Hello, All!

According to http://www.msnbc.msn.com/id/21033161/
it looks like Excel is having trouble either displaying or using
2^16. I'd really like to see the sloppy code that causes this!


James Silverton
Potomac, Maryland

E-mail, with obvious alterations:
not.jim.silverton.at.verizon.not

And apparently discovered / reported by a " Molham Serry" in the
Public.Excel group on the 22nd.
http://tinyurl.com/338gjd (146 messages)
--
Jim Cone
San Francisco, USA
http://www.realezsites.com/bus/primitivesoftware
(Excel Add-ins / Excel Programming)



"James Silverton"
wrote in message
Hello, All!
According to http://www.msnbc.msn.com/id/21033161/
it looks like Excel is having trouble either displaying or using
2^16. I'd really like to see the sloppy code that causes this!
James Silverton
Potomac, Maryland
E-mail, with obvious alterations:
not.jim.silverton.at.verizon.not

Jim wrote on Fri, 28 Sep 2007 15:38:51 -0700:


JC And apparently discovered / reported by a " Molham Serry"
JC in the Public.Excel group on the 22nd.
JC http://tinyurl.com/338gjd (146 messages)
JC --
JC Jim Cone

I wasn't implying that I had discovered anything about the
binary representation except to wonder if it might be some sort
of overflow problem and a prime example of warning programmers
"If it ain't broke don't fix it".


James Silverton
Potomac, Maryland

E-mail, with obvious alterations:
not.jim.silverton.at.verizon.not

Perhaps it was "fixed" to allow multi-threaded calculations, or to improve
performance of calculations in the new large grid. I don't think MS has the
resources to rewrite code for the sake of rewriting code.

- Jon
-------
Jon Peltier, Microsoft Excel MVP
Tutorials and Custom Solutions
Peltier Technical Services, Inc. - http://PeltierTech.com
_______


"James Silverton" wrote in message
...
Jim wrote on Fri, 28 Sep 2007 15:38:51 -0700:


JC And apparently discovered / reported by a " Molham Serry"
JC in the Public.Excel group on the 22nd.
JC http://tinyurl.com/338gjd (146 messages)
JC --
JC Jim Cone

I wasn't implying that I had discovered anything about the binary
representation except to wonder if it might be some sort of overflow
problem and a prime example of warning programmers "If it ain't broke
don't fix it".


James Silverton
Potomac, Maryland

E-mail, with obvious alterations: not.jim.silverton.at.verizon.not

Let's have a contest to see who can write the best algorithm to parse
the arithmetic with up to 64 levels of function nesting from a text
string. You first. We'll post it up on some OpenSource discussion
board for them to decide how sloppy it is.



On Sep 28, 6:06 pm, "James Silverton"
wrote:
Hello, All!

According tohttp://www.msnbc.msn.com/id/21033161/
it looks like Excel is having trouble either displaying or using
2^16. I'd really like to see the sloppy code that causes this!

James Silverton
Potomac, Maryland

E-mail, with obvious alterations:
not.jim.silverton.at.verizon.not

»iliace« wrote:
Let's have a contest to see who can write the best algorithm to parse
the arithmetic with up to 64 levels of function nesting from a text
string. You first. We'll post it up on some OpenSource discussion
board for them to decide how sloppy it is.

I doubt anyone will be able to replicate the bug... becaust it is not
caused by a sloppy parser or a sloppy calculation engine, but by a buggy
threshold value in some data representation decision that is primarily
done to save memory. So perhaps some == check was done with an epsilon
that should have been an exact ==.

Maybe reading the OOXML specs could give a hint about this
(specifically, reading how Excel stores numbers in files), but I don't
think so (I just expect OOXML to store numbers as text). The specs of
the binary XLS formats could more likely give it away, but these are not
public.

On Sep 28, 6:06 pm, "James Silverton" wrote:
Hello, All!

According tohttp://www.msnbc.msn.com/id/21033161/
it looks like Excel is having trouble either displaying or using
2^16. I'd really like to see the sloppy code that causes this!