Loading ...
Sorry, an error occurred while loading the content.
 

Re: [dita-users] Re: difference marking in DITA

Expand Messages
  • Tony Chung
    ... My tools expert co-worker (you know who, Jim) encountered issues when comparing Docbook XML files that had conditional text. He had to run the compare
    Message 1 of 7 , Aug 1, 2008
      On Thu, July 31, 2008 15:59, Deborah Pickett wrote:
      > --- In dita-users@yahoogroups.com, "Jim Tivy" <jimt@...> wrote:
      >> Has the DITA standard contemplated the idea of differencing two
      >> documents and ongoing track changes.

      > It's my opinion that tracking changes should be managed at a higher
      > level than the XML document that forms the DITA source.


      My tools expert co-worker (you know who, Jim) encountered issues when
      comparing Docbook XML files that had conditional text. He had to run the
      compare after at least one set of processing to bring the documents in
      line with each other. This was pre-publishing. Where he found problems is
      that the XML document is no longer valid, but at least we can see the
      changes.

      It's one of those things we need to live with. Change tracking is a
      temporary stage, and validation shouldn't be required. However, a best
      practice would bring the comparison tool onto the scene after stripping
      out the conditional text.

      > I don't want to dissuade those who want change tracking to take a stab
      > at it. But I did want to make sure that it's understood just how much
      > work it will be.

      If a tool could handle conditional processing in real-time and generate a
      document to view on screen with change bars, I think we'd spread the
      adoption of XML as a content-authoring standard, and finally be done with
      Word and all it's formatting and renumbering glitches.

      Cheers,


      Tony Chung
      Multimedia Production and Online Communications
      e-mail: tonyc@... • cel phone: 604.710.5164
      _______________________________________________________________________________

      web: http://tonychung.ca • cel: 604.710.5164 • fax: 604.468.1360
      skype: tonychung.ca • gtalk & msn: tonychung.ca@... • LinkedIn:
      http://www.linkedin.com/in/tonychung
    • Jim Tivy
      Well Lets try the cousin Clem approach... I think you can accomplish change marking with processing instructions. If you imagine only two things insertPI and
      Message 2 of 7 , Aug 2, 2008

        Well

         

        Lets try the cousin Clem approach...

         

        I think you can accomplish change marking with p roc essing instructions.  If you imagine only two things insertPI and deletePI, time stamped and author stamped. The p roc essing instructions can be parenthetical for inserted content and for deleted content they can contain within the PI that deleted content thus removing it from the content.  In that way the content model is always valid and you don't get a bunch of extra <ph> content containers or other tags from some other namespace. Tools would have to interpret the PIs at render time to mark changes.   You don't want to get hung up on doing it purely in Xml and then having tortured and invalid content.  So given that it might be easy - we need to develop a spec - we want interoperability?.  I guess we also have to imagine something XSLT can p roc ess.  Can XSLT do PIs or would we have to write a SAX node to put it into Xml?

         

        I think Eliots package idea is good too, but does not have to solve this problem.

         

        Jim

         


        From: dita-users@yahoogroups.com [mailto:dita-users@yahoogroups.com] On Behalf Of Deborah Pickett
        Sent: Thursday, July 31, 2008 4:00 PM
        To: dita-users@yahoogroups.com
        Subject: [dita-users] Re: difference marking in DITA

         

        --- In dita-users@yahoogro ups.com, "Jim Tivy" <jimt@...> wrote:

        > Has the DITA standard contemplated the idea of differencing two
        documents
        > and ongoing track changes.

        I don't remember any discussions specifically, but it would obviously
        have its uses.

        One of the hurdles is that tracking changes invariably means that you
        have to relax the content model or live with invalid documents.
        Imagine a task where someone has entirely replaced the <steps> element
        with their own <steps-unordered> element. The document that tracks
        this change has to have both elements, which violates the requirement
        that a <taskbody> contains one and only one
        <steps>-or-< steps-unordered> . You could relax the content model to
        allow any number of elements, but then you lose the ability to
        validate the document against a "real" DITA DTD. Relaxing the content
        model isn't even sufficient in some cases, where the order of elements
        is important.

        It's my opinion that tracking changes should be managed at a higher
        level than the XML document that forms the DITA source. One candidate
        might be the proposed DITA interchange format, spearheaded by Eliot.
        Standardizing the structure of change-tracked documents would permit
        those using different editors to agree on how to flatten a tracked
        document back down to DITA. This would need an awful lot of
        cooperation from authoring tool vendors, some of which have their own
        home-baked change-tracking mechanisms already.

        I don't want to dissuade those who want change tracking to take a stab
        at it. But I did want to make sure that it's understood just how much
        work it will be.

      • Cavicchio_Rob@emc.com
        ... That seems like a really good idea. ... XSLT can process PIs. ************************* Rob Cavicchio Principal Technical Writer EMC Captiva EMC
        Message 3 of 7 , Aug 4, 2008
          "Jim Tivy" jimt@... jimtivy wrote:


          > I think you can accomplish change marking with processing
          > instructions. If
          > you imagine only two things insertPI and deletePI, time
          > stamped and author
          > stamped. The processing instructions can be parenthetical for inserted
          > content and for deleted content they can contain within the
          > PI that deleted
          > content thus removing it from the content. In that way the
          > content model is
          > always valid and you don't get a bunch of extra <ph> content
          > containers or
          > other tags from some other namespace. Tools would have to
          > interpret the PIs
          > at render time to mark changes. You don't want to get hung
          > up on doing it
          > purely in Xml and then having tortured and invalid content.

          That seems like a really good idea.


          > Can XSLT do PIs
          > or would we have to write a SAX node to put it into Xml?

          XSLT can process PIs.


          *************************
          Rob Cavicchio
          Principal Technical Writer
          EMC Captiva
          EMC Corporation
          10145 Pacific Heights Boulevard, 6th Floor
          San Diego, CA 92121-4234

          P: (858) 320-1208
          F: (858) 320-1010
          E: Cavicchio_Rob@...

          The opinions expressed in this message are my own and should in no way
          be interpreted to reflect the opinions of EMC.
        Your message has been successfully submitted and would be delivered to recipients shortly.