Originally posted by twhiteheadEven the companies that make mother boards send the parts they think are
But you do agree that you do not need to talk to the designer to spot a design flaw.
flawed back to the company who made them so they can tell them what they
are seeing is indeed a flaw of some type, design or otherwise. You can say you
think you found one, that does not mean it is, then again you can say anything
that much is sure. Flaws are meaured in DPM defects per million so finding a
flaw is something, spotting a design flaw would be a very big deal.
Kelly
Originally posted by KellyJayI don't know why you continue to argue a point that you initially conceded. If anything it was your argument from the very beginning. You claimed that one would have to have designed something equivalent in order to spot a design flaw. Your initial challenge was to ask what life forms we had designed.
Even the companies that make mother boards send the parts they think are
flawed back to the company who made them so they can tell them what they
are seeing is indeed a flaw of some type, design or otherwise. You can say you
think you found one, that does not mean it is, then again you can say anything
that much is sure.
You even stated that you would trust AMDs opinion on Intels CPU designs.
Now you seem to be backing away from that stance and saying that it doesn't really matter whether we are designers, we still have to go back to the original designer.
Flaws are meaured in DPM defects per million so finding a
flaw is something, spotting a design flaw would be a very big deal.
Kelly
That is irrelevant. The types of defects you are referring too could in many cased be distinguished from design flaws. For example, manufacturing defects would almost never result in extra transistors sitting around where they shouldn't be.
Originally posted by gtbiking4life“…There are components on a CPU that are currently not connected to anything. certain pins come to mind. …”
There are components on a CPU that are currently not connected to anything. certain pins come to mind. In order to know without a doubt it is a design flaw, you would indeed have to talk with the designer. How else would you know if there is something you didn't think of? I disagree of course with you.
Yes, I am aware of that. And we know those pins are there so that the CPU more firmly grips onto its socket that it is to be plugged into. Since we can make a plausible hypothesis of the purpose of such pins, that doesn’t really count here as a definite flaw. And we can make such a plausible hypothesis without talking to the designer.
But, back to my example, what about a microscopically small transistor that is not connected to anything and is embedded deep within the circuitry of the CPU (so no practical way to access it from outside)? As I said before, you don’t have to “to talk to the designer” to know about how electricity works and to therefore be able to deduce that, all the time there are no electrical connections to a transistor, it isn’t going to do much! Would you dispute that fact?
If so, can you give a credible example of how such a transistor not connected to anything deep in a CPU can have a purpose?
“…How else would you know if there is something you didn't think of?...”
How would you know there is nothing (of relevance) you didn’t think of?
Also; if you like you can forget about the CPU example because I have a better example; better because there is nothing hypothetical about it because this actually happened! I gave this example to Kelly more than once but he refuses to respond to it presumably because he has no answer to it:
I remember a particular occasion when I showed a computer program I created that was not working properly to a student at a university I attended and he immediately spotted and point out that, in the very first line of my program. I had inserted in the letter ‘O’ in a integer literal where I should have inserted in the numerical digit ‘0’ in it and he spotted this flaw BEFORE I had a chance to even begin to explain to him what the program is for and how it is supposed to work. So he did NOT have much knowledge of my program! And yet he instantly and correctly spotted a flaw and then I corrected the flaw and found the program then worked perfectly thus confirming this WAS a flaw!
So, here is my question:
Would you deny that this is a clear example of something (a program in this case) that was designed (by me in this case) and somebody (another student in this case) CORRECTLY and rationally spotted and identified a flaw in it (the flaw being an invalid literal in this case)?
If you don’t refute this, then wouldn’t this prove you don’t have to talk to the designer to correctly and confidently spot a flaw?
Originally posted by Andrew HamiltonI'm going to work now but I'll reply later to your question. ๐
“…There are components on a CPU that are currently not connected to anything. certain pins come to mind. …”
Yes, I am aware of that. And we know those pins are there so that the CPU more firmly grips onto its socket that it is to be plugged into. Since we can make a plausible hypothesis of the purpose of such pins, that doesn’t really count here a ...[text shortened]... n’t this prove you don’t have to talk to the designer to correctly and confidently spot a flaw?
Originally posted by twhiteheadI conceded the point that someone would build a CPU to fit into a socket, as I pointed
I don't know why you continue to argue a point that you initially conceded. If anything it was your argument from the very beginning. You claimed that one would have to have designed something equivalent in order to spot a design flaw. Your initial challenge was to ask what life forms we had designed.
You even stated that you would trust AMDs opinion on ...[text shortened]... efects would almost never result in extra transistors sitting around where they shouldn't be.
out I was not thinking about a copy cat CPU where a company would force its designers
to use the foot print of their part to fit into the space someone one else came up
with, and to lay out their pin definitions to the layout someone else came up with.
That said, if they screw that up, it would be easier to spot a flaw; however, I doubt
anyone who would go through the trouble of doing that would put a ground where
it shouldn’t be a so on.
My point about designers was and still is even a design engineer is not as gifted as
you claim to be! You said you could spot a design flaw by looking at the part, I have
been telling you that design engineers require a battery of tests to find them, that is
many of the design engineers sole reason for being, that is all they focus on and they require
data galore, they require high-end testing equipment, they require a battery of tests
to identify and resolve those types of flaws, unlike you who has claimed by looking
at the part can spot them.
Kelly
Originally posted by KellyJayAnd that is not the point I was referring to.
I conceded the point that someone would build a CPU to fit into a socket, ....
..... I have been telling you that design engineers require a battery of tests to find them......
So you are essentially saying that this group of engineers, with a battery of tests are capable, without questioning the initial designer, to spot a flaw. ie you concede the point that questioning the initial designer is not a requirement.
So when I posted:
But you do agree that you do not need to talk to the designer to spot a design flaw.
You could simply have responded in the affirmative, but you chose not to, even though you do apparently agree with the point.
And you continue to apparently challenge it without actually contradicting it. Its really a straight forward 'yes' or 'no' assertion. Either the original designers opinion is required or it is not.
Originally posted by Andrew HamiltonI'll reply more later - all I'll say for now is I could easily spot that flaw in a program as well without knowing what the program is for and how it is supposed to work. The student knew at least a little about programming and could understand what an integer is and how variables are declared. It certainly would not take a genius to figure it out. All someone needs to know is a little about programming and can spot a simple mistype. They don't even have to know what the program did to point that out. I'm not sure what this is supposed to prove honestly.
“…There are components on a CPU that are currently not connected to anything. certain pins come to mind. …”
Yes, I am aware of that. And we know those pins are there so that the CPU more firmly grips onto its socket that it is to be plugged into. Since we can make a plausible hypothesis of the purpose of such pins, that doesn’t really count here a n’t this prove you don’t have to talk to the designer to correctly and confidently spot a flaw?
I have to ask how you can spot a microscopically small transistor that is not connected to anything. Do you take a microscope with you when you buy a CPU? If not than you would not be able to point out that flaw.
I believe my point still stands and that is you would not be able to point that flaw out unless you had a very good knowledge of CPUs or the designer told you. You would also need a microscope ๐
Your program had a very simple mistake to notice and point out. You mind as well ask me if I'm standing next to a cliff.
Originally posted by twhiteheadInput required.
And that is not the point I was referring to.
[b]..... I have been telling you that design engineers require a battery of tests to find them......
So you are essentially saying that this group of engineers, with a battery of tests are capable, without questioning the initial designer, to spot a flaw. ie you concede the point that questioning the i ...[text shortened]... ard 'yes' or 'no' assertion. Either the original designers opinion is required or it is not.[/b]
KJ
Originally posted by twhiteheadLook at it this way, say I have a batch of parts (DUT) devices under test, and you
And that is not the point I was referring to.
[b]..... I have been telling you that design engineers require a battery of tests to find them......
So you are essentially saying that this group of engineers, with a battery of tests are capable, without questioning the initial designer, to spot a flaw. ie you concede the point that questioning the i ...[text shortened]... ard 'yes' or 'no' assertion. Either the original designers opinion is required or it is not.[/b]
see them placed on a piece of test equipment and you find the whole batch of
parts all fail for shorts. Can you say that was due to a design flaw, or maybe
something else? What if you put a CPU on a mother board and it shorts out, you
put another on the same mother board and it shorts out, what do you know?
Kelly
Originally posted by gtbiking4life“….all I'll say for now is I could easily spot that flaw in a program as well without knowing what the program is for and how it is supposed to work…..”
I'll reply more later - all I'll say for now is I could easily spot that flaw in a program as well without knowing what the program is for and how it is supposed to work. The student knew at least a little about programming and could understand what an integer is and how variables are declared. It certainly would not take a genius to figure it out. All ...[text shortened]... e mistake to notice and point out. You mind as well ask me if I'm standing next to a cliff.
So you would answer “no” to that question; you would not deny it just like I would not deny it; that is a relief!
I am not sure if Kelly would not deny it though! he does not answer; can you think why?
“….I'm not sure what this is supposed to prove honestly….”
It proves Kelly’s argument wrong. His argument was that you could not correctly spot a design flaw in a giraffe’s neck without consulting the designer even if it was a really obvious stupid flaw such as like the dumb mistake I made in that computer program. Surely a “no” answer to my question would logically contradict that stance? –don’t you think? If you can correctly spot a really stupid flaw in my program without consulting me then why couldn’t you correctly spot a really stupid flaw in the giraffe’s neck?
“….I have to ask how you can spot a microscopically small transistor that is not connected to anything. Do you take a microscope with you when you buy a CPU? If not than you would not be able to point out that flaw…..”
The ability to point out a flaw isn’t relevant here. A flaw is still a flaw even if it is practically impossible to spot.
I don’t carry around a microscope but, in the very unlikely event that I did and in the even more unlikely event of me having the means to systematically strip away the circuitry one microscopic layer at a time to see what is underneath (not even sure what equipment I would need for that! ), then, hypothetically, I could spot such a flaw.
The CPU example is not a good example for several reasons –better to stick to my program example ๐
Incidentally, we don’t need a microscope to spot the flaw in a giraffe’s neck because that nerve can clearly be seen with the unaided eye so this problem with the CPU example couldn’t work against the giraffe’s neck example anyway.
“…I believe my point still stands and that is you would not be able to point that flaw out unless you had a very good knowledge of CPUs or the designer told you…..”
So tell me, in your opinion, what possible reason would a designer have to place a transistor deep within the CPU circuitry where it could never be accessed from outside and for it to be connected to absolutely nothing?
“…Your program had a very simple mistake to notice and point out….”
Exactly! As simple and obvious as the flaw in the giraffe’s neck!
Originally posted by KellyJayIt may not be a design flaw.
Look at it this way, say I have a batch of parts (DUT) devices under test, and you
see them placed on a piece of test equipment and you find the whole batch of
parts all fail for shorts. Can you say that was due to a design flaw, or maybe
something else? What if you put a CPU on a mother board and it shorts out, you
put another on the same mother board and it shorts out, what do you know?
Kelly
But you are still not being at all clear in your answer.
Are you saying that if you have no access to the original designer of a CPU, you can never ever spot a design flaw?
Originally posted by twhiteheadThis is what knowing why the part is laid out the way it is does for you to name a few.
It may not be a design flaw.
But you are still not being at all clear in your answer.
Are you saying that if you have no access to the original designer of a CPU, you can never ever spot a design flaw?
You know what is there and why, or what is supposed to be there and why.
You know what is supposed to happen and not.
You know what is to be expected in all conditions.
You know what the conditions are supposed to be each time you test.
You know what the tolerances are supposed to be each time you test.
Without that type of knowledge you’re guessing, EVEN WITH this knowledge you
have to setup tests to reveal what is really going on, you have to make sure that
while you’re testing the test environment is exactly like it is supposed to be
without varying conditions out of spec. If you make a bad assumption on seeing
test results and you would't know that for example you were testing at one voltage
or temp instead of another all your results are not what you think they mean.
There is a lot to spotting a design flaw and all other types so your and Andrew
claims just by looking at a part you two can see a design flaw isn't something I
agree with.
Even reading a document on what a part is supposed to do, ONLY means that the
part and the document disagree, is it the part or the document that is flawed and
if there is a flaw, is it a design issue or some other condition in play your not
aware of yet?
Kelly
Originally posted by KellyJay“….There is a lot to spotting a design flaw and all other types so your and Andrew
This is what knowing why the part is laid out the way it is does for you to name a few.
You know what is there and why, or what is supposed to be there and why.
You know what is supposed to happen and not.
You know what is to be expected in all conditions.
You know what the conditions are supposed to be each time you test.
You know what the tolerance ...[text shortened]... is a flaw, is it a design issue or some other condition in play your not
aware of yet?
Kelly
claims just by looking at a part you two can see a design flaw isn't something I
agree with…..”
You can forget the CPU example: My example of a correctly identified flaw in my program that you so far ignored completely debunks all your arguments here. The other student correctly identified that flaw without knowing what my program does nor how it works nor much else! gtbiking4life basically agreed that those are the facts of what happened.
Originally posted by Andrew HamiltonIf I were talking about your program, it would be a different subject. I'm not
“….There is a lot to spotting a design flaw and all other types so your and Andrew
claims just by looking at a part you two can see a design flaw isn't something I
agree with…..”
You can forget the CPU example: My example of a correctly identified flaw in my program that you so far ignored completely debunks all your arguments here. The other st ...[text shortened]... orks nor much else! gtbiking4life basically agreed that those are the facts of what happened.
saying you HAVE to have designer knownledge in ALL things, but at some point
it is required. If you have a program where everyone knows the code, everyone
knows what it is there for and why, that is different than something so complex
that even with designer knowledge you still need more help.
Kelly
Originally posted by Andrew HamiltonI said I wasn't finished ๐
“….all I'll say for now is I could easily spot that flaw in a program as well without knowing what the program is for and how it is supposed to work…..”
So you would answer “no” to that question; you would not deny it just like I would not deny it; that is a relief!
I am not sure if Kelly would not deny it though! he does not answer; can you thin o notice and point out….”
Exactly! As simple and obvious as the flaw in the giraffe’s neck!
Well you are comparing apples to oranges with your computer program example. The problem is the program was not working, so it obviously had a flaw.
We are discussing something that is already working just fine. There is a difference. I believe my point still stands that in order to know what a flaw is or is not in something that is working fine like it is, you would need a designer's input.