Thank you for pointing this out before I had to. The bracket tables are only there to make the math easier, because we don't trust people to correctly calculate 15% of the first 20K+ 20% of everything between 20-30K +25% of eveything above (numbers completely made up). The tables simplify that for people and reduce mistakes.
And such a system is grossly unfair to the low wage earners, would make poverty line people struggle to feed themselves while being a windfall for the wealthy, and would shatter the economy as 75% of homewoners, who depend on the mortgage deduction, would go bankrupt.
A lot of deductions can definitely go away, and probably should. But the idea of "moving to a flat tax solves all problems" is so naive that to even suggest it you have to be a fucking idiot or have major external motivations.
I've worked outside of the large corporate world for more than half my career. I've never been IT. Never been asked to be IT. Why? Because I'm too fucking expensive to be IT in more than an emergency case- you can find IT guys for 3/4 to 1/2 my pay. It's a waste of time and money. The closest I've come to being IT is installing software on my own computer.
And even if for some dumb reason I had to be (which unless you're in a company of 10 or so you shouldn't), if you're hiring a programmer, you care about his programming skills. You're not going to be looking for an experienced dev with IT certs, because you won't find any. Two totally different career paths- there's no reason for anyone with any experience in programming to pick up one of those certs. The most you'd get is someone trying to change careers from one to the other, which is not the type of person you want as a dev at a very small company- too big a risk he can't do the job.
I actually did vote for every new tax on Washington's last ballot (and against raising the requirements to institute new taxes). I lost on all of them, unfortunately. And all but one of them would have impacted me.
I'm not going to prop up a broken system by paying more when those richer than me don't. I'm happy to pay my fair share, but not so someone better off than me can live higher on he hog. I gave to charity instead.
CE= computer engineering? Not sure from context. Where I went to school (UIUC), CE was basically like a major in CS and a minor in EE or vice versa, depending on how you did your electives. If they took a more CS version, they're qualified to do anything a CS person is, although perhaps a bit weaker at algorithm development (since they'd have to take some pretty deep electives for that). This type also makes a very good fit for a systems programmer or embedded programmer. If they took a more EE route, they're weaker on architecture but very well set up for an embedded programmer or doing something like DSP work. And of course they can always do circuits instead.
IT!=programming. IT doesn't need a degree, it would be like getting a degree to be an auto mechanic. Programming does take years of study to do efficiently and well. That's what the majority of the graduates go into, and the number of non-degreed people in it is low and dropping. Probably because most of the non-degreed people I've met knew the basics, but never bothered to learn the underlieing concepts of discrete math, complexity, or data structures- they just learned enough APIs to limp along.
I don't pay enough income tax. I actually pay less as a percentage than my sister, who earns 1/3 what I do. It's ridiculous that I pay less than 20% in taxes making a 6 figure income. And since I buy almost nothing, that means I only pay pretty much that 20%. I should be paying, and would vote to pay, at least double that.
My problem is that rich people and corporations pay even less as a percentage than I do. We need to bring back the Eisenhower tax rates. If you make more than 1 million a year, 90% of that should be taken away to pay for society (actual rates back then were 94%- we're actually the least taxed generation since WWII).
You're quite frankly out of your mind. There's a hell of a lot of programming jobs open in WA right now. I average a cold call every 2-3 days from recruiters- and I haven't updated my resume in over a year. If you're anything above the awful level it's pretty easy to get a job here right now, programming is in a boom cycle.
Took me about 2 minutes to write, more time spent on formatting then I would in a real text editor, due to writing in slashcode. Now to test I need to hook it up to the unit test framework (unknown time depending on framework, lets say it's negligible). I need to write a test for the null case, a test for starting letters, a test for embedded letters, a test for trailing letters. The same for non-letter characters. If I wanted to be truly cautious I'd want to do the same with non-printable characters, in case the idiot who wrote it did something really stupid. I'd want to embed a few nulls (0 values) just to make sure the author didn't assume that they ended the data stream (unless it's supposed to- in which case you should test that it does). I'm assuming ascii here, if that's not the case the above code is wrong, and the test suite should include tests for non ascii utf-8 characters.
I spent more than 2 minutes just thinking up the test cases for a trivial function. Actually writing them all would take me a good 10 minutes with a poor framework, probably 2-5 with a good one (admittedly I'm slower at coding unit tests than writing code, because I find unit tests boring. Consider that a personality flaw). And that's about the smallest number of test cases I'd consider for this trivial function to have remotely good coverage.
Writing unit tests is slow and takes time to do it right. There's payoff in the end, enough to make it worth it. But don't assume it's negligible time, it's really a huge investment.
There's also the people who got promoted who thought they were star programmers but were really B or C grade. Those are the worst of the bunch- they think they're much better than they are, and don't listen to their programmers. The B and C grades who are realistic about their skill levels will defer on technical issues.
Formal code reviews are worthless- nitpicky, political, and way too slow. Here's the rules for a good code review:
1)There must be a web based tool to upload the code to, which automatically emails when comments are added.
2)Said tool must have an easy way to say "Approved".
3)Code reviews are to look for logic bugs and possible side effects. They are not the place to complain about how someone likes to name variables (unless an extreme problem) or that someone forgot a space before a + sign. That's just a waste of everyone's time.
4)Code reviews are 2nd in priority only to ship blocking bugs. Get them done fast so the guy can move on to the next issue.
5)Code reviews should go out to a wide audience (so they can get potential bugs from anyone who knows the code), but do not need more than 1 approval. That approval may not be from a junior.
6)If you have roles in your code review other than submitter and reviewer, you're doing it wrong.
Test an out of memory error, without crashing the test harness.
Test a race condition.
Test every error condition that could possibly occur in your program. This includes every OS error return of every OS call.
Test all network errors, on every socket read.
All of these are reasons why 100% isn't possible. High 80s/90s is only possible in languages and code that make heavy use of exceptions, and then are artificially high (you aren't really fully testing all the error generating code, you're testing a small amount of it but the automated tools can't tell).
80% code coverage costing less than 10% development time? Only if you're either ignoring all errors, or have massive try catch blocks that catch everything generically (which means your code coverage numbers are garbage- if you aren't testing each way of generating each type of error, you only have a fraction of that coverage).
I'm not going to deny the usefulness of unit tests, but in my experience writing a good unit test suite takes 50-200% of the time to code the function, depending on the complexity of the code being tested.
No, I really don't. If you're an executive, you may get a credit check. Maybe for the defense industry. Regular jobs? Nope, not going to happen. I've been asked exactly once, said no, and still got the job.
Personal references do sometimes get called. But not generally- I've asked the ones I've used, they almost never get one. Nor is there really a point- you're giving the references yourself, why would you pick anyone who wouldn't give a top notch reference?
Employment history check to see you don't exaggerate your salary? No ex-employer gives that information out- most will barely say you worked for them. In fact, it's illegal to give that information out in many places unless authorized. Nor have I ever told a company what I made at my last job- I tell them what I want now. They're willing to pay it, or they aren't.
Umm, what planet do you come from where most companies have the money and time to do that for all potential hires? At most, the vast majority of companies will call the last 1 or 2 employers and make sure they worked there. Maybe they'll call/email a reference. They won't do more than that.
And how the hell would you background check code? You can throw a line into google and see if it has any hits, but there's no way to know if a friend wrote it for them. If they aren't writing the code during the interview, it can't be trusted.
I was like that 10 years ago. Since then I had a close call with a nervous breakdown from overwork and personal life crisis, and I haven't worked more than a dozen or so hours a year on personal programming projects since. Too much stress.
For that matter, too many other things to do. I have a dating life, friends, and other interests. When I want to do something intellectual, I'd rather study history, physics, math, or a field of engineering I know little about (such as mech eng) rather than work on a bullshit side project. Especially one as boring as writing a mail server- I can go and buy a dozen of those, what's the point of writing it? Unless you can do it an order of magnitude better, you're wasting your time.
Am I a bad programmer? My boss doesn't think so- I've received a 18% raise this year, and almost doubled my equity stake on top of that. I've been at the job 1 year and been put in charge of their highest volume, highest revenue driving platform. I was lead programmer on the 3.0 effort. So either they think I'm good or they're truly desperate.
In fact, all of the best programmers I know wouldn't be caught dead working on personal programming projects. Life's too short to spend more than the 40-50 hours a week you already do on it.
Are you insane? The AI is so piss poor that I beat it on the highest setting in under a week. I've never beaten Monarch on any of the others. It's completely broken.
Those are perfectly legitimate ways to question, so long as you give credit for all viable answers. The idea then being how you can respond and come up with additional ideas. For example, many of the algorithm questions I use have an obvious n^2 answer. If you come up with that, you get points for thinking of the obvious, but I still want you to solve it more efficiently. Many of them also have n^2 space algorithms. Again, you get points, but I want to see you solve it the real way.
By the way, I'd accept your answer on the cake, but only if the cake was unfrosted/untopped. What did they want, cut once, put the 2 pieces on top of each other, cut again, repeat?
Please tell me where you work and if you hire. I want to avoid ever working there.
If you don't want bottom of the barrel people (and Google doesn't) you need to be much more specific than that. I don't work at google, but my technique is to ask very difficult but not standard CS course algorithmic problems. Candidates are supposed to struggle, I want to see them thinking and see what approaches they take to the problem. Being given hints is ok, but they need to be able to solve it within a reasonable time. If they can't I don't want to hire them- I'll spend more time fixing their work than I would doing it myself.
I'm sure some good candidates slip away, due to having a bad day or just not getting the particular problem of the day. I know I've done it myself at 1 or 2 jobs. But I'd rather miss out on a good fit than hire a bad one.
ANd this is pretty much industry standard. I've worked at HP, Amazon, and multiple west coast startups. They all do interviews like this. What the hell type of jobs are you doing that they don't?
Because every contracting job I've ever seen pays about the same or less, has no stock, no benefits, is the first to be fired, gets treated like shit, has no vacation, and has to track hours. No thanks. Full time at a startup is the only way to go.
And IT isn't programming at all. IT is the domain of setting up computers and networks and sys admins. Programmers are their own category, and lumping them into "IT" just confuses every conversation since their jobs are so different. Its like calling a mech e who designs engines and a mechanic "automotive technologists". At that point the name is meaningless.
Knock the gold and silver off your list, they'll be worse than useless. In that event, there's only a few things anyone will take in trade- ammo, food, medicine, sex, drugs. Noone will take gold or silver because they can't do jack shit with them. Those are only valuable in the case that society is likely to reform- in which case plain old dollars will suffice.
Slashdot Mirror
Thank you for pointing this out before I had to. The bracket tables are only there to make the math easier, because we don't trust people to correctly calculate 15% of the first 20K+ 20% of everything between 20-30K +25% of eveything above (numbers completely made up). The tables simplify that for people and reduce mistakes.
And such a system is grossly unfair to the low wage earners, would make poverty line people struggle to feed themselves while being a windfall for the wealthy, and would shatter the economy as 75% of homewoners, who depend on the mortgage deduction, would go bankrupt.
A lot of deductions can definitely go away, and probably should. But the idea of "moving to a flat tax solves all problems" is so naive that to even suggest it you have to be a fucking idiot or have major external motivations.
I've worked outside of the large corporate world for more than half my career. I've never been IT. Never been asked to be IT. Why? Because I'm too fucking expensive to be IT in more than an emergency case- you can find IT guys for 3/4 to 1/2 my pay. It's a waste of time and money. The closest I've come to being IT is installing software on my own computer.
And even if for some dumb reason I had to be (which unless you're in a company of 10 or so you shouldn't), if you're hiring a programmer, you care about his programming skills. You're not going to be looking for an experienced dev with IT certs, because you won't find any. Two totally different career paths- there's no reason for anyone with any experience in programming to pick up one of those certs. The most you'd get is someone trying to change careers from one to the other, which is not the type of person you want as a dev at a very small company- too big a risk he can't do the job.
Certs? Once again, IT!=programming. Certs are for IT, if anything having a cert loses you value for programming.
I actually did vote for every new tax on Washington's last ballot (and against raising the requirements to institute new taxes). I lost on all of them, unfortunately. And all but one of them would have impacted me.
I'm not going to prop up a broken system by paying more when those richer than me don't. I'm happy to pay my fair share, but not so someone better off than me can live higher on he hog. I gave to charity instead.
CE= computer engineering? Not sure from context. Where I went to school (UIUC), CE was basically like a major in CS and a minor in EE or vice versa, depending on how you did your electives. If they took a more CS version, they're qualified to do anything a CS person is, although perhaps a bit weaker at algorithm development (since they'd have to take some pretty deep electives for that). This type also makes a very good fit for a systems programmer or embedded programmer. If they took a more EE route, they're weaker on architecture but very well set up for an embedded programmer or doing something like DSP work. And of course they can always do circuits instead.
IT!=programming. IT doesn't need a degree, it would be like getting a degree to be an auto mechanic. Programming does take years of study to do efficiently and well. That's what the majority of the graduates go into, and the number of non-degreed people in it is low and dropping. Probably because most of the non-degreed people I've met knew the basics, but never bothered to learn the underlieing concepts of discrete math, complexity, or data structures- they just learned enough APIs to limp along.
I don't pay enough income tax. I actually pay less as a percentage than my sister, who earns 1/3 what I do. It's ridiculous that I pay less than 20% in taxes making a 6 figure income. And since I buy almost nothing, that means I only pay pretty much that 20%. I should be paying, and would vote to pay, at least double that.
My problem is that rich people and corporations pay even less as a percentage than I do. We need to bring back the Eisenhower tax rates. If you make more than 1 million a year, 90% of that should be taken away to pay for society (actual rates back then were 94%- we're actually the least taxed generation since WWII).
You're quite frankly out of your mind. There's a hell of a lot of programming jobs open in WA right now. I average a cold call every 2-3 days from recruiters- and I haven't updated my resume in over a year. If you're anything above the awful level it's pretty easy to get a job here right now, programming is in a boom cycle.
Actually, the smaller and more modular your code is the more time you need to properly unit test it as a percentage of time. For example, the C code
Took me about 2 minutes to write, more time spent on formatting then I would in a real text editor, due to writing in slashcode. Now to test I need to hook it up to the unit test framework (unknown time depending on framework, lets say it's negligible). I need to write a test for the null case, a test for starting letters, a test for embedded letters, a test for trailing letters. The same for non-letter characters. If I wanted to be truly cautious I'd want to do the same with non-printable characters, in case the idiot who wrote it did something really stupid. I'd want to embed a few nulls (0 values) just to make sure the author didn't assume that they ended the data stream (unless it's supposed to- in which case you should test that it does). I'm assuming ascii here, if that's not the case the above code is wrong, and the test suite should include tests for non ascii utf-8 characters.
I spent more than 2 minutes just thinking up the test cases for a trivial function. Actually writing them all would take me a good 10 minutes with a poor framework, probably 2-5 with a good one (admittedly I'm slower at coding unit tests than writing code, because I find unit tests boring. Consider that a personality flaw). And that's about the smallest number of test cases I'd consider for this trivial function to have remotely good coverage.
Writing unit tests is slow and takes time to do it right. There's payoff in the end, enough to make it worth it. But don't assume it's negligible time, it's really a huge investment.
There's also the people who got promoted who thought they were star programmers but were really B or C grade. Those are the worst of the bunch- they think they're much better than they are, and don't listen to their programmers. The B and C grades who are realistic about their skill levels will defer on technical issues.
Formal code reviews are worthless- nitpicky, political, and way too slow. Here's the rules for a good code review:
1)There must be a web based tool to upload the code to, which automatically emails when comments are added.
2)Said tool must have an easy way to say "Approved".
3)Code reviews are to look for logic bugs and possible side effects. They are not the place to complain about how someone likes to name variables (unless an extreme problem) or that someone forgot a space before a + sign. That's just a waste of everyone's time.
4)Code reviews are 2nd in priority only to ship blocking bugs. Get them done fast so the guy can move on to the next issue.
5)Code reviews should go out to a wide audience (so they can get potential bugs from anyone who knows the code), but do not need more than 1 approval. That approval may not be from a junior.
6)If you have roles in your code review other than submitter and reviewer, you're doing it wrong.
Test an out of memory error, without crashing the test harness.
Test a race condition.
Test every error condition that could possibly occur in your program. This includes every OS error return of every OS call.
Test all network errors, on every socket read.
All of these are reasons why 100% isn't possible. High 80s/90s is only possible in languages and code that make heavy use of exceptions, and then are artificially high (you aren't really fully testing all the error generating code, you're testing a small amount of it but the automated tools can't tell).
80% code coverage costing less than 10% development time? Only if you're either ignoring all errors, or have massive try catch blocks that catch everything generically (which means your code coverage numbers are garbage- if you aren't testing each way of generating each type of error, you only have a fraction of that coverage).
I'm not going to deny the usefulness of unit tests, but in my experience writing a good unit test suite takes 50-200% of the time to code the function, depending on the complexity of the code being tested.
No, I really don't. If you're an executive, you may get a credit check. Maybe for the defense industry. Regular jobs? Nope, not going to happen. I've been asked exactly once, said no, and still got the job.
Personal references do sometimes get called. But not generally- I've asked the ones I've used, they almost never get one. Nor is there really a point- you're giving the references yourself, why would you pick anyone who wouldn't give a top notch reference?
Employment history check to see you don't exaggerate your salary? No ex-employer gives that information out- most will barely say you worked for them. In fact, it's illegal to give that information out in many places unless authorized. Nor have I ever told a company what I made at my last job- I tell them what I want now. They're willing to pay it, or they aren't.
Umm, what planet do you come from where most companies have the money and time to do that for all potential hires? At most, the vast majority of companies will call the last 1 or 2 employers and make sure they worked there. Maybe they'll call/email a reference. They won't do more than that.
And how the hell would you background check code? You can throw a line into google and see if it has any hits, but there's no way to know if a friend wrote it for them. If they aren't writing the code during the interview, it can't be trusted.
I was like that 10 years ago. Since then I had a close call with a nervous breakdown from overwork and personal life crisis, and I haven't worked more than a dozen or so hours a year on personal programming projects since. Too much stress.
For that matter, too many other things to do. I have a dating life, friends, and other interests. When I want to do something intellectual, I'd rather study history, physics, math, or a field of engineering I know little about (such as mech eng) rather than work on a bullshit side project. Especially one as boring as writing a mail server- I can go and buy a dozen of those, what's the point of writing it? Unless you can do it an order of magnitude better, you're wasting your time.
Am I a bad programmer? My boss doesn't think so- I've received a 18% raise this year, and almost doubled my equity stake on top of that. I've been at the job 1 year and been put in charge of their highest volume, highest revenue driving platform. I was lead programmer on the 3.0 effort. So either they think I'm good or they're truly desperate.
In fact, all of the best programmers I know wouldn't be caught dead working on personal programming projects. Life's too short to spend more than the 40-50 hours a week you already do on it.
Are you insane? The AI is so piss poor that I beat it on the highest setting in under a week. I've never beaten Monarch on any of the others. It's completely broken.
Those are perfectly legitimate ways to question, so long as you give credit for all viable answers. The idea then being how you can respond and come up with additional ideas. For example, many of the algorithm questions I use have an obvious n^2 answer. If you come up with that, you get points for thinking of the obvious, but I still want you to solve it more efficiently. Many of them also have n^2 space algorithms. Again, you get points, but I want to see you solve it the real way.
By the way, I'd accept your answer on the cake, but only if the cake was unfrosted/untopped. What did they want, cut once, put the 2 pieces on top of each other, cut again, repeat?
Please tell me where you work and if you hire. I want to avoid ever working there.
If you don't want bottom of the barrel people (and Google doesn't) you need to be much more specific than that. I don't work at google, but my technique is to ask very difficult but not standard CS course algorithmic problems. Candidates are supposed to struggle, I want to see them thinking and see what approaches they take to the problem. Being given hints is ok, but they need to be able to solve it within a reasonable time. If they can't I don't want to hire them- I'll spend more time fixing their work than I would doing it myself.
I'm sure some good candidates slip away, due to having a bad day or just not getting the particular problem of the day. I know I've done it myself at 1 or 2 jobs. But I'd rather miss out on a good fit than hire a bad one.
ANd this is pretty much industry standard. I've worked at HP, Amazon, and multiple west coast startups. They all do interviews like this. What the hell type of jobs are you doing that they don't?
Because every contracting job I've ever seen pays about the same or less, has no stock, no benefits, is the first to be fired, gets treated like shit, has no vacation, and has to track hours. No thanks. Full time at a startup is the only way to go.
Now factor in inflation and the increased number of people in the country. Doubling the cost 10 years later seems quite reasonable to me.
And IT isn't programming at all. IT is the domain of setting up computers and networks and sys admins. Programmers are their own category, and lumping them into "IT" just confuses every conversation since their jobs are so different. Its like calling a mech e who designs engines and a mechanic "automotive technologists". At that point the name is meaningless.
Sure you do- walk out. That's why you keep a years pay in the bank.
Knock the gold and silver off your list, they'll be worse than useless. In that event, there's only a few things anyone will take in trade- ammo, food, medicine, sex, drugs. Noone will take gold or silver because they can't do jack shit with them. Those are only valuable in the case that society is likely to reform- in which case plain old dollars will suffice.