[OASIS Issue Tracker] Commented: (OFFICE-2216) Clarify patterns(regular expressions and wildcards)

[OASIS Issues Tracker <workgroup_mailer@lists.oasis-open.org>]

    [ http://tools.oasis-open.org/issues/browse/OFFICE-2216?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16686#action_16686 ] 

Dennis Hamilton commented on OFFICE-2216:
-----------------------------------------

There are two significant differences in where comparison/pattern-strings show up in ODF and specifically in OpenFormula.

1. In ODF, apart from formula-valued attributes, the main non-OpenFormula use of regular expressions is as the value of an attribute.   So the attribute value is the comparison/pattern-string.  If we come up with a strict regular-expression syntax that shall be supported, it is still possible to allow for possible alternate forms, including extensions to the base form, provided there might be a way to tell that is happening and whose alternative it is.   We normally allow for this with prefixing.  I do not know whether the currently-implemented versions of these patterns is one that allows unambiguous introduction of a prefix.  The current drafts need to be reviewed for this and for whether or not an explicit pattern datatype is defined and used.  Note that translation of these comparison strings from one version of an implementation to another is conceivable, including from legacy ODF 1.1 documents, assuming that the ODF 1.2 approach is not completely backward compatible.

2. In OpenFormula, the comparison/pattern-string values show up as strings that are used as operands in formulas.  These strings may be recorded literally in the formula, with suitable quotation and escapes to introduce quotation characters literally.  These operand strings may also be obtained by reference to table-cell values, field values, and other indirect means (including as the result of functions and other operations that are evaluated to achieve the operand value).  This means that, in general, it is not possible to always know when a string is intended to serve as a pattern; the format of the pattern may be wired into the logic of the OpenFormula calculation and never visible as a literal string value.  This means that translation between regular-expression forms is not assured when moving between legacy spreadsheet-formula expressions and OpenFormula formula expressions that involve comparison/pattern-strings as operands.  (There is a similar problem with regard to replacement of calculations employing numeric, ordinal-day values with calendar-based date-time datatype values having non-numeric datatype.  Complete translation is not always possible and ordinal-day numeric computations must be preserved in such cases.)  This also doesn't appear to be a place where prefixing on the string is appropriate.  We cannot avoid malformed or unrecognizable comparison/pattern-string values having to be detected at run time in the midst of a recalculation.

3.  It would seem, at a first level, that (2) requires a rather strict and unforgiving definition of the syntax and use of comparison/pattern-string operands in OpenFormula.  Whether that is also tolerable for the case under (1) may be a separate matter, since extension mechanisms are more easily dealt with in (1) and translation would appear to be possible so long as the necessary features are available in both directions that might introduce translation.

> Clarify patterns (regular expressions and wildcards)
> ----------------------------------------------------
>
>                 Key: OFFICE-2216
>                 URL: http://tools.oasis-open.org/issues/browse/OFFICE-2216
>             Project: OASIS Open Document Format for Office Applications (OpenDocument) TC
>          Issue Type: Improvement
>          Components: Database, OpenFormula, Table
>    Affects Versions: ODF 1.2
>         Environment: This issue applies to OpenFormula, how open formula is profiled for ODF 1.2, and also how regular expression and wildcard patterns are used as features in OpenFormula and in unrelated parts of ODF 1.2 Part 1.
>            Reporter: David Wheeler 
>            Assignee: Dennis Hamilton
>             Fix For: ODF 1.2
>
>
> Clarify the specification of pattern processing (regular expressions and wildcards).
> Section 3.3 briefly notes wildcards; need to define more clearly.
> Need to clearly note that table:use-regular-expressions and table:use-wildcards are mutually exclusive (at most one can be true).
> Database criteria specification (section 4.9.8) needs to be clearer.
> Each function affected by patterns (other than the database functions which are governed by the database criteria text) need to be clearly identified as such.
> David A. Wheeler will help with this.

-- 
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