OASIS Mailing List ArchivesView the OASIS mailing list archive below
or browse/search using MarkMail.

 


Help: OASIS Mailing Lists Help | MarkMail Help

office message

[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index] | [List Home]


Subject: [OASIS Issue Tracker] Issue Comment Edited: (OFFICE-2334)NEEDS-DISCUSSION: Reference - Rewording needed



    [ http://tools.oasis-open.org/issues/browse/OFFICE-2334?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=18399#action_18399 ] 

Dennis Hamilton edited comment on OFFICE-2334 at 4/13/10 12:56 PM:
-------------------------------------------------------------------

In the SC Chat Discussion today, 2010-04-13, it was clear that this issue needs discussion.  Also, there is an important consideration on how the sheet ordering is established and how it is known.

The orderings have to be related to the persisted form of the document that hosts the formula.  When a producer emits a document that hosts the formula, it must be possible to unambiguously determine the cuboid in the reference, as recorded, and consumers should recognize that cuboid unambiguously based on information in that document format.

This, as does the resolution of row and column references, depends on the hosting implementation and the hosting specification.  We need to look at how rigorously we can establish the relevant ambient properties.

Finally, I think it is important to understand that OpenFormula says nothing about what happens when the structure of an existing spreadsheet document is manipulated, how formulas may be rewritten automatically or manually, etc.  We only know what should be possible to assert about how a produced document is to be understood by a consumer.   

[Side note: Recalculation is odd here, so we probably have to assume, just to be specific, that the effect for a document that is produced after manipulation of a consumed document, is that it is as if recalculations are always kept current, whether or not that is actually the case and whether or not there is an indicator to a consumer that recalculation is not current.)  It seems to me we have some processing model assumptions that we want to be able to rely on without getting into the details in describing what a cuboid is.  But there is more that needs to be said for the OpenFormula to work in a given hosting and domain of interoperability.]

For a definition of cuboid, I think the idea is to establish what cells, if they exist, are included in the cuboid range and the details of how such matters as sheet dimensions and sheet ordering need to be dealt with in what is expected in terms of ambient properties.  

[More Side note: How the division of labor is worked out conceptually between the hosting implementation/specification and the evaluator should be dealt with elsewhere.  One thing to consider is whether the evaluator needs to be able to enumerate individual cell references in its reliance on references or whether, in terms of the conceptual model, the hosting implementation can be thought of as providing the enumeration in some manner.  In the second case, it would seem that references and ":" operations would then have to be handled by the hosting implementation and not the evaluator.  Something else to think about.  To keep ourselves sane we might want to include a resolver in the processing model, and let it be responsible for enumerations of cells too, perhaps.  I think one reason for hypothesizing a resolver is that it needs to be able to tell that an array is in hand, etc., and there need to be ways to talk about the rows and columns of that array and its cells without concern for the actual references that correspond to those individual cells.]

[PS: I used up my time thinking about this.  I will return with a proposed statement just on the cuboid definition before 2010-04-13T20:00Z.]

      was (Author: orcmid):
    In the SC Chat Discussion today, 2010-04-13, it was clear that this issue needs discussion.  Also, there is an important consideration on how the sheet ordering is established and how it is known.

The orderings have to be related to the persisted form of the document that hosts the formula.  When a producer emits a document that hosts the formula, it must be possible to unambiguously determine the cuboid in the reference, as recorded, and consumers should recognize that cuboid unambiguously based on information in that document format.

This, as does the resolution of row and column references, depends on the hosting implementation and the hosting specification.  We need to look at how rigorously we can establish the relevant ambient properties.

Finally, I think it is important to understand that OpenFormula says nothing about what happens when the structure of an existing spreadsheet document is manipulated, how formulas may be rewritten automatically or manually, etc.  We only know what should be possible to assert about how a produced document is to be understood by a consumer.   

[Side note: Recalculation is odd here, so we probably have to assume, just to be specific, that the effect for a document that is produced after manipulation of a consumed document, is that it is as if recalculations are always kept current, whether or not that is actually the case and whether or not there is an indicator to a consumer that recalculation is not current.)  It seems to me we have some processing model assumptions that we want to be able to rely on without getting into the details in describing what a cuboid is.  But there is more that needs to be said for the OpenFormula to work in a given hosting and domain of interoperability.]

For a definition of cuboid, I think the idea is to establish what cells, if they exist, are included in the cuboid range and the details of how such matters as sheet dimensions and sheet ordering need to be dealt with in what is expected in terms of ambient properties.  

[More Side note: How the division of labor is worked out conceptually between the hosting implementation/specification and the evaluator should be dealt with elsewhere.  One thing to consider is whether the evaluator needs to be able to enumerate individual cell references in its reliance on references or whether, in terms of the conceptual model, the hosting implementation can be thought of as providing the enumeration in some manner.  In the second case, it would seem that references and ":" operations would then have to be handled by the hosting implementation and not the evaluator.  Something else to think about.]
  
> NEEDS-DISCUSSION: Reference - Rewording needed
> ----------------------------------------------
>
>                 Key: OFFICE-2334
>                 URL: http://tools.oasis-open.org/issues/browse/OFFICE-2334
>             Project: OASIS Open Document Format for Office Applications (OpenDocument) TC
>          Issue Type: Bug
>          Components: OpenFormula
>    Affects Versions: ODF 1.2 Part 2 CD 1
>            Reporter: Patrick Durusau
>             Fix For: ODF 1.2 Part 2 CD 1
>
>
> Reference currently reads:
> "A reference refers to a cuboid of cells called a range. A special case is a single cell reference, a reference to a single cell."
> Rather:
> "A reference specifies a contiguous set of one or more cells. Such as set of cells may be referred to as a range."
> I am assuming that "cuboid" was used to indicate that a reference can refer only to a contiguous range of cells. 

-- 
This message is automatically generated by JIRA.
-
If you think it was sent incorrectly contact one of the administrators: http://tools.oasis-open.org/issues/secure/Administrators.jspa
-
For more information on JIRA, see: http://www.atlassian.com/software/jira




[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index] | [List Home]