Re: Bug in selection intervals??

["Daniel P. Dougherty" <doughe57 msu edu>]

On Monday, February 14, 2011 10:18:14 pm Andreas Guelzow wrote:
> Gnumeric is intended to produce essentially the same calculation results
> as Excel (+- some obvious bug fixes.) As a consequence empty cells have
> to produce 0 when used in a calculation.
>
> I am looking forward to see your spreadsheet implementation.

Why? As you mentioned there is already Excel, Minitab, SPSS, SAS JMP, and 
other enterprise solutions out there.  Many of these either have computational 
errors, are not user friendly or like SAS JMP have too many bells-and-whistles 
to make them useful for entry-level/instructional purposes and many just don't 
work as well as Gnumeric even though they are (quite) expensive.  The nice 
thing about Gnumeric is that it has and continues to fill the niche left open 
by these other (non-open source/non-free) spreadsheet apps by simply being 
better at what it does.  I hope that finding smart, intuitive, and user-
friendly ways of overcoming the deficiencies of these other apps will continue 
to be a primary development goal of Gnumeric.

This 2008 paper discusses some of these very issues but also highlights the 
niche that Gnumeric actually does fill and points out the weaknesses of Excel 
2007:

Computational Statistics & Data Analysis
Volume 52, Issue 10, 15 June 2008, Pages 4602-4606  

As a Gnumeric developer, do you really mean to say this?
> Gnumeric is intended to produce essentially the same calculation results
> as Excel (+- some obvious bug fixes.)

As an end-user using Gnumeric in an instructional setting,  all I can do 
(supposed to do?) when  I  (or my students) find something appearing in a new 
version that is non-intuitive, difficult, or seeming to give unintended results 
is to report them back to the developers.  For most users, Excel compatability 
is just a plus and likewise R compatability is just a plus  -- not the real 
deal maker/breaker.   I doubt most Gnumeric users see these as the main 
reasons they use Gnumeric.  1) Gnumeric is _much_ easier to use than Excel and 
tends to give more correct analyses 2) it's a spreadsheet app and so although 
limited in some ways much easier to get into than learning R/Matlab/SAS or 
whatever.

If Gnumeric starts getting away from 1) and 2) then it's losing the niche.  I 
doubt Excel/R/Matlab/SAS or whatever would ever bother to fill the niche 
however.  They've had all the time and resources and still have not wanted or 
been able to do it.  Probably some other open-source thing would end-up coming 
along -- but I doubt anybody would view that turn of events as a success.  
Gnumeric is already here -- why not try to make it as good as it can possibly 
be at filling the niche it does?

P.S.

About 1/3 of my students now use Mac... and that number is growing each year. 
Xcode+Macports+Gnumeric continues to provide for them a much needed (if not at 
times harrowing) lifeline to Gnumeric.