[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index] | [List Home]
Subject: Re: [office-collab] Tables vs. spreadsheets
On Tue, 2012-05-29 at 18:17 -0600, robert_weir@us.ibm.com wrote: > <office-collab@lists.oasis-open.org> wrote on 05/18/2012 08:09:55 AM: > > > From: Patrick Durusau <patrick@durusau.net> > I am curious what > > is the perceived difference between tables and spreadsheets? > > > > I thought about it early this morning and reasoned that the cells in > a > > table change one by one, whereas in a spreadsheet, I may have a > formula, > > the changing of which changes values in many cells. > > > > OK, but then I thought, how is the spreadsheet example different > from a > > global save and replace where I have lots of tables and cells in > each > > table are changed in the save and replace? > > > > True, not a formula change but it is still a change that is > performed by > > the application across a potentially large number of locations in a > > document. > > > > Or is the problem how do you capture changes to the formula? That > is > > while I am changing "+" to "/" does that populate tracked changes in > all > > the cells? Well, don't do dumb things while you have change tracking > on. > > ;-) Or, make change tracking in calculated cell values the result > of > > applying formulas. > > > > This is a recurring question, not only with spreadsheet formulas, but > with references, table of contents, indices, and a few other places. > The change to one element can trigger a change in another element. > Depending on your change tracking model, you could handle this a > number of ways. But I think that it is most natural to treat a > contingent chain of changes that are reproducible from some root user > action as being subsumed into a single change event. For spreadsheets the situation can be even more complicated in that there may be changes that are not necessarily triggered by a user action. I am specifically thinking of recalculation of formulas that may lead to changed results, (RAND, NOW, ...) and possibly volatile function that may trigger there own recalculation, eg. functions that track stock market prices. I am not sure that such changes should in fact be receorded other than a generic recalculation flag (ie. it doesn't really matter whether there were 5 or 50 recalculations I think). Andreas -- Andreas J. Guelzow, PhD, FTICA Concordia University College of Alberta
Attachment:
signature.asc
Description: This is a digitally signed message part
[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index] | [List Home]