Chris R. Lee said:
I've downloaded StarOffice. It doesn't have as many bells & whistles as
Microsoft products, but I can't see anything wrong with it as long as you're
only looking for a clone that has the basic functionality of the current
market leader. That's a pity, because, if I may go slightly OT, some of us
need something completely different.
....
That's somewhat unfair. While OOo/SO Calc lacks some 3D functionality, its
worksheets are much more a third dimension than are Excel's. If worksheets
were in the sequence A, B, C, and B.A2 (OOo/SO addressing syntax) contained
the formula =A.B5, copying that cell and pasting it into C.B3 would make
that cell's formula =B.C6. Worksheet references can be frozen using $ just
like rows and columns. There are a few other things OOo/SO Calc does better
than Excel. Search the ng archives - I've listed some of them before.
What we really need is a (frustratingly) simplified spreadsheet with
formulae but with nothing that prevents you from totally checking
mission-critical applications. People working in regulated industries such
as pharmaceutical production use specific software that has taken years to
validate for routine calculations. Anything unexpected or otherwise
non-routine has to be done by hand, because XL and StarOffice are so
complicated you can never be sure there isn't a program/formula error
lurking somewhere.
To an extent this is a criticism of all spreadsheets, all functional
languages, and indeed all programming languages generally. It's even
possible to miskey entries on pocket calculators, rendering this complaint
almost meaningless.
Spreadsheets generally are one heck of a lot easier to use than, say, Lisp
or APL. Spreadsheets are susceptible to programming errors just like any
other form of programming, and studies have shown spreadsheet error
densities to be comparable to other programming lahguages. The point is that
programming takes care, and like it or not so does programming using
spreadsheets.
Others also have experience writing software in various languages, and they
can share their experiences and impressions, but IMO as long as formulas are
kept fairly simple (no more than, say, 5 function calls per cell), *current*
spreadsheets are simpler to use *and* verify than most of the alternatives.
I suppose the same issue applies to businesses that rely on homespun
spreadsheet developments for their accounts.
Simple bookkeeping in spreadsheets is one thing - it compares favorably to
paper, pencil and calculator. Full fledged double entry accounting systems
are a PITA to try to implement using spreadsheets. Larger companies use true
accounting systems. Spreadsheets are used for bugetting and forecasting, not
financial reporting.
No doubt many contributors wonder why formulae other than simple descriptive
statistics are so complicated and counter-intutitive. It's an inevitable
consequence of the restricted environment - a bit like the old batch-file
language of MS-DOS - and we'll have to wait for a new generation of
spreadsheets that work by drag-and-drop or something.
Got some *SPECIFIC* examples?
How would drag & drop work? The spreadsheet would read your mind and do what
you want it to do? At some point you, the human, have to tell the computer
what to do. Sometimes that means telling it *how* to do it.
No doubt our different perspective come from different experience using
computers. If you've ever spent time debugging programs submitted on paper
punch cards, you'd think very nice thoughts about spreadsheets.