This. If you're hiring a coder, ASK CODING QUESTIONS. At my job, we need embedded C programmers. And whether you're coming in for an internship or have 10 years experience, you're going to be asked to write whiteboard code. Find the least set bit in a 32 bit word, linked list insert, basic stuff. Someone in college should be able to manage and someone with experience should have no problem at all.
What's disturbing is the number of people who have great resumes, years of experience, can talk in depth about their projects, etc., but can't do basic pointer manipulation or know what "volatile" means. It boggles the mind.
Seriously. Even on my Mac, all that does is highlight the story for no apparent reason. Why on earth is that region clickable, anyway? It makes it damn near impossible to hit the REAL links in the middle of the giant one.
That's not going to work. The keys they use are derived from IDs stored in both the card and the tuner. And I think they may do some sort of pseudo-randomization on the tuner ID because I seem to recall that yanking it and putting it back in my TiVo made me go through the activation process again.
That said, my Comcast guys don't care what device I put my M-Card in. I just need to call them and read off the numbers from my TiVo. And then do the same thing for the next couple days while rebooting the box several dozen times before it actually will decrypt anything.
Why? Can't you just link to wikipedia pages where appropriate? OK, my company has an internal server we link through to sanitize referrer info so our internal wiki titles don't get all over teh interwebs. But if the wiki users can't figure out "hey, this article is too specific - maybe wikipedia has more general information that would help me," you've got bigger problems than your wiki management.
Hey, I was in the first year of ECE 189 back in '03 (also in the first graduating class of their newfangled CE major). Very cool class.
And hey: my project was a portable* laptop drive-based MP3 player and five years later, I'm working one one again. But this time it's for a well-known company in Cupertino...
* "Portable" MP3 player turned out to be about half the size of a lunchbox. Don't forget mounting holes in your PCB: other mounting solutions are difficult and large!
The number is even more meaningless when you consider what a "failure" is (as pointed out my Feynman after Challenger). We only consider a shuttle mission a failure if the situation gets so FUBAR that the vehicle explodes.
But all those times that the O-rings showed wear that implied they weren't operating as designed? Not considered failures.
Just because the thing took off and landed again without killing anyone doesn't mean that it was working as designed and certainly doesn't mean it was operating safely.
How exactly did the state violate the constitution if it was given permission? If a cop knocks on my door, he can ask "May I have a look around?" and if I say "go ahead", well, whatever's in there is fair game. If I say "no" and he looks anyway, that's a whole other story.
The 4th just means the government can't FORCE you to cooperate. It doesn't mean they can't ask.
Looking into your config? There shouldn't be anything I could possibly configure on the thing that should make it require a reboot.
It should 1) Not allow you to do that in the first place, 2) Fail gracefully (possibly reverting the part that's a problem) if it DOES let you, and 3) Have a watchdog make it reset ITSELF if it's found itself unable to keep routing.
No, requiring a manual reset is a sign of people who don't know how to develop an embedded system for crap.
That's still "bad" in my book. Maybe I'm just not a good geek, but I don't need or want my own DNS. I've got a laptop, a desktop that I never use anymore, an iPhone, and a Tivo on my network. It shouldn't be too much to ask that they just be able to get to the Internet.
Indeed. I do firmware for semiconductor startups, so I'm firmly into the "C everyday" camp. Yeah, I learned C/C++ in college, but I have to say, I learned about as much in the 2.5 years at my first job as I did getting my BS. Probably because I quickly realized I didn't know jack and had to bust my ass to get up to speed.
Anyway, if there's anything that'll help you stand out when you graduate, it's cool projects and internships. I was a Computer Engineering major (thought I'd be a digital hardware guy, didn't happen that way) and we had to do an embedded system project senior year. That turned out to be the key that got me into that first entry-level position. Yeah, in hindsight our design was lousy, we chose an inappropriate CPU for the job, and had to bit-bang our ATA harddrive because we screwed up a couple signals. But it still worked, had real hardware interaction, and was real C and ASM with no operating system to help. So at least they guys hiring me knew I understood what a pointer was.
From then I was just a matter of unlearning all the bad habits I'd developed (seriously, a while ago I found some of my old code from school and it was damn-near unreadable)
Exactly. And people would be wise to remember that. Take the whole USB vs 1394 arguments that come up from time to time. Some people insist that since high speed USB 2.0 runs at 480Mbps versus Firewire's 400Mbps, USB is *obviously* faster. Nevermind the fact that 480/400 are signalling rates that have absolutely nothing to do with throughput. Speaking as a USB developer, the spec just says that when a host or device sends a 512 byte USB packet, it goes down the wire at a speedy 480Mbps.
But the spec doesn't care how often you send one of those speedy packets. So in reality, going thru the whole MS disk and USB stacks, you're looking at *maybe* 200Mbps of actual throughput.
Hell, the fastest USB device I ever saw was a piece of hardware that did nothing but consume data as fast as its little memory bus could run with a driver that sent the biggest payloads as fast as it could and it crapped out at 320Mbps (limited by the PC).
Remember: higher signalling rates don't always lead to higher throughput!
Well, I'm a CE grad pushing three years out in the real world. Personally, I think CE was the best path for me. My program was very open-ended. There were analog hardware classes, digital hardware, software, VLSI, OSes, you name it. I focused mainly on digital HW and embedded sw stuff. My final senior project was an mp3 player: I was in a group with 3 other guys and we did everything (circuit design, board layout, firmware, PC software, etc). It was kind of flakey, but it worked.
The thing is, at the end of 4 years, I had two resumes: my embedded software resume, and my digital design/VLSI resume. I felt comfortable applying to entry-level for either path. Embedded software worked out. CE opened a lot of doors that I woudln't get with a CS degree. CS has other paths CE doesn't but none of them appealed to me (I had no desire to be a game programmer). Once you pick a door and get into the workplace, however, it's not particularly important either way. Experience dominates from then on.
A major point to consider is that it's engineering, not science. If research is your thing, I'd suggest CS. My math and algorithm coursework was adequate, but kind of weak compared to CS guys. Similarly if you like working on big apps (windows stuff, games, business apps, etc.) I'd say go CS.
But if embedded systems, drivers, OS work, etc. sounds like your thing, CE is definitely the way to go. Most CS guys who come thru can't work an oscilloscope to save their life.
While I'm at it, I'd also suggest you look into internships and if you can take a project/capstone class, do it. We get tons of entry-level resumes, most of which look the same. GPA's not a big deal to us so long as you graduated. What sets people apart is really experience (i.e. internships) and interesting/relevant projects. It was that mp3 player that got me my first job more than my major.
How much does your prepaid service cost? Looks like you're in the UK, but here in the US, prepaid tends to run in the $0.25/min range (just from browsing a couple providers). Figure 10min of calls a day (~300min/month) and we're looking at $75/month. In contrast, for $45, I'm getting 1000 minutes per month.
*checks bill*
Looks like last month I used about 500 minutes...makes the decision to stick with a contract pretty easy. They'd have to get the price down to around $0.11 for me to break even and below that to overcome the equipment subsidies.
All the kernel mode stuff is downright maddening. Maybe I'm just stupid, the whole notion of an IRP just seems like a pain in the ass. Ok so it's asynchronous. The the code to deal with it is huge and if things don't work, it's damn near impossible to figure out *why*.
And, as a USB developer, it boggles my mind that the XP DDK comes with a "simple" USB BULK transfer driver example: it's 8492 lines of code in 6 C files and 7 headers. 2751 of those lines are for Windows PnP support. 1686 are for power management (USB only supports three states! Connected, suspended, disconnected!)
And it does damn near the same thing as the 349 line usb-skeleton.c in Linux (essentially allows simple read()/write() access to a bulk endpoint pair).
Well-designed my ass. We're talking a factor of 24x more code to do the same thing.
Yes, you always need probably cause, but things like slow driving laws vary from state to state. For example, in California you need a minimum speed sign posted to have a minimum speed. I don't think I've *ever* seen one in CA (though I have seen them in other states).
But we've still got this handy section in the books:
22400. (a) No person shall drive upon a highway at such a slow speed as to impede or block the normal and reasonable movement of traffic, unless the reduced speed is necessary for safe operation, because of a grade, or in compliance with law.
Out here, all you need to do is hold up people behind you and you're fair game. I knew a guy that got cited for that when he was (stupidly) doing 50 on a 65 freeway.
The dealer can say all they want (even here in California), but it doesn't matter unless they can show that your filter was directly responsible for any damage to your car.
See the Magnuson Moss Warranty Act (15 USC 2301-2312). Good stuff to be aware of, especially when dealing with cars. It applies to other consumer items as well, though.
People really need to stop going to vendors for information on what is and is not allowed with their products. Just because a car dealer says you can't change a filter or a software vendor prints something in an EULA doesn't make it so, no matter how much they scream and wave their arms.
Crazy, huh? Looking at property here in the Bay Area I discovered that with an income of about 62k, I qualify for low-income housing assistance in a lot of towns here. Truely absurd.
Re:I wondered what the hell that stuff was...
on
Hacking Vodka
·
· Score: 1
Ah yes, the joys of cheap vodka. One minute you're sitting in a friend's dorm room watching Army of Darkness. Next minute it's 8am. You're on the floor. In your underwear. Next to an empty bottle of Popov. Covered head-to-toe with crude drawings of male genitalia.
Thank you. Why *does* this need a warning? He knows he's not supposed to do it. When I was working for a large defense contractor in IT, they were always quite clear: "DO NOT FUCK WITH THE SERVERS" and especially "DO NOT FUCK WITH THE GOVERNMENT SERVERS". It was in the handbook. It said it when I logged into the machines. My boss told me. That's three warnings right there. This was *not* a point open to debate.
And if I was stupid enough to install rogue software onto a server there, I certainly wouldn't expect a warning/second chance. I know for certain I would immediately have my admin and root access pulled and would probably be fired too.
It blows my mind that someone would think that this was a good idea. The boss should have kept his mouth shut with respect to the "intelligent life" comment, but frankly I agree with his assessment.
You can get gcc for AVRs and Atmel has their own compiler as well. There are also a couple commerical ones available.
I did some work with the ATMega128 a couple years back. Used an STK500/STK501 eval board to get started. Easy enough. Nothing scary in the board design. The datasheet tells you what you need to do (in terms of required passives, crystal, etc.) Just make sure to get the in-circuit programming stuff on there so you can modify the flash when it's on the board. And if you haven't done anything before, remember: headers are your friends. You'll probably screw *something* up and it's really nice to be able to probe things and have posts ready to wrap to when you need to bust out the blue wires. My first board (mp3 player, school project) looked like a porcupine with all the headers. And good thing, too, cuz we made some *really* stupid mistakes.
Never had a chance to use PICs (used AVRs as a hobbiest and more 32-bit DSPs at work...and I'm really a firmware guy so I don't get much input in the board designs anyway)
even though they're shooting from the Mojave desert and there's no chance of it happening.
Speaking as someone who had the misfortune to have to spend six years living in the Mojave desert due to my dad being stationed at Edwards AFB, I feel I should remind you that there is *some* civilization out there. The towns of Lancaster and Palmdale come to mind...
Not that I'd shed any tears if that area was incinerated, but still...
*mutters something incoherent about the stupid desert and those damn tumbleweeds*
Slashdot Mirror
This. If you're hiring a coder, ASK CODING QUESTIONS. At my job, we need embedded C programmers. And whether you're coming in for an internship or have 10 years experience, you're going to be asked to write whiteboard code. Find the least set bit in a 32 bit word, linked list insert, basic stuff. Someone in college should be able to manage and someone with experience should have no problem at all.
What's disturbing is the number of people who have great resumes, years of experience, can talk in depth about their projects, etc., but can't do basic pointer manipulation or know what "volatile" means. It boggles the mind.
Seriously. Even on my Mac, all that does is highlight the story for no apparent reason. Why on earth is that region clickable, anyway? It makes it damn near impossible to hit the REAL links in the middle of the giant one.
That's not going to work. The keys they use are derived from IDs stored in both the card and the tuner. And I think they may do some sort of pseudo-randomization on the tuner ID because I seem to recall that yanking it and putting it back in my TiVo made me go through the activation process again.
That said, my Comcast guys don't care what device I put my M-Card in. I just need to call them and read off the numbers from my TiVo. And then do the same thing for the next couple days while rebooting the box several dozen times before it actually will decrypt anything.
Why? Can't you just link to wikipedia pages where appropriate? OK, my company has an internal server we link through to sanitize referrer info so our internal wiki titles don't get all over teh interwebs. But if the wiki users can't figure out "hey, this article is too specific - maybe wikipedia has more general information that would help me," you've got bigger problems than your wiki management.
Hey, I was in the first year of ECE 189 back in '03 (also in the first graduating class of their newfangled CE major). Very cool class.
And hey: my project was a portable* laptop drive-based MP3 player and five years later, I'm working one one again. But this time it's for a well-known company in Cupertino...
* "Portable" MP3 player turned out to be about half the size of a lunchbox. Don't forget mounting holes in your PCB: other mounting solutions are difficult and large!
The number is even more meaningless when you consider what a "failure" is (as pointed out my Feynman after Challenger). We only consider a shuttle mission a failure if the situation gets so FUBAR that the vehicle explodes.
But all those times that the O-rings showed wear that implied they weren't operating as designed? Not considered failures.
Just because the thing took off and landed again without killing anyone doesn't mean that it was working as designed and certainly doesn't mean it was operating safely.
How exactly did the state violate the constitution if it was given permission? If a cop knocks on my door, he can ask "May I have a look around?" and if I say "go ahead", well, whatever's in there is fair game. If I say "no" and he looks anyway, that's a whole other story.
The 4th just means the government can't FORCE you to cooperate. It doesn't mean they can't ask.
Looking into your config? There shouldn't be anything I could possibly configure on the thing that should make it require a reboot.
It should 1) Not allow you to do that in the first place, 2) Fail gracefully (possibly reverting the part that's a problem) if it DOES let you, and 3) Have a watchdog make it reset ITSELF if it's found itself unable to keep routing.
No, requiring a manual reset is a sign of people who don't know how to develop an embedded system for crap.
That's still "bad" in my book. Maybe I'm just not a good geek, but I don't need or want my own DNS. I've got a laptop, a desktop that I never use anymore, an iPhone, and a Tivo on my network. It shouldn't be too much to ask that they just be able to get to the Internet.
Indeed. I do firmware for semiconductor startups, so I'm firmly into the "C everyday" camp. Yeah, I learned C/C++ in college, but I have to say, I learned about as much in the 2.5 years at my first job as I did getting my BS. Probably because I quickly realized I didn't know jack and had to bust my ass to get up to speed.
Anyway, if there's anything that'll help you stand out when you graduate, it's cool projects and internships. I was a Computer Engineering major (thought I'd be a digital hardware guy, didn't happen that way) and we had to do an embedded system project senior year. That turned out to be the key that got me into that first entry-level position. Yeah, in hindsight our design was lousy, we chose an inappropriate CPU for the job, and had to bit-bang our ATA harddrive because we screwed up a couple signals. But it still worked, had real hardware interaction, and was real C and ASM with no operating system to help. So at least they guys hiring me knew I understood what a pointer was.
From then I was just a matter of unlearning all the bad habits I'd developed (seriously, a while ago I found some of my old code from school and it was damn-near unreadable)
As has been mentioned before, cops are cops, and renta-cops are not.
UCPD are *not* rent-a-cops. They're real CA state police with all the same responsibilities as the CHP.
Exactly. And people would be wise to remember that. Take the whole USB vs 1394 arguments that come up from time to time. Some people insist that since high speed USB 2.0 runs at 480Mbps versus Firewire's 400Mbps, USB is *obviously* faster. Nevermind the fact that 480/400 are signalling rates that have absolutely nothing to do with throughput. Speaking as a USB developer, the spec just says that when a host or device sends a 512 byte USB packet, it goes down the wire at a speedy 480Mbps.
But the spec doesn't care how often you send one of those speedy packets. So in reality, going thru the whole MS disk and USB stacks, you're looking at *maybe* 200Mbps of actual throughput.
Hell, the fastest USB device I ever saw was a piece of hardware that did nothing but consume data as fast as its little memory bus could run with a driver that sent the biggest payloads as fast as it could and it crapped out at 320Mbps (limited by the PC).
Remember: higher signalling rates don't always lead to higher throughput!
Well, I'm a CE grad pushing three years out in the real world. Personally, I think CE was the best path for me. My program was very open-ended. There were analog hardware classes, digital hardware, software, VLSI, OSes, you name it. I focused mainly on digital HW and embedded sw stuff. My final senior project was an mp3 player: I was in a group with 3 other guys and we did everything (circuit design, board layout, firmware, PC software, etc). It was kind of flakey, but it worked.
The thing is, at the end of 4 years, I had two resumes: my embedded software resume, and my digital design/VLSI resume. I felt comfortable applying to entry-level for either path. Embedded software worked out. CE opened a lot of doors that I woudln't get with a CS degree. CS has other paths CE doesn't but none of them appealed to me (I had no desire to be a game programmer). Once you pick a door and get into the workplace, however, it's not particularly important either way. Experience dominates from then on.
A major point to consider is that it's engineering, not science. If research is your thing, I'd suggest CS. My math and algorithm coursework was adequate, but kind of weak compared to CS guys. Similarly if you like working on big apps (windows stuff, games, business apps, etc.) I'd say go CS.
But if embedded systems, drivers, OS work, etc. sounds like your thing, CE is definitely the way to go. Most CS guys who come thru can't work an oscilloscope to save their life.
While I'm at it, I'd also suggest you look into internships and if you can take a project/capstone class, do it. We get tons of entry-level resumes, most of which look the same. GPA's not a big deal to us so long as you graduated. What sets people apart is really experience (i.e. internships) and interesting/relevant projects. It was that mp3 player that got me my first job more than my major.
As someone who just bought a MacBook a couple weeks ago, I think I'd cry if they announced a 30% sale...
How much does your prepaid service cost? Looks like you're in the UK, but here in the US, prepaid tends to run in the $0.25/min range (just from browsing a couple providers). Figure 10min of calls a day (~300min/month) and we're looking at $75/month. In contrast, for $45, I'm getting 1000 minutes per month.
*checks bill*
Looks like last month I used about 500 minutes...makes the decision to stick with a contract pretty easy. They'd have to get the price down to around $0.11 for me to break even and below that to overcome the equipment subsidies.
I wish I could say that about *any* stack.
All the kernel mode stuff is downright maddening. Maybe I'm just stupid, the whole notion of an IRP just seems like a pain in the ass. Ok so it's asynchronous. The the code to deal with it is huge and if things don't work, it's damn near impossible to figure out *why*.
And, as a USB developer, it boggles my mind that the XP DDK comes with a "simple" USB BULK transfer driver example: it's 8492 lines of code in 6 C files and 7 headers. 2751 of those lines are for Windows PnP support. 1686 are for power management (USB only supports three states! Connected, suspended, disconnected!)
And it does damn near the same thing as the 349 line usb-skeleton.c in Linux (essentially allows simple read()/write() access to a bulk endpoint pair).
Well-designed my ass. We're talking a factor of 24x more code to do the same thing.
But we've still got this handy section in the books:
Out here, all you need to do is hold up people behind you and you're fair game. I knew a guy that got cited for that when he was (stupidly) doing 50 on a 65 freeway.
The dealer can say all they want (even here in California), but it doesn't matter unless they can show that your filter was directly responsible for any damage to your car.
See the Magnuson Moss Warranty Act (15 USC 2301-2312). Good stuff to be aware of, especially when dealing with cars. It applies to other consumer items as well, though.
People really need to stop going to vendors for information on what is and is not allowed with their products. Just because a car dealer says you can't change a filter or a software vendor prints something in an EULA doesn't make it so, no matter how much they scream and wave their arms.
Crazy, huh? Looking at property here in the Bay Area I discovered that with an income of about 62k, I qualify for low-income housing assistance in a lot of towns here. Truely absurd.
Ah yes, the joys of cheap vodka. One minute you're sitting in a friend's dorm room watching Army of Darkness. Next minute it's 8am. You're on the floor. In your underwear. Next to an empty bottle of Popov. Covered head-to-toe with crude drawings of male genitalia.
Good times.
Props to a fellow Natty Light drinker!
Breakfast of champions...
Thank you. Why *does* this need a warning? He knows he's not supposed to do it. When I was working for a large defense contractor in IT, they were always quite clear: "DO NOT FUCK WITH THE SERVERS" and especially "DO NOT FUCK WITH THE GOVERNMENT SERVERS". It was in the handbook. It said it when I logged into the machines. My boss told me. That's three warnings right there. This was *not* a point open to debate.
And if I was stupid enough to install rogue software onto a server there, I certainly wouldn't expect a warning/second chance. I know for certain I would immediately have my admin and root access pulled and would probably be fired too.
It blows my mind that someone would think that this was a good idea. The boss should have kept his mouth shut with respect to the "intelligent life" comment, but frankly I agree with his assessment.
You can get gcc for AVRs and Atmel has their own compiler as well. There are also a couple commerical ones available.
I did some work with the ATMega128 a couple years back. Used an STK500/STK501 eval board to get started. Easy enough. Nothing scary in the board design. The datasheet tells you what you need to do (in terms of required passives, crystal, etc.) Just make sure to get the in-circuit programming stuff on there so you can modify the flash when it's on the board. And if you haven't done anything before, remember: headers are your friends. You'll probably screw *something* up and it's really nice to be able to probe things and have posts ready to wrap to when you need to bust out the blue wires. My first board (mp3 player, school project) looked like a porcupine with all the headers. And good thing, too, cuz we made some *really* stupid mistakes.
Never had a chance to use PICs (used AVRs as a hobbiest and more 32-bit DSPs at work...and I'm really a firmware guy so I don't get much input in the board designs anyway)
Hmm....I4100 here.
broken power cord (insulation/shield seperated at connector): check
pink screen thing: check
funky trackpad: check
But you forgot:
broken bezel and keyboard mounts from ~1ft fall onto plush carpet (when landing flat)
having to jiggle the ethernet cable to get a link
bad keyboard interface so two columns of keys in the middle stop working
display that can't close securely
weird squeeking sounds from the speakers
even though they're shooting from the Mojave desert and there's no chance of it happening.
Speaking as someone who had the misfortune to have to spend six years living in the Mojave desert due to my dad being stationed at Edwards AFB, I feel I should remind you that there is *some* civilization out there. The towns of Lancaster and Palmdale come to mind...
Not that I'd shed any tears if that area was incinerated, but still...
*mutters something incoherent about the stupid desert and those damn tumbleweeds*