ISSUE: sub-string APS restrictions in WebCGM

[Lofton Henderson <lofton@rockynet.com>]

[...1st of 2 new issues about sub-string APS in WebCGM...]

ISSUE:  Should WebCGM introduce any restrictions, beyond those in the ISO 
corrigendum, on APS usage in sub-strings?

DISCUSSION:  I was looking at whether any substantive, normative changes 
needed to be made in generating the 2.1 text from the 2.0 text, in order to 
enable "sub-string hotspots" in 2.1.  I did find a number of informative 
improvements that would help 2.1 implementors to understand what needed to 
be done and where to find more information.

ISO license terms forbid us to circulate the published ISO corrigendum, but 
it resulted from and is technically identical to our submitted "Defect 
Report Number:  8632-1 Defect 2007-01" [0].

[0] http://lists.oasis-open.org/archives/cgmo-webcgm/200706/msg00020.html

CGM:1999 is corrected to clearly allow an APS to occur within partial text 
state, i.e., an APS may occur within Restricted Text (RT) and may contain 
an Append Text (AT) element.  According to CGM:1999 (corrected), this can 
be *any* APS.  I.e., there are no further restrictions.

Looking at the APS subsection of Chapter 3 in WebCGM 2.0 [1], it becomes 
clear that it would be useful somewhere in there to refine the 'gdata' 
construct and explain explicitly about the special case of RT.

[1] http://docs.oasis-open.org/webcgm/v2.0/OS/WebCGM20-IC.html#webcgm_3_2_1

Before we can do that, we need to identify what restrictions, if any, we 
want to put on the APS types within RT.  Right now, CGM:1999 itself would 
allow any APS type within RT.

See for example the discussion that we started last summer [2].  [2] notes 
that it could be a substring within a subpara within an RT within a 
para.  And there could be 'content' ApsAttrs on the para and the 
subpara.  This is probably close to the use case that we had in mind.  Like 
this for example:

<para><content>
...RT
.......[substring nf]
.......<subpara><content><linkuri>
...........[substring nf]
.......</subpara>
.......[substring fin]
</para>

But CGM:1999 would also allow, unless we restrict it, a substring within a 
grobject within an RT within a grobject.  (But of course 2.0 doesn't allow 
'content' on 'grobject'.)

[2] http://lists.oasis-open.org/archives/cgmo-webcgm/200706/msg00041.html

OPTIONS:

1.) no further restrictions beyond CGM:1999.
2.) only 'layer' and 'grnode' are disallowed as substring APS
3.) only subpara may occur as substring APS
4.) only subpara and para may occur as substring APS.
5.) other

RECOMMENDATION:  #3, only subpara may occur as substring APS.

(I'm open to arguments for other options, but this one seems to satisfy our 
use case, without creating unnecessary burden on the implementors.

Regards,
-Lofton.

At 05:45 PM 6/21/2007 -0600, Lofton Henderson wrote:
>One further comment.  I have thought more about this (from minutes)...
>
>[[[
>Discussion:  [...]  This only works easily with grobjects.  Things would 
>have to be carefully implemented due to model of para/subpara.
>]]]
>
>Assuming the ISO defect correction, this would be currently legal 
>according to 2.0 (I think!!! I haven't yet found anything in 2.0 to 
>prohibit it, but haven't looked carefully enough yet):
>
>BegAPS 'grobject'
>   RT x,y extent notfinal substring
>   BegAPS 'grobject'
>     ApsAttr 'linkuri'
>      AT notfinal substring
>   EndAPS
>   AT final substring
>EndAPS
>
>And this bould be 2.0-legal:
>
>BegAPS 'para' content="whole big string"
>   RT x,y extent notfinal "whole "
>   BegAPS 'subpara' content="big"
>     ApsAttr 'linkuri'
>      AT notfinal "big"
>   EndAPS
>   AT final " string"
>EndAPS
>
>And one could have a 'grobject' on the outside with a 'para' on the 
>inside.  But ... you can't have a 'grobject' outside and 'subpara' 
>inside.  And can't have 'content' on 'grobject'.
>
>Regards,
>-Lofton.
>
>At 05:39 PM 6/21/2007 -0600, Lofton Henderson wrote:
>>One comment and question about the minutes...
>>
>>[[[
>>Sub-string hotspot identification
>>Response from US S1000D graphics subproject indicated that the defect 
>>resolution to the CGM standard is the correct approach and would solve 
>>the issue.
>>Discussion:  Defect to the standard clears the way to 
>>implementation.   Will this require an erratum to WebCGM 2.0?
>>]]]
>>
>>Is it our intention to retroactively recognize sub-string hotspotting in 
>>2.0?  Or do we mean it to be new in 2.1?  The first might require an 
>>erratum to 2.0, the second (possibly) might require broadening the 2.0 
>>content rules for 2.1.
>>
>>It is actually more subtle than that.  The way that ISO defect 
>>corrections work is that they are considered to apply as of the 
>>publication date of the standard.  I.e., when the sub-string APS tagging 
>>defect correction is approved, then that is considered to always have 
>>been a part of the conformance landscape of CGM:1999.
>>
>>Therefore, if there is nothing in the current 2.0 text that prohibits 
>>sub-string hotspotting, e.g., nothing in 2.0's content model that 
>>prevents sub-string APS tagging, then sub-string APS tagging 
>>automatically is a feature of 2.0 by virtue of the ISO defect correction.
>>
>>So, strictly speaking, we might need a 2.0 erratum to *prevent* APS 
>>tagging within sub-strings, if we wish sub-string hotspotting to be 
>>2.1-legal but not 2.0-legal.
>>
>>Regards,
>>-Lofton.
>>
>>At 08:02 PM 6/21/2007 +0000, david.w.cruikshank@boeing.com wrote:
>>>The document named WebCGM TC telecon minutes 20070620 
>>>(20070620_WebCGM_TC_Telecon_minutes.doc) has been submitted by Mr. david 
>>>cruikshank to the OASIS CGM Open WebCGM TC document repository. Document 
>>>Description: Minutes of telecon View Document Details: 
>>>http://www.oasis-open.org/apps/org/workgroup/cgmo-webcgm/document.php?document_id=24466 
>>>Download Document: 
>>>http://www.oasis-open.org/apps/org/workgroup/cgmo-webcgm/download.php/24466/20070620_WebCGM_TC_Telecon_minutes.doc 
>>>PLEASE NOTE:  If the above links do not work for you, your email 
>>>application may be breaking the link into two pieces.  You may be able 
>>>to copy and paste the entire link address into the address field of your 
>>>web browser. -OASIS Open Administration
>>
>
>
>