I'm not saying the current sensors caused a problem. But:
For that matter, how could he screw up the circuit in such a way as to blow the motor without blowing it up immediately? You sure can't wire the circuit to 220 volts without doing something very funny with the common wire.
Imagine a long run of conduit from a panel to a room. In the conduit are one netural and two hots, on opposite phases. In the room are two machines - one, on the first phase, draws 20 amps. The other, on the second phase, draws 1 amp.
If the neutral wire has 1 ohm resistance, the first machine sees 101 volts and the second sees 139. Many variations on this ideas are possible.
My particular issue with C++ is that which BS seems to be warning against -
unlike him I already think there's far too much in the C++ language that
should properly have gone into the standard library.
I understand most of your wishlist. But how do you survive without copy constructors? Do we always default to bitcopy? What if there are pointers in the object?
What I meant was, PLCs lend themselves to verification. Maybe that's the key to building other kinds of verifiable systems - write a lower layer that's reusable and QA it heavily; then write the upper layer which is somewhat self-evidently verifiable.
I posted another comment about this, but it's almost certainly a PLC programmed in ladder logic. Ladder logic is mostly equivalent to a list boolean equations defining outputs of the PLC. So if we state, "output17=input3 and input15 and (not input 11)", it's pretty straightforward to understand when output 17 will be turned on. Compare that to a procedural program, where many lines of code could affect output17.
So yes, each and every line has been verified, but not much math is involved. The engineers will calculate the effect of each possible failure, such as a sensor wire breaking, and make sure it leads to an E-STOP.
Generally in machine design (which ride design is part of) it is always OK for the controller to send a 0 on an output. OK, meaning it may impair functionality, but won't cause a disaster. It is not OK to send a 1 when a 0 is needed. An E-STOP causes all outputs to become 0, usually by a contactor disconnecting 24v power from the PLC.
This makes a machine/ride slightly simpler to design than a missile, for instance, because a missile has no "safe state" once it launches.
In Europe, most ride control is done with relays. This actually makes sense - you only need about one relay per block zone. A relay control system has less room for bugs and hidden effects. Remember, people get injured or killed on coasters, and engineers will have to testify in court as to the system design and whether the system still adhered to the design when the accident occurred.
A PLC typically simulates relay logic via "ladder logic" programming. This means that a mechanical engineer with no programming knowledge can view the ladder logic and understand all failure modes.
PLCs are an excellent fit for their problem space, which is primarily machine control.
Thanks - that was interesting. However, you seem to imply that a successful company can't have any non-obvious values. I will argue that Microsoft has at least one: "eat your own dogfood". By taking the short-term pain of using their own horrible products, they've enjoyed long-term gain, as feedback inside the company will be better informed and better heard than outsider feedback.
That page is such utter BS. The author is mad that Scott Adams doesn't harness his strip to promote some tired leftist ideology. Guess what? If he did, the strip would not be as popular. Dilbert is aimed at the educated middle-income knowledge worker. We don't generally share the views on that page. In fact, some of these leftist views have become part of the ridiculous PC fabric of the corporation. For example, the whiner mentions "sexual harrassment." I've seen huge fusses over sexual harrassment. I've seen good people fired for viewing porn after hours, under the guise of sexual harrassment. I've yet to see a valid case of sexual harrassment in the corporate world. Not saying it doesn't happen - just saying this whiner wants to move Dilbert further away from accurately depicting what I see.
Look, whiner, make your own cartoon about your favorite leftist memes. Instead of the pointy haired boss, you can have a top-hatted capitalist. Don't be surprised when nobody reads your strip because you completely failed to pick up on what a modern office is like.
I agree, except I must have missed the good years. Every time I saw that thing, I reread it to make sure I wasn't missing the point. No, it was as stupid and pointless as I thought.
But then, I think all newspaper comics are utter crap, except for Dilbert. They all leave me wondering how someone with at least one neuron firing bothered to create something so inane.
Sure, a few pinko liberals might have complained...
If we're talking about Stalin, it was the pinkos (communist sympathisers) who took his side in the west. People like Walter Duranty cheered him on and covered up his crimes.
OK, I can't resist. What happens when some nation, such as Iran, develops a suitcase nuke for use by ignorant terrorists. It is designed to be super-stable and maintenance free until deliberately detonated. A civilian collector gets his hand on one. Should it be illegal to possess?
Remember, some civilian collectors have 50 cal machine guns, which are devestating weapons, but they are rarely or never used in crime.
Must be bad universities, I for one had the luck to attend a lot of very good and deep courses about system programming, computer architecture and digital design.
I'm glad you had a good education. Unfortunately, there are a lot of CS majors emerging these days with a Java-only education. Certain universities are notorious for this. You might think that once someone knows CS, learning additional languages is trivial. But this is not really true, especially if he only knows one language. It's also easier to go from C to java than java to C.
The reduction of the use of C in our profession marks the decline of real knowledge in the average programmer.
Speaking just for Silicon Valley, I can't accept your gloom-and-doom scenario. C is very important right now. Even if a job does not involve C coding, C is often used as a measuring stick in the interview. I think a lot of people are using C to screen out the semi-skilled "web engineers" who arrived during the boom.
OTOH, if you only know C, your opportunities are limited.
OK, maybe you're right about the average programmer, but the number of programmers has increased. I'm not upset if there's a huge number of java-only programmers writing internal apps at insurance companies.
I agree that the candidate should ask questions to learn about the company - with some caveats. Don't ask questions which the web could answer - you should already know. Ask questions later in the process, after you've proven some desirability. And show some sensitivity to which group is interviewing you. It's annoying to sound like you're more interested in a different part of the company.
But none of this says an employer is obligated to "sell" to a candidate.
No one tried to sell him on either the company or the job, he said. He withdrew his application.
Why does this guy expect to be "sold"? I think it's the candidate's job to learn about the company. Anything a recruiter says is just a sales pitch.
I just finished a job search. I got called by many recruiters and went through many phone screens. I didn't feel surprised or offended by that - obviously the first step is sto phone screen a candidate.
Sorkin seems to think that if Microsoft is calling him, they must have already made the decision to hire. Not so. Even if you locate a strong-seeming candidate, it's unwise to skip any steps in interviewing.
Then Sorkin seems offended that Microsoft called him again. Why? People's situations change. It makes sense for an employer to keep phone numbers and try later. Maybe there was bad chemistry with one group but there could be good chemistry with another.
I guess the news story is that Sorkin is offended by totally ordinary hiring practices.
Thanks. I am getting sick of this simplistic urban legend. But I think there's more to it.
When you design a cast metal part, you have to make sure the molten metal will reach all the corners. The easiest thing to cast is a sphere; the hardest is something with thin parts sticking off. It's probably easier to cast round manhole covers with a lower reject rate.
Also, sewer and storm drain manholes are generally shaped like big bottles. A skinny neck to enter through, and a large body to work inside. Given a round neck, wouldn't it be awkward to attach a square frame? (Power and phone manholes are usually rectangular, with short round necks supporting the mh cover frame.)
And also, a precast manhole may be installed at an angle to the center line of a street. The round cover disguises this, but a square cover would look odd if rotated from the CL.
This is a tricky question, and I think both the candidate and you missed the key. The key is design. You have to elicit the requirements from the interviewer and design around them. Talking about WIFI at the end is just a way of saying, "you forgot to ask me what I want." This question tests whether you realize that design must be responsive to requirements. Most geeks don't.
I agree. I don't know any unemployed programmers in the valley right now. In fact, most of my friends are looking for good programmers to hire. My most unemployable friend (a sysadmin who never smiles) just got a job.
I'm not sure I blame political correctness - I just don't think NYT has picked up on that story. I do think they enjoy reporting some bad news about the economy.
OK, I don't think I've read any business method patents. But I do hope that when you read them, you understand some basics. The scope of a patent is determined by its claims. If you don't infringe at least one claim, you don't infringe the patent, no matter what is said elsewhere in the patent.
To infringe a claim, you must basically have every element in the claim. (There are exceptions.)
Focus on the independent claims - those that do not refer to another claim. You can't infringe a dependent claim without also infringing an independent claim.
I think that this lens will reduce the number of "broad and stupid" patents you see.
I live in Silicon Valley. It seems like everyone is hiring right now. Startups are springing up like mushrooms and huge companies are funding new efforts. In my recent job search, I got five offers locally. But you have to know your stuff. Most companies grilled me extensively on languages, databases and algorithms. It's fortunate that my ongoing interest in computers led me to keep learning while I was working at my last job.
We are phonescreening a lot of candidates, and almost all are unsatisfactory.
To combine my experience with the idea that "employment is not rising," I guess that the Valley still has many unqualified programmers. As companies get better at screening, these people will be unemployed more often.
If you are a good programmer with the right skills, the Valley is a very exciting place to be right now.
I call bullshit on your calling bullshit. Slashdot has created a mythology about software patents - the mythology evaporates if you learn the basics of patents and read the patents under discussion.
Slashdot nearly always overstates the breadth of a patent's claims.
I have read several file wrappers and I feel that the PTO does a very good job of prior art search. Not "investigation" - they are not required to "investigate", only to search their own records.
Not to mention that the slashdot understanding of a patent's claims is almost always wrong. It is based on the headline. For a while I read every patent application that upset slashdot and was astonished at how wrong slashdot is. Now I don't bother.
Whatever this patent covers, it's definitely not what slashdot says.
While I partially agree, I have to mention the counterpoint. MySQL is very good across a certain domain of problems. Where I work now, we pick between Oracle and MySQL for specific projects. If we don't see transactional requirements, writer contention, or huge volumes of data, we pick MySQL. It's faster and generally easier to work with.
I just don't care that MySQL isn't SQL-92 compliant. By the way, none of the applications I have worked on on Oracle could be trivially ported to a different db, standards or no. I agree that MySQL is friendlier than other dbs. Most obvious example - meaningful error messages.
I am not excited by Innodb or most of the features being added to MySQL. These changes, which add Oracle-like capabilities, also add Oracle-like problems, such as complexity, slowness, lock buildups, etc. We already have Oracle. We use MySQL where it fits.
As for hammers and nails: the guy with only one tool has a problem; so does the guy who spends all his time learning new tools and never becomes expert with one. There is not a logical niche for every technology - many of them are not worth touching, period. It's better to invest time in learning a few tools deeply.
But slashdot told me that IP laws only apply in the US, and by passing and enforcing them the US will just be at a disadvantage compared to "the rest of the world".
Could it be that all civilized countries act together in these matters? Could slashdot really be wrong?
Slashdot Mirror
Imagine a long run of conduit from a panel to a room. In the conduit are one netural and two hots, on opposite phases. In the room are two machines - one, on the first phase, draws 20 amps. The other, on the second phase, draws 1 amp.
If the neutral wire has 1 ohm resistance, the first machine sees 101 volts and the second sees 139. Many variations on this ideas are possible.
For example?
I understand most of your wishlist. But how do you survive without copy constructors? Do we always default to bitcopy? What if there are pointers in the object?
What I meant was, PLCs lend themselves to verification. Maybe that's the key to building other kinds of verifiable systems - write a lower layer that's reusable and QA it heavily; then write the upper layer which is somewhat self-evidently verifiable.
I posted another comment about this, but it's almost certainly a PLC programmed in ladder logic. Ladder logic is mostly equivalent to a list boolean equations defining outputs of the PLC. So if we state, "output17=input3 and input15 and (not input 11)", it's pretty straightforward to understand when output 17 will be turned on. Compare that to a procedural program, where many lines of code could affect output17.
So yes, each and every line has been verified, but not much math is involved. The engineers will calculate the effect of each possible failure, such as a sensor wire breaking, and make sure it leads to an E-STOP.
Generally in machine design (which ride design is part of) it is always OK for the controller to send a 0 on an output. OK, meaning it may impair functionality, but won't cause a disaster. It is not OK to send a 1 when a 0 is needed. An E-STOP causes all outputs to become 0, usually by a contactor disconnecting 24v power from the PLC.
This makes a machine/ride slightly simpler to design than a missile, for instance, because a missile has no "safe state" once it launches.
I don't know anything about this coaster, but most ride control in the US is done with PLCs and programmed in ladder logic. Here's a picture of an Allen Bradley PLC from some wastewater treatment page.
In Europe, most ride control is done with relays. This actually makes sense - you only need about one relay per block zone. A relay control system has less room for bugs and hidden effects. Remember, people get injured or killed on coasters, and engineers will have to testify in court as to the system design and whether the system still adhered to the design when the accident occurred.
A PLC typically simulates relay logic via "ladder logic" programming. This means that a mechanical engineer with no programming knowledge can view the ladder logic and understand all failure modes.
PLCs are an excellent fit for their problem space, which is primarily machine control.
Thanks - that was interesting. However, you seem to imply that a successful company can't have any non-obvious values. I will argue that Microsoft has at least one: "eat your own dogfood". By taking the short-term pain of using their own horrible products, they've enjoyed long-term gain, as feedback inside the company will be better informed and better heard than outsider feedback.
That page is such utter BS. The author is mad that Scott Adams doesn't harness his strip to promote some tired leftist ideology. Guess what? If he did, the strip would not be as popular. Dilbert is aimed at the educated middle-income knowledge worker. We don't generally share the views on that page. In fact, some of these leftist views have become part of the ridiculous PC fabric of the corporation. For example, the whiner mentions "sexual harrassment." I've seen huge fusses over sexual harrassment. I've seen good people fired for viewing porn after hours, under the guise of sexual harrassment. I've yet to see a valid case of sexual harrassment in the corporate world. Not saying it doesn't happen - just saying this whiner wants to move Dilbert further away from accurately depicting what I see.
Look, whiner, make your own cartoon about your favorite leftist memes. Instead of the pointy haired boss, you can have a top-hatted capitalist. Don't be surprised when nobody reads your strip because you completely failed to pick up on what a modern office is like.
I agree, except I must have missed the good years. Every time I saw that thing, I reread it to make sure I wasn't missing the point. No, it was as stupid and pointless as I thought.
But then, I think all newspaper comics are utter crap, except for Dilbert. They all leave me wondering how someone with at least one neuron firing bothered to create something so inane.
If we're talking about Stalin, it was the pinkos (communist sympathisers) who took his side in the west. People like Walter Duranty cheered him on and covered up his crimes.
OK, I can't resist. What happens when some nation, such as Iran, develops a suitcase nuke for use by ignorant terrorists. It is designed to be super-stable and maintenance free until deliberately detonated. A civilian collector gets his hand on one. Should it be illegal to possess?
Remember, some civilian collectors have 50 cal machine guns, which are devestating weapons, but they are rarely or never used in crime.
I'm glad you had a good education. Unfortunately, there are a lot of CS majors emerging these days with a Java-only education. Certain universities are notorious for this. You might think that once someone knows CS, learning additional languages is trivial. But this is not really true, especially if he only knows one language. It's also easier to go from C to java than java to C.
Speaking just for Silicon Valley, I can't accept your gloom-and-doom scenario. C is very important right now. Even if a job does not involve C coding, C is often used as a measuring stick in the interview. I think a lot of people are using C to screen out the semi-skilled "web engineers" who arrived during the boom.
OTOH, if you only know C, your opportunities are limited.
OK, maybe you're right about the average programmer, but the number of programmers has increased. I'm not upset if there's a huge number of java-only programmers writing internal apps at insurance companies.
I agree that the candidate should ask questions to learn about the company - with some caveats. Don't ask questions which the web could answer - you should already know. Ask questions later in the process, after you've proven some desirability. And show some sensitivity to which group is interviewing you. It's annoying to sound like you're more interested in a different part of the company.
But none of this says an employer is obligated to "sell" to a candidate.
Why does this guy expect to be "sold"? I think it's the candidate's job to learn about the company. Anything a recruiter says is just a sales pitch.
I just finished a job search. I got called by many recruiters and went through many phone screens. I didn't feel surprised or offended by that - obviously the first step is sto phone screen a candidate.
Sorkin seems to think that if Microsoft is calling him, they must have already made the decision to hire. Not so. Even if you locate a strong-seeming candidate, it's unwise to skip any steps in interviewing.
Then Sorkin seems offended that Microsoft called him again. Why? People's situations change. It makes sense for an employer to keep phone numbers and try later. Maybe there was bad chemistry with one group but there could be good chemistry with another.
I guess the news story is that Sorkin is offended by totally ordinary hiring practices.
Thanks. I am getting sick of this simplistic urban legend. But I think there's more to it.
When you design a cast metal part, you have to make sure the molten metal will reach all the corners. The easiest thing to cast is a sphere; the hardest is something with thin parts sticking off. It's probably easier to cast round manhole covers with a lower reject rate.
Also, sewer and storm drain manholes are generally shaped like big bottles. A skinny neck to enter through, and a large body to work inside. Given a round neck, wouldn't it be awkward to attach a square frame? (Power and phone manholes are usually rectangular, with short round necks supporting the mh cover frame.)
And also, a precast manhole may be installed at an angle to the center line of a street. The round cover disguises this, but a square cover would look odd if rotated from the CL.
This is a tricky question, and I think both the candidate and you missed the key. The key is design. You have to elicit the requirements from the interviewer and design around them.
Talking about WIFI at the end is just a way of saying, "you forgot to ask me what I want."
This question tests whether you realize that design must be responsive to requirements. Most geeks don't.
I agree. I don't know any unemployed programmers in the valley right now. In fact, most of my friends are looking for good programmers to hire. My most unemployable friend (a sysadmin who never smiles) just got a job.
I'm not sure I blame political correctness - I just don't think NYT has picked up on that story. I do think they enjoy reporting some bad news about the economy.
OK, I don't think I've read any business method patents. But I do hope that when you read them, you understand some basics. The scope of a patent is determined by its claims. If you don't infringe at least one claim, you don't infringe the patent, no matter what is said elsewhere in the patent.
To infringe a claim, you must basically have every element in the claim. (There are exceptions.)
Focus on the independent claims - those that do not refer to another claim. You can't infringe a dependent claim without also infringing an independent claim.
I think that this lens will reduce the number of "broad and stupid" patents you see.
I live in Silicon Valley. It seems like everyone is hiring right now. Startups are springing up like mushrooms and huge companies are funding new efforts. In my recent job search, I got five offers locally. But you have to know your stuff. Most companies grilled me extensively on languages, databases and algorithms. It's fortunate that my ongoing interest in computers led me to keep learning while I was working at my last job.
We are phonescreening a lot of candidates, and almost all are unsatisfactory.
To combine my experience with the idea that "employment is not rising," I guess that the Valley still has many unqualified programmers. As companies get better at screening, these people will be unemployed more often.
If you are a good programmer with the right skills, the Valley is a very exciting place to be right now.
I call bullshit on your calling bullshit. Slashdot has created a mythology about software patents - the mythology evaporates if you learn the basics of patents and read the patents under discussion.
Slashdot nearly always overstates the breadth of a patent's claims.
I have read several file wrappers and I feel that the PTO does a very good job of prior art search. Not "investigation" - they are not required to "investigate", only to search their own records.
Not to mention that the slashdot understanding of a patent's claims is almost always wrong. It is based on the headline. For a while I read every patent application that upset slashdot and was astonished at how wrong slashdot is. Now I don't bother.
Whatever this patent covers, it's definitely not what slashdot says.
You can only patent new things. If you can come up with a business method that isn't used yet, you may be able to patent it.
While I partially agree, I have to mention the counterpoint. MySQL is very good across a certain domain of problems. Where I work now, we pick between Oracle and MySQL for specific projects. If we don't see transactional requirements, writer contention, or huge volumes of data, we pick MySQL. It's faster and generally easier to work with.
I just don't care that MySQL isn't SQL-92 compliant. By the way, none of the applications I have worked on on Oracle could be trivially ported to a different db, standards or no. I agree that MySQL is friendlier than other dbs. Most obvious example - meaningful error messages.
I am not excited by Innodb or most of the features being added to MySQL. These changes, which add Oracle-like capabilities, also add Oracle-like problems, such as complexity, slowness, lock buildups, etc. We already have Oracle. We use MySQL where it fits.
As for hammers and nails: the guy with only one tool has a problem; so does the guy who spends all his time learning new tools and never becomes expert with one. There is not a logical niche for every technology - many of them are not worth touching, period. It's better to invest time in learning a few tools deeply.
But slashdot told me that IP laws only apply in the US, and by passing and enforcing them the US will just be at a disadvantage compared to "the rest of the world".
Could it be that all civilized countries act together in these matters? Could slashdot really be wrong?