this and all milestones estimated in number of weeks says Jane. Corey notes the original schedule should still hold. Jane would like to see the schedule and estimates for the next cycle. Corey will work with Jane on that for the next cycle.

  • Jane adds the next Tooling liaison meeting should be the third Thursday in the month, 16 February. To ask for Board funding for this cycle we need updates on estimates in the next two weeks. Corey and Michael will talk and get with Jane next week. Corey asks for an example: ACTION ITEM: Andy and Jane will draft such an estimate by Saturday. Milestones, deliverables, and effort estimates are what they need, for an approved Project Scope Statement to request funding for an iteration in a cycle. Woody will evaluate their artifacts for MIF without changing their Work Flow as perhaps a back end conversion for Publishing. Jane says there is a sub project on the list to gain MIF representation for what they need. Woody notes that we already have a number of GForge trackers on which people can add requirements that are not used.
  • Jane will be arranging a dialogue with the Eclipse architect for 2 days virtual meetings to look at what we have, which will include Woody and Lloyd with the Tooling Cochairs. The evaluation is not just SMD but also MDHT.
• Mr Farkus is director of interoperability at CHI seeking to improve standards distribution in a way to add value.
• Jane reports as liaison for OHT and Templates. Templates will pass the templates registry business process, for how to register templates and the information needed. This is being registered with OHT in the developing HEART repository which is intended for the global health community to seek artifacts. HL7 is intended to be an early adopter. She intends to use the templates registry business process as the starting process for Healthcare Exchange Asset Repository T. We can expect to see target architecture, information required and so on for the next update.
• HL7 Tooling Challenge - attendees are invited to participate; Tooling will review the suggestions on February 2, selection April 2nd.
• Andy will reach out to Attila for Canadian tools for the Tuesday tooling night.

Adjourned 12:36 PM CST

Meeting Outcomes

Actions .

Next Meeting/Preliminary Agenda Items Next meeting on February 2 - Andy and Jane are not available Jan 26. Review Tooling Challenge.

Return to Tooling

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