On Sat, 09 May 2009 20:41:52 GMT, Owen Duffy wrote:

Jeff Liebermann wrote in
:

...
It's an INTENTIONAL Microsoft bug. See:

It is certainly popular to blame Microsoft with lots of things, whether
they were responsible or not matters little.

Sorry. Bad choice of wording. From my perspective, a bug is
something that works in a manner that would be considered unexpected
or fails some form of standards compliance. The order and precedence
of operators was well established and includes no distinction between
negation and subtraction. I cannot imagine an example where a
distinction would be necessary (although I am willing to be
enlightened). Whether MS can be blamed for creating the distinction
is subject to some debate, however I doubt that MS can be praised for
creating it.

Keep in mind that Microsoft did not 'design' the algebraic operator
hierarchy for Excel, Excel was released with a claim of 100% cell formula
compatibility with the then leading spreadsheet Lotus 123. (Microsoft's
compatibility was so good, it was subject of a famous court case.)

Chuckle. I once made good money cleaning up a mess of old Lotus 1-2-3
and Symphony spreadsheets so that they would run under Excel. This is
for Excel 2003:
http://office.microsoft.com/en-us/excel/HP051986941033.aspx
There are also comparisons between other version of Excel and Lotus
1-2-3, but I can't find them. Different versions seem to have
somewhat different differences, which makes me wonder if Excel has
"evolved" their standards.

Quoting:
"Lotus 1-2-3 evaluates the exponentiation operator (^) before
the negation operator (–); Excel evaluates negation first."
So it is written, so it must be. Web pages are never wrong.

It was much later that Microsoft conceived VBA and added it to their apps.
IIRC, Visual Basic for Applications inherits its algebraic operator
hierarchy from the BASIC language which was conceived around 1964 and
enriched progressively.

Yep. Even standards change with time. It's the de facto standard of
the moment.

The "intentional Microsoft bug" perspective looks like just prejudice.

Nope. I like Microsoft. If MS actually produced a bug free,
reliable, and fully functional product, I would be out of business. As
it stands, I expect to see considerable business from MS customers, as
new versions seem to introduce more features and functions, than fixes
to old bugs. After all, features and functions sell upgrades, but bug
fixes do not. 2.999999 cheers for Microsoft.

Incidentally, the company motto is "If this stuff worked, you wouldn't
need me". None of my customers have ever disagreed.

Owen


--
# Jeff Liebermann 150 Felker St #D Santa Cruz CA 95060
# 831-336-2558
# http://802.11junk.com
# http://www.LearnByDestroying.com AE6KS

On Sat, 9 May 2009 14:15:19 -0500, "Ralph Mowery"
wrote:

I think Intel had to recall a bunch of chips because of an error in the math
coprocessor part at one time.

Close. See:
http://en.wikipedia.org/wiki/Pentium_FDIV_bug
for the details. Never mind that the fixed chips arrived after the
release of the next generation of Pentium processors, making
replacements for the older and slower chips a waste of effort. I had
several servers running the buggy Pentium 60 and 66Mhz chips.
Incidentally, they ran unusually hot and required extra cooling. I
applied to Intel for 3 replacement chips. By the time they arrived,
the server motherboards had been replaced with something better and
faster, so the new chips just sat around.

Microsoft products are so full of 'problems' that if they ever put out an
error free product it would seem to be a mistake.

I beg to differ. Microsoft bashing seems to be the national sport in
computers. Yet, they're the most successful computah company in
history. In addition, they did it without any ties to proprietary
hardware. They must be doing something right.

In my never humble opinion, 99% or more of what MS releases is done
correctly and works well. The 1% that doesn't is what we're all
complaining about. Because MS has such a huge number of products and
technologies, it's fairly easy to find bugs and problems. However, if
you compare the MS bug lists with those from other companies, the
ratio of bugs to product complexity is very favorable for MS products.
I have had to deal with OS's and apps from smaller companies. Methinks
they're far worse than MS. Also, there may be plenty to complain
about, but most products are sufficiently functional to be usable for
the intended purpose. What bothers me about MS is not the quantity of
bugs, it's their tendency to add features and functions instead of
fixing bugs. This tends to make the product grow into a bloated
monstrosity of useless features, with far too many semi-permanent
bugs.



--
# Jeff Liebermann 150 Felker St #D Santa Cruz CA 95060
# 831-336-2558
# http://802.11junk.com
# http://www.LearnByDestroying.com AE6KS

On Sat, 09 May 2009 19:32:57 -0700, Jeff Liebermann
wrote:

http://en.wikipedia.org/wiki/Pentium_FDIV_bug

Also, the Foof bug:
http://en.wikipedia.org/wiki/F00f


--
# Jeff Liebermann 150 Felker St #D Santa Cruz CA 95060
# 831-336-2558
# http://802.11junk.com
# http://www.LearnByDestroying.com AE6KS

"Jeff Liebermann" wrote in message
...

I beg to differ. Microsoft bashing seems to be the national sport in
computers. Yet, they're the most successful computah company in
history. In addition, they did it without any ties to proprietary
hardware. They must be doing something right.


Microsoft got so big the same way Walmart did. They put out a cheeper
product.

Digital Research had a much beter product when IBM produced the PC. I think
they wanted about $ 150 for it and MS wanted $ 50 for their product. They
basically put DRI out of business and also some other companies that had
their ideas incorporated in to the MS product line.

On May 8, 12:36*pm, "Antonio Vernucci" wrote:
I think this message can be of interest for those using Excel for antenna
calculations.

Write in one cell: *=(-A1^2 + 8) * Note: the exponent of A1 is 2 (and not
2+8=10) because Excel performs squaring before summing

Write in another cell: *=(8 - A1^2)

They look pretty much the same

But give A1 any non-zero value and see what happen.

73

Tony I0JX
Rome, Italy

Computer Science 101 taught me that various computer compilers do not
always handle complex equations as I would expect and to break
equations down into steps using multiple lines of code. Compilers dont
even follow basic rules of mathmatics, if they did you would never be
able to code a = a + 1. They allow the programer to ASSIGN values to
variables and it is up to the programer to follow the rules.

Jimmie

JIMMIE wrote:

On May 8, 12:36Â*pm, "Antonio Vernucci" wrote:
I think this message can be of interest for those using Excel for antenna
calculations.

Write in one cell: Â*=(-A1^2 + 8) Â* Note: the exponent of A1 is 2 (and not
2+8=10) because Excel performs squaring before summing

Write in another cell: Â*=(8 - A1^2)

They look pretty much the same

But give A1 any non-zero value and see what happen.

73

Tony I0JX
Rome, Italy

Computer Science 101 taught me that various computer compilers do not
always handle complex equations as I would expect and to break
equations down into steps using multiple lines of code. Compilers dont
even follow basic rules of mathmatics, if they did you would never be
able to code a = a + 1.

The expression is an equation in mathematics and an assignment statement in
programming. Two different concepts on two different contexts. The meaning
of an expression can easily vary based on the context.

Consider 'x'; it can mean any of the following.
An abbreviation for a religious figure
An abbreviation for trans (xmit)
multiplication (2x3)
location of a pirate's treasure on a map
a kiss
one letter in a word

So, expecting a spreadsheet to follow the all rules of math is unrealistic.


The original post reflected confusion over operator precedence. Operator
precedence is usually documented. Not paying attention to how a spreadsheet
works is the user's fault. If there is doubt, parentheses can be added to
make sure the calculations are carried out as desired. Verifying the
calculations by hand is also a good idea.



They allow the programer to ASSIGN values to
variables and it is up to the programer to follow the rules.

Jimmie

On Sat, 9 May 2009 22:51:37 -0500, "Ralph Mowery"
wrote:


"Jeff Liebermann" wrote in message
.. .

I beg to differ. Microsoft bashing seems to be the national sport in
computers. Yet, they're the most successful computah company in
history. In addition, they did it without any ties to proprietary
hardware. They must be doing something right.


Microsoft got so big the same way Walmart did. They put out a cheeper
product.

They weren't always the biggest and baddest company in town. The
software departments of the major big iron makers were much larger
than MS in both manpower and revenue for most of the 1980's. Any one
of them could have produced a consumer grade operating system and
usable apps at the time and wiped MS off the map. They didn't because
they didn't believe that there was money to be made in essentially
consumer retail (i.e. off the shelf) operating systems and apps. They
also didn't know how to do it. I still recall the DEC Rainbow, where
customers were expected to buy pre-formatted floppies from DEC at
outrageous prices.

MS may also be very economical for OEM PC operating systems and
desktop apps. However, I note that a superior and totally free
operating system, while quite popular, has not produced much of a dent
in Microsoft's OS dominance. MS is also not currently the cheapest
OS. Apple OS/X Leopard retails for $130 while Vista Ultimate is $219.
Digital Research had a much beter product when IBM produced the PC.
I think
they wanted about $ 150 for it and MS wanted $ 50 for their product. They
basically put DRI out of business and also some other companies that had
their ideas incorporated in to the MS product line.

Yep. In 1981, CP/M-86 was better than PC-DOS 1.0. I was there.
CP/M-86 sold for $150. PC-DOS 1.0 sold for $60. Most of the early
IBM PC 5150 adopters bought both. I vaguely recall paying about
$4,000 for mine. $100 difference wasn't going to make a huge
difference.

CP/M-86 did more, but was more difficult to use. PC-DOS (er... QDOS)
was crude and simple. At the time everyone was waiting for DRI to
clean up the OS or at least make it more user friendly, while PC-DOS
was treated as a temporary expedient so IBM could sell PC's that were
suppose to run mostly apps in BASIC. Also note that PC-DOS included
MSBASIC, while CP/M-86 would sorta run the older CP/M-80 apps. CBASIC
came later. The IBM PC 5150 came with cassette BASIC in ROM. However
BASIC in ROM was not easily accessible from CP/M-86. Within months of
introduction, there were literally hundreds of new and ported apps for
PC-DOS arriving at Computerland. Meanwhile CP/M-86 was still
struggling with porting CP/M-80 apps. I had customers running some
bookkeeping application on CP/M-86 well into the late 1980's. It was
a struggle under CP/M-86. When they finally purged the machines and
switched to PCDOS, things went more smoothly. For example, relinking
the CP/M-86 operating system to install a new device driver was not my
idea of fun. With PC-DOS, it was just adding a line in config.sys.

All this has something to do with ham radio antennas, but the
connection escapes me for the moment.






--
Jeff Liebermann
150 Felker St #D http://www.LearnByDestroying.com
Santa Cruz CA 95060 http://802.11junk.com
Skype: JeffLiebermann AE6KS 831-336-2558

On Sun, 10 May 2009 16:19:22 +0000, Jim Higgins
wrote:

I disagree. Negation is not a subtraction operation; it's a
multiplication operation.

It varies by position. At the front of a string of arithmetic
operations, it's multiplication. In between the terms of an equation
or values, it's subtraction. This article covers some of the problem:
http://www.physicsforums.com/showthread.php?t=109516
Humans can usually make the distinction by context. Too bad computers
can't do the same.

Either way Microsoft Excel implements it incorrectly because in either
case exponentiation has a higher precedence in the science of
mathematics.

Agreed.

Agreed that it's too late to go back now, but bugs perpetuate because
they aren't fixed promptly when encountered. This issue isn't new and
it was fixable when first encountered in the very first release of
Excel.

I can sorta tolerate perpetuating mistakes. However, Microsoft's
attitude toward precedence operations reeks of damage control and of
trying to create a secondary standard by sheer number of users. The
first step to fixing the problem should have been to admit that they
were wrong. That never seems to have happened.

--
Jeff Liebermann
150 Felker St #D http://www.LearnByDestroying.com
Santa Cruz CA 95060 http://802.11junk.com
Skype: JeffLiebermann AE6KS 831-336-2558

On Sun, 10 May 2009 16:21:02 +0000, Jim Higgins
wrote:

I could be wrong, but I don't recall a case involving MS and Lotus
over Excel/123. I recall Lotus vs Borland over 1-2-3/Quattro
though... and that wasn't about details like precedence of operations

Nope. It was Borland versus Lotus over the look and feel of the
spreadsheet.
http://en.wikipedia.org/wiki/Lotus_v._Borland

There was also Apple versus Microsoft over the Windoze look and feel:
http://en.wikipedia.org/wiki/Apple_v._Microsoft

However, none of this had anything to do with the operation of the
spreadsheet. That's because the original spreadsheet operation was
defined by Visicalc:
http://en.wikipedia.org/wiki/Visicalc
Visicalc and spreadsheets were NOT patented and therefore not the
subject of litigation over function or operation:
http://techdirt.com/articles/20050812/1835229_F.shtml

Fun reading on who sued whom:
http://en.wikipedia.org/wiki/Apple_Inc._litigation
http://en.wikipedia.org/wiki/Microsoft_litigation


--
Jeff Liebermann
150 Felker St #D http://www.LearnByDestroying.com
Santa Cruz CA 95060 http://802.11junk.com
Skype: JeffLiebermann AE6KS 831-336-2558

Jim Higgins wrote in
:

....
I could be wrong, but I don't recall a case involving MS and Lotus
over Excel/123. I recall Lotus vs Borland over 1-2-3/Quattro

Lotus did not like Excel's compatibility with 123, and targeted
specifically the / command key in Excel as knocking off Lotus' IP. The /
key was removed in later versions, and IIRC the help sections that were
designed for 123 users.

Yes, those were the days of people suing people for too much looking and
feeling. Ashton Tate were right up there in the DBase lookalike race, but
pretty much everyone was into it, and just recently we hear that SCO
didn't really own the rights to their flavour of UNIX.

The early versions of Excel did have very good compatibility. One cannot
say that about OpenOffice, last time I looked it did not support VBA, and
I use VBA to a great extent in many of by spreadsheets, so not being
prepared to port and test all the content, the brave new OpenOffice is
for beginners who don't have a large investment in VBA code.

Irrespective of the heritage of these things, it is an inexperienced and
naive programmer who doesn't check the operator precedence for the
language environment being used.

Owen