On Mon, May 27, 2013 at 2:27 PM, Dennis E. Hamilton <dennis.hamil...@acm.org> wrote: > Without concern for rationale, it is part of the ODF 1.2 OpenFormula syntax > that unary minus has precedence over all binary operators. > > It is not uncommon for there to be such a rule in programming languages > (C/C++, Java, and JavaScript, for example). > > Some languages that have an exponentiation operator (not that common) might > have different precedence cases, so that > > A^-b might be either illegal or A^(-b) > > -a^b might be either (-a)^b or -(a^b), the second being the case for > Fortran > > In Openformula both forms are legal and the unary - has higher precedence (is > done first) over ^. > > It is unfortunate that the built-in Help defines the operators in Calc but it > does not specify the precedence. > > - Dennis > > PS: There are times when INVALID and WON'T FIX are inappropriate responses. > One can be taken as saying something about the submitter, the other says > something about the developer, although WON'T FIX is appropriate for a > trivial defect that has no serious consequences or for an > implementation-specific case that will be preserved. My preference over > INVALID is BY DESIGN (it is meant to be that way) and even SPECIFIED or TO > SPEC or STANDARD (when it is required to be that way). >
We use INVALID for two main things today: - user didn't submit a defect report at all, but posted a help request, "how to" question, etc. - user submitted a defect report but the behavior was not a defect WONTFIX mean something is a actually a valid defect, but for one reason or another it won't be fixed. IMHO, if what we're doing is in accordance with ODF, and is needed for compatibility with MS Office and past OpenOffice spreadsheets, then the defect report is INVALID. There is not other way to treat it, if we want to be consistent. Note: the distinction is important if we track defects by components, etc. Knowing how many real (non-invalid) defects are in a component, or are found in a given release, etc., is good information to know. We lose the value of that data if we mix it up by calling something WONTFIX for political reasons. This should be an engineering decision, not a public relations effort. Making best thing is to rename "INVALID" so it is less offensive? Maybe NOT-A-BUG or something like that? But still preserve the distinction between that at WONTFIX. Regards, -Rob > -----Original Message----- > From: Rory O'Farrell [mailto:ofarr...@iol.ie] > Sent: Monday, May 27, 2013 05:21 AM > To: dev@openoffice.apache.org > Subject: Re: Question about: OpenOffice Calc. "=-1^2" like expressions > > On Mon, 27 May 2013 14:08:15 +0200 > Andrea Pescetti <pesce...@apache.org> wrote: > >> Andre Fischer wrote: >> > In any case, this is a problem with no good solution. >> >> Indeed. Before this becomes another 0 ^ 0... there is an issue for this, >> it dates back to 2004 and it is marked invalid. That's it. >> >> https://issues.apache.org/ooo/show_bug.cgi?id=24271 >> >> If someone has strong feelings he can post comments there, even though I >> don't believe that marking the issue as REOPENED would have any >> practical effects. >> >> Regards, >> Andrea. >> > I think whether this behaviour be mathematically correct or incorrect we must > follow the behaviour of MS Excel. > > -- > Rory O'Farrell <ofarr...@iol.ie> > > --------------------------------------------------------------------- > To unsubscribe, e-mail: dev-unsubscr...@openoffice.apache.org > For additional commands, e-mail: dev-h...@openoffice.apache.org > > > --------------------------------------------------------------------- > To unsubscribe, e-mail: dev-unsubscr...@openoffice.apache.org > For additional commands, e-mail: dev-h...@openoffice.apache.org > --------------------------------------------------------------------- To unsubscribe, e-mail: dev-unsubscr...@openoffice.apache.org For additional commands, e-mail: dev-h...@openoffice.apache.org
