Message that M43112609 is prime! Wd1: 5B377150,00000000 - Page 2

petrw1 · 2012-01-29, 20:06

Quote:

Originally Posted by xilman

I could make a reasonable claim that this Dartmouth BASIC program

Code:

10 REM NO BUGS
20 END

is bug free.

Paul

Depends on how far you want to take it.
Based on today's implementation of BASIC I will take your word for it that it is syntactically correct and **should** function as expected:

However:
- What guarantees do you have that the OS you run this on is bug free?
- What guarantees do you have that the HW is bug free?
- What guarantees do you have that the Dartmouth developers won't make a change to BASIC in the future (intentionally or OMG a bug of their own) that could cause your program to NOT work the same?

Christenson · 2012-01-30, 13:17

Just having lost a computer power supply, shorted out with a chip smoked by a fuzzy caterpillar-like bug, I assure you, no computer program is guaranteed bug free.

Further, while it *may* be possible to write error-free programs, I think stating "bug free" is a little like measuring the amount of cash in an economy as a primary driver in the economy. It oversimplifies things.

By way of evidence, I cite the Pentium FDIV bug. There was a reward out there, in a place with some extremely intelligent folks, to find someone actually affected by this widespread problem. It went unawarded -- the FDIV bug affected the 4th or 5th significant figure, and the vast majority of us (computational number theorists are a distinct and small minority) don't have measurements of physical quantities that precise to begin with, and if the fifth significant figure matters, the problem has a very poor condition number, and any decent numerical modeler will find this out anyway.

And given Peterw1's qeustions, I think it is only possible to discuss systems that are asymptotically bug free...or highly reliable and robust, as it were. Yes, we can get to the point where no more bugs are found in the source code, and issues are more often due to hardware, but there's a raft of other probems, even if the source code is, in some sense, perfect:
1) Was the specification for the program correct?
2) Is my hardware and OS going to react properly to disturbances, such as cosmic rays, random errors, and loss of power?
3) How would a frankly complex program/computing system detect that it has misbehaved? (Note that P95 hass done significant work in that regard, but mfaktc has not)
4) What about leaking memory? In the system that runs the program? Will this force an eventual shutdown? In an interval that matters? (Salesmen of ferroelectric ram like to cite the wearout mechanism as becoming important in 1700 years, if the device is written to constantly, every few millseconds or so).

xilman · 2012-01-30, 17:25

Quote:

Originally Posted by petrw1

Depends on how far you want to take it.
Based on today's implementation of BASIC I will take your word for it that it is syntactically correct and **should** function as expected:

However:
- What guarantees do you have that the OS you run this on is bug free?
- What guarantees do you have that the HW is bug free?
- What guarantees do you have that the Dartmouth developers won't make a change to BASIC in the future (intentionally or OMG a bug of their own) that could cause your program to NOT work the same?

I neither have nor make any guarantees of that nature.

Equally, I make no guarantee that I won't quantum tunnel to Mars some time tomorrow.

xilman · 2012-01-30, 17:33

Quote:

Originally Posted by petrw1

Depends on how far you want to take it.
Based on today's implementation of BASIC I will take your word for it that it is syntactically correct and **should** function as expected:

However:
- What guarantees do you have that the OS you run this on is bug free?
- What guarantees do you have that the HW is bug free?
- What guarantees do you have that the Dartmouth developers won't make a change to BASIC in the future (intentionally or OMG a bug of their own) that could cause your program to NOT work the same?

Going into more detail:

I would be surprised if any OS on which a Dartmouth BASIC interpreter has ever run has been bug free.

I would be even more surprised if any hardware on which a Dartmouth BASIC interpreter has ever run has been bug free.

As for the Dartmouth developers making a change of the nature you suggest, surprise is no longer adequate and I am forced to resort to astonishment(*).

Paul

(*) At this point, I suggest that you start Googling. You should find at least two distinct items of interest in the third claim.

Christenson · 2012-01-31, 13:47

But Paul....the first question is where do I *get* an implementation of "Dartmouth" BASIC these days???? I somehow *expect* it to have a *very high* degree of support!!!!! (grin)

What??!!!??? It was only available on megauploads.com?????!*!?!?

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
.dat file message	mnh001	YAFU	7	2015-02-07 07:37
TF: Wd2: 00000000	tichy	Software	6	2010-12-01 18:53
Log Out Message	storm5510	Forum Feedback	3	2009-08-15 07:11
PRIME.INI question to address error message	JHagerson	Software	1	2006-01-26 18:34
error message help?	AurKayne	Hardware	3	2005-08-29 09:13

2012-01-30, 13:17	#13
Christenson Dec 2010 Monticello 5×359 Posts	Just having lost a computer power supply, shorted out with a chip smoked by a fuzzy caterpillar-like bug, I assure you, no computer program is guaranteed bug free. Further, while it may be possible to write error-free programs, I think stating "bug free" is a little like measuring the amount of cash in an economy as a primary driver in the economy. It oversimplifies things. By way of evidence, I cite the Pentium FDIV bug. There was a reward out there, in a place with some extremely intelligent folks, to find someone actually affected by this widespread problem. It went unawarded -- the FDIV bug affected the 4th or 5th significant figure, and the vast majority of us (computational number theorists are a distinct and small minority) don't have measurements of physical quantities that precise to begin with, and if the fifth significant figure matters, the problem has a very poor condition number, and any decent numerical modeler will find this out anyway. And given Peterw1's qeustions, I think it is only possible to discuss systems that are asymptotically bug free...or highly reliable and robust, as it were. Yes, we can get to the point where no more bugs are found in the source code, and issues are more often due to hardware, but there's a raft of other probems, even if the source code is, in some sense, perfect: 1) Was the specification for the program correct? 2) Is my hardware and OS going to react properly to disturbances, such as cosmic rays, random errors, and loss of power? 3) How would a frankly complex program/computing system detect that it has misbehaved? (Note that P95 hass done significant work in that regard, but mfaktc has not) 4) What about leaking memory? In the system that runs the program? Will this force an eventual shutdown? In an interval that matters? (Salesmen of ferroelectric ram like to cite the wearout mechanism as becoming important in 1700 years, if the device is written to constantly, every few millseconds or so).

2012-01-31, 13:47	#16
Christenson Dec 2010 Monticello 703₁₆ Posts	But Paul....the first question is where do I get an implementation of "Dartmouth" BASIC these days???? I somehow expect it to have a very high degree of support!!!!! (grin) What??!!!??? It was only available on megauploads.com?????!*!?!?