XML-MIF Publishing Process

XML-MIF Project Background

This project arose from the tactical analysis of how to improve and stabilize the V3 Publishing process that is based opn XML and currently trying to migrate to MIF representations of all V3 artifacts. The tooling projects forwarded to the HL7 Board state:

The Tooling Work Group and the Publications Work Group has determined that a third foundational project should be performed at this time. In the last two years we have standardized on using MIF2 as our model (i.e., artifacts) exchange format. Working with the Publishing Work Group we have determined that it is time that we upgrade the XML that we use in our internal publishing processes to make use of MIF2. This requires that we produce MIF2 export from our Publications Database and convert our existing XML publishing process to be MIF2 enabled.

This will allow HL7 to upgrade its existing XML stream in the publishing process which is needed to participate in the new Specification Publication Process. We also need to document the implied requirements so that we have solid information as we start producing the other new tools necessary to support our complete tooling plan.

For the remainder of this year we plan to get all publishing process MIF2 enabled for the 2010 V3 publication. Paul Biron and Andy Stechishin (publishing WG) are prepared and capable of doing this work.

Roles/Style

• Plan/Oversee/Design Critic - WB under current HL7 contract
• Oversee Publishing Results, processes, etc. - DL as HL7 responsibility
• MIF Review, Schema Design, Transform Design - PB - new contract
• Transform Development, Testing, Process Design & Documentation - AS - new contract

Steps

Basic Assumptions

1. Everything is (already) in XML - we need to "unify", "improve" it by moving it into a defined, formal set of schemas - MIF
2. Focus on two paths - PubDb-based content; and manually-edited content. (Twist is that data types defined in ideosyncratic format.)
3. Deliverables become:
   1. Plan for execution and buy-in from Quinn, Publishing (and Tooling) - needs time to "come up to speed"
   2. Means to get "old" content to "new" (MIf-based) -- for each kind of content
   3. Means to do basic publishing from the new content.
   4. Means to manage the process
4. Change proposals for MIF

Contracting strategy

Use time and materials not to exceed x per deliverable for the starter plan, with the specifics the for rest of the budget to come forward as part of the plan.

• [Note: the xhtml namespace conversions (a.k.a. MIF 2.1.5) was tackled and completed by GWB over weekend of 8/1/9.]

Planning and Design Decisions (AS, WB, PB)

1. Explore WYSIWYG Editor Opportunities from NHS (AS)
   1. Determine whether to include this integration in the current project
   2. Current NHS implementations are (I believe):
      1. based on TinyMCE
      2. Also documented on OHT see Step 2-5 here
   • Progress
2. Review, critique, document strategy current usage of MIF to Publish (PB, AS, WB)
   1. Examples
      1. RIM
         • Current Published Content derived from
         • Source as rim.mif and rim-publication.mif on SVN at http://gforge.hl7.org/svn/v3-ballot-pkg/trunk/V3BallotsFeb2007/input/infrastructure/rim/
         • using transforms v3mRimMIF2HTML.xsl and v3mRimMIF2HTML-detail.xsl in publishing xsl directory
      2. Vocabulary
         • Current Published Content derived from
         • Source as vocabulary.mif and vocabulary-publication.mif on SVN at http://gforge.hl7.org/svn/v3-ballot-pkg/trunk/V3BallotsFeb2007/input/infrastructure/vocabulary/
         • using transform v3mVocabularyMIF2HTML.xslin publishing xsl directory
   2. (If needed) seek amendment to MIF2 (also ongoing)
   • Progress
3. Map (plan) current DTD-based Structures to MIF (PB, AS)
   • NOTE: - All of the support files - DTDs, ent, xsl, Ant scripts are in a single "support" directory hierarchy on SVN
   1. PubDb XML
      • Example is uvpa.xml file from Patient Administration
   2. v3mDomain.dtd (same as above??)
   3. v3m.dtd (??) Generic publishing DTD
   • Progress
4. Select PubDb strategy (WB, PB, AS)
   • Transform current xml to
   • Generate from DB to MIF
     • Java - Eclipse
     • RoseTree (VB6)
   • Progress
5. Review Current Publication Format and consider changes to more readily align with MIF (WB, DL, AS, PB)
   1. Document Style changes
   2. DECIDE/Document needed transforms to either
      • Publish in old form from MIF
      • Migrate old content to new Form
   • Progress
6. Phase II Plan
7. Notes from Discussions

Build Useful Stuff

• Next Nine Months
  1. Build and Test PubDb to MIF process (AS, PB [WB?])
  2. Design/document new Transforms needed for Publishing from MIF content (PB, AS)
  3. Build and Test Transforms for Publishing (AS, DL)
• Stage Two
  1. Design and integrate Publishing transformation processes into current process (AS, DL, PB,WB)
     • First phase of implementation complete.
       • PA Domain for review here in ZIP
       • ZIP of all domains (but with image only for PA) is here
       • Don Lloyd has created a scratch site for the review of the new generated artifacts go [here]
  2. Design and assemble an editing platform (using WYSIWYG editor) plus a MIF-driven structure to create HL7 Specifications directly in MIF (AS, DL, PB,WB)

Other Opportunity to "Do Good"

1. CDA Transform from Static Model to X-path & Schematron constraint statements (AS, PB, CDA)