To comment on the following update, log in, then open the issue:
http://www.openoffice.org/issues/show_bug.cgi?id=5658





------- Additional comments from [EMAIL PROTECTED] Mon Nov 24 21:58:16 +0000 
2008 -------
discoleo->cornouws
> For all needing 100% Excel compatability: make sure you have all
> the same version of Excel as well as the same CPU. Both are known
> sources for differences in more complex spreadsheets ...

This is the kind of comment that doesn't bring anything useful to this issue. By
the way, if you really want absolutely identical spreadsheet results, be also
warned that different versions of OOo Calc, and gnumeric and probably any other
major spreadsheet beyond MS Excel will yield different results, so you NEED to
stick to exactly the same version of the same spreadsheet on the same processor.
[Yes, a SPARC will handle floating point very different than most MIPS 
processors.]

The problem however is very different than EXCEL compatibility: it is really the
broken philosophy in Calc, that allows to compute WRONG results, without
notifying the user. So, in the end, the user might well be unsuspecting that his
calculations are utterly wrong, and might be so catastrophically wrong.

To paraphrase a different user:
> The question, also, is not “does OOo behaviour really cause more
> real problems than incompatibility to Excel.”  The incompatibility
> already exists as Issue 5658 shows.  Both spreadsheets do illegal
> operations, *they just do them differently*.
>
> The real question is, “Do we want to have a spreadsheet that we know
> is faulty, where the faults  may be  very serious and potentially
> life threatening.

So, the real issue is that Calc allows me to do:
= 1 + '1,000,000
and get a "1" and have NO feedback that the result is probably utterly nonsense.
The unsuspecting user will likely miss this wrong calculations. How often did
you look into a 10,000 rows spreadsheet if some numbers are actual text?

Be warned that Calc often imports csv as text. You end sometimes even with a
number transformed to text. Format e.g. a column as text and enter now numbers.
They end up as *strings*, and any mathematical operation is broken. You may well
miss this on a foreign spreadsheet, and even inspecting the cells will devoid
you of any useful information, because the apostrophe in front of the text is
missing (for cells formatted as text - how often did you check that the cell was
previously formatted as text?). And have you thought of a 50,000 rows
spreadsheet. I already work with >100,000 rows (well, obviously not in Calc).
What chance do I have to find the faulty string in this spreadsheet?
[see e.g. http://www.openoffice.org/issues/show_bug.cgi?id=85328 - I was well
unaware that various cells were text - and this was a happy example that I
noticed is wrong. !4! different users beyond me did NOT recognize that the cell
was actually text.]

I would bet, NONE.

And on a different note: THERE IS NO SUCH STRING AS 1!!!!!!!!

If you mean the string 'one' than this is different from the number '1'. There
is only one number '1' and NO string '1'. This is purely a programmers
distinction with NO real backup in linguistics. '1' is a number. Basta.

Not to confuse the issue with '1 is not a number'. This is indeed a composite
string, but simple numerical symbols are always numbers. I already described
more advanced ways to handle string-input in a previous post:
http://www.openoffice.org/issues/show_bug.cgi?id=5658#desc123

Hope this clarifies some issues.

---------------------------------------------------------------------
Please do not reply to this automatically generated notification from
Issue Tracker. Please log onto the website and enter your comments.
http://qa.openoffice.org/issue_handling/project_issues.html#notification

---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]


---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]

Reply via email to