But this is just a DSP. It has some on-board stuff to help the code do the protocol, but you still gotta write the code.
Furthur, this only deals with demodulating the signal once it has been received. You still need the RF front-end to do:
Narrowband FM at 800MHz for AMPS cellular (US cellphones)
Wideband CDMA at 900MHz for IS-96
Narrowband TDMA at 900MHz for GSM
Narrowband TDMA at 1800MHz for GSM-II
Narrowband TDMA at 1900MHz for PCS
8-QPSM at 800MHz, 900MHz and 1900MHz for GPRS
GMSK at 800MHz for CDPD
and that ain't easy.
Then, you have to decode the protocol (that's where the part comes in) plus any higher level signalling (usually, you'd used another microcontroller for this).
This isn't a one-chip phone by any means. A multiprotocol phone is going to be a mother of a beast to code.
Now, I work in the comms field, designing test equipment that uses the Motorola 56300 family of DSPs, and I am less than impressed. I hope this new chip fixes the bugs in the chip, the stupid DMA controller (MOTO DSP GUYS: ever hear of scatter/gather?), the brain-dead instruction set (who ever heard of a DSP without a saturating add instruction!) and oddball word length (let's move from 24 bits to 32, shall we?), and small stack size. Not to mention that this DSP is a fixed point only DSP, no floating point on board.
If anyone out there has a good link to this chip's data sheet (I looked on Moto's site, no dice), please post it.
The source for the entire Amiga OS, from the lowest levels all the way up to WorkBench and Intuition are open sourced
The code is joyously clean, well structured, and easy to hack
The code will build with gcc
OK, this helps us how? Without the Amiga chipset, much of what made the Amiga interesting is not possible. So, to what end can I spend my free time on this code?
Making Workbench into an X window manager/desktop? WB was ok, but there are betters wm's out there now. Why duplicate the effort?
Making a better Amiga emulator for [Linux|Be|Windows|*]? Possibly, but again, emulating the Amiga chipset isn't going to be easy. Have they Open Source'd the specs for those chips (not just the programming specs, but the actual implementation)?
Making compatibility layer for AmigaOS to allow it to run X apps? Maybe, but without virtual memory (68000 based Amiga 500s need not apply) you aren't going to easily port most X apps.
Now, don't get me wrong. I love the idea of open sourcing the Amiga OS, just as I loved the idea of Caldera open sourcing GEM and Atari open sourcing TOS. However, how much activity has the latter generated?
All I am saying is that this is interesting, but not earth shattering. Perhaps, if the rumors are true, and Amiga and Transmeta release an ubercomputer, this will be a seminal moment in personal computer history, but unless that happens, I fear that this won't even be a footnote. The hacker community won't take up the code and make it wonderful, it will just sit around taking up FTP server space.
Perhaps it is time to unleash the mighty power of the/. effect upon Apple: Why doesn't everybody who has bitched about not having Sorenson support under Xanim e-mail Apple, and ask them nicely to allow Sorenson to allow Mr. Podlipec to make a codec plug-in for Xanim. Since Apple is claiming to support Open Source, lets give them a chance to walk the walk, and not just talk the talk.
With DSL, though, I think everything is built for 24/7 connectivity, and even your voice calls are handled with data packets.
No, that's not how it works. The wires your phone uses can handle signals from DC to about 1MHz. The voice line uses 300Hz - 3kHz. Then there is a gap to about 10kHz, then DSL uses from about 20kHz to 1MHz. There's a splitter at your house, and the high frequency signals go to the DSL modem, and the low freqs to your phone. There's the same at the telco office: the low freqs go to the phone switch, and the high freqs to another DSL modem. So, your voice is still analog (at least until it hits the switch), and your DSL data never touches the phone network. That's why the phone companies love DSL: your switch usage goes back to the "old model" of usage (lots of short calls), and your data usage is handled seperately.
in the way FTP, HTTP, or HTML are. Allow me an analogy: C++ defines a language, but by the time you define a few macros, some overloaded operators, and a few new libraries, you can generate a language that is completely different from stock C++. In the same way, all XML does is give you a framework to define your tags in. If I define a tag < Unit Price >, and you define the same meaning as < Price per unit >, then we have an incompatability.
Now, look at what MS is doing:
"Extend" the HTTP standard
Use XML, so they can define their own tags willy-nilly.
That has about as much chance of being "open" as the Windows API does: Sure, you may be able to read & write data (more or less, given how much of the extended HTTP standard you get), and you may be able to parse some of the tags, but MS can simple define new tags that only they understand.
I'll beleive in the promise of XML when good, well defined standards exists for defining tags, and standard tags are defined for standard things.
Microsoft's latest efforts on Capitol Hill will have little or no impact on the department's antitrust case against the software giant, and for that reason they seem somewhat unusual. While companies regularly ask lawmakers to block federal agencies from implementing specific policies, it is more uncommon to seek an across-the-board cut in a department's budget, especially in the middle of a major court battle.
This is an example of Microsoft's long range planning: gut the department now, then start acting up again.
And nonprofit organizations that receive financial support from the company have also urged key congressional appropriators to limit spending for the division when they begin their final negotiations on the Justice Department budget, possibly as early as Monday.
More "astroturf" style grassroots marketing.
Well, all I can say is write your Congressdrones and complain.
What I hate about this stuff is that there is no option to just freaking download the damn video!
Look, I live in the country. I have about as much chance of getting broadband access this millenium (i.e. before 2001 because that is the millenium dammit) as Bill Gates using Linux for his house. (less, actually). All I want to do is start a download at the paultry 28.8 kbps I get, and walk away. I'll watch the video when it is finished, whenever that is.
Some things they stream to keep you from saving a copy, but something by maddog? Where's the "open source" in that? There ought to be an FTP site with the actual.RA file, so that we can just download it.
This is one of the many areas that Geeks In Space gets right: I can just download the file, and listen when I get it!
I am an software engineer, and work with several other software engineers who still don't understand mounting disks. I'd set up the autofs automounter, but for the following problem: until the directory is mounted, it does not show up.
Somebody (if I ever get the time I'll look into it myself) needs to modify autofs so that a mountpoint can be specified, and will show up in the automount directory, but won't be mounted until you actually try to enter that directory and retrieve the file list. That way, under [Gnome|KDE|Enlightenment|*], there will be an icon for drooling user boy to click on. This would go a long way to making Linux easier to use.
Now, re: Loki's install. The biggest problem with the install script that I had was that it was not tagged as executable on the CD. That means that:
The icon in Gnome won't be the "executable" icon, corn-fusing the user.
You won't be able to launch the install from a GUI and you won't be able to just launch it from the shell. Instead, you have to do a "sh install", which is non-intuitive to the average luser.
The TCL/TK script the installer runs tries to to a grab, which if you are also running something else that wants to do a grab, the installer will error out. Not good. (BTW, I like being able to switch away from the installer and do something productive, which you cannot do under Windows. Point for us!).
Now, I agree that autostart on programs is a BAD IDEA. I turn it off on my W9bluescreen systems, but it does make it easier for luserboy to run his stuff. I have a friend who has a five year old. She can play her games on the computer, since they all autostart. She finds the CD, puts it in the drive, and there is her game. There is no way my friend will leave that system in Linux and have to deal with a cranky kid! But, autostart should be an option (perhaps a daemon that is launched (or not) on a per user basis when you log in AT THE CONSOLE.
Lastly, the LSB needs to push some sort of standardized installer, be it RPM, DEB, or supercilfragoulishexpealidous. But it needs to be standardized, it needs to be able to set up any and all desktop managers (i.e. the various GUIs need to settle on a standard, common way to set file associations, icons, etc.), and it needs to work both as a GUI and CLI app.
Make no mistake: if we do not make Linux desktop ready, Bill will bide his time, gather his strength, and when he is ready, destroy us. In his world view, "There can be only one."
What YOU don't realize is: Z80 - $2 RAM - $2 Glue logic - $2 ROM - $2 Board space - priceless, if your project must be small (hand-held), light (space), or both.
Furthur, if I already HAVE to have the FPGA for other reasons, incremental cost of the additional gates for the core is only a few dollars.
It's the same trick as the Cellular community uses: give the hardware away and sell the service.
How do they make money?
They prevent you from going to a cable modem or DSL. This prevents their competition from getting a foothold, and helps keep them profitable.
They short-change the hardware. $8/month from 100 people will go a long way to paying for three old 28.8Kbps modems. So what if you never get in? (Works for AOL!)
They wait a year, and jack the prices up. "Oh, well you are connecting at 56K. Read your service contract, and you will see we only guaranteed 28.8, so you are using our premium service, which is $20/month."
Don't go there, don't buy that, get a real computer or get nothing at all, you will be happier in the long run.
Alan Greenspan will tell you that he is a macropsychologist more than an economist
I cannot remember the show, other than it was on PBS, but they were interviewing a FED economist and he said something that cause me to have an epiphany. What he said was something like:
We have all this information, and we should use it to provide feedback...
Now, I'm an electrical engineer, and feedback is something I understand far better than any economist. One of the things you learn in control systems theory is that a system with lots of phase lag (i.e. it takes a long time from control input to response output) will be unstable if the loop feedback isn't lowpass filtered. The economy has a great deal of lag to it, but the FED's reactions to the vagarities of the market is almost instantanious. Also, you don't hit a system with step functions (sudden changes from one value to another) unless you want it to ring. The FED always makes step changes in the prime rate. Result: instablity.
My solution: The FED should be required to make all changes over a long period (say, six months), with a small increment each week. This would do two things:
The market would not panic everytime Greenspan clears his throat.
The power of the FED would be reduced
The system would be more stable
Of course, since the FED would lose power under my plan, that is exactly why it will never happen.
What many of you don't realize is the importants of the statement "Synthesizable core". That means that I can get the VHDL description of the device, and drop it into an FPGA or ASIC (field programmable gate array and application specific integratted circuit) and make a single-chip solution to some embedded problem. Who would buy the actual silicon from Zilog? But give me a small, simple core in my ASIC and I'll be a happy man.
BTW, I too learned my assembly on a Z80. I learned real quick what happens when an ISR takes too long by writing code to blink the screen on a TRS-80 from interrupt, a lesson that I have carried forward into my years as a professional embedded software designer.
No need to topple Microsoft, I think, just stand back, they're doing quite well all on their own. Does one yell
"Software!" when a behemoth starts to fall?
I've been an EE and embedded systems designer for twelve years now. I'm one rung from the top of the ladder where I work. If a resumee comes through for someone applying in my department, I'm one of the people who reviews it, and interviews the person when they come in.
Having established my credentials, here's my $2.00 worth (because I think my opinion is worth more than $0.02):
1) If you are going to be a bit banger or electron pusher (for anything other than life support or power distribution), a PE is a luxury, not a necessity. This is not to say luxuries are bad, but: 2) Of more importance (and quicker payoff) are the other certifications you can get in other areas. Get certified on Real time OS's, CAD packages, VHDL, SDL, etc. (things a potential employer would otherwise have to train you in) Things like CNE, MSCE, etc. are useful only if you want to go into Information Services (i.e. be a network admin.) 3) Consider getting an amatuer radio license. Especially if you plan on going into any company that does anything remotely pertaining to radio, this is a very good (and cheap) way to break the ice. Plus, you will get hands on experience in making electronic things. 4) Whatever certs you get, whatever you put on your resumee, be damn sure you can actually "walk the walk." Yes, you have to make yourself look good in the resumee, but I cannot count the number of times I've seen "Experience in C++" on a resumee and the person couldn't even tell me what a virtual function was! "Bzzzt! Thank you for playing" 5) If you want to make the most money, get a lobotomy, errr, MBA (Masters of Business Administration). Sad to say, but being a creator doesn't pay as well in the long run as being a Pointy Haired Boss. 6) Learn your people skills. Learn to speak clearly in presentations, learn to manage people well, learn to write clearly. Even if you stay in a technical track, you WILL eventually be a project lead. Learn to take criticism well, learn to critisize well. Learn the "leading question": instead of saying "Your design is crap because if Vcc drops.5V, the switching regulator will blow", say "I'm confused, what happens to the regulator if Vcc drops?" 7) Learn personal integrity (this is part of what the PE covers): You boss pays you for your opinion: if he cannot trust you, why is he paying you. Never lie when giving your professional opionon. If you don't know, say so (the three wisest words in English are "I don't know"). Take responsiblity for your actions: If you screwed up, say "I screwed up, and this is how I plan on fixing the problem." 8) If you do get a PE, you had better carry malpractice insurance. Just like a doctor.
The last piece of advise I can give any graduating EE: You may thing you are Gods give to engineering, but remember this: That technician, the guy who has a six-week certificate from DeVry, has been doing this for years longer than you. He remembers more about the mechanics of electrical design after his first cup of coffee in the morning than you know. You are a first leutenant, he is a seargent with hash marks down his sleeve. Respect him, listen to him, and learn!
My shop has also hired contractors instead of staff. The experiences I've had have been almost universally bad: contractors who do not know what they claim (e.g. a DSP "contractor" who did not know what a MAC (multiply and accumulate, a staple of DSP) instruction was), contractors who teach themselves something on our dollar, etc. However, this can be explained by my bosses screwing up and hiring bozos rather than checking the people out carefully.
However, the basic issue is the "rent vs. buy" issue: do you rent a skill (contractor) or buy it (employee). You use the same metric for this as you do for capital equipment:
If the need is short term or the cost is high, you rent.
If the need is long term or the cost is low, you buy.
For example, you buy a car in your home town, since you are going to need it for a long time. However, you rent a car if you are on a business trip, since you need it only for a few days.
With skills, you ask yourself "Am I going to need this skill in a year?" If the answer is yes, you hire an employee. If you can honestly say "No", then contracting makes sense. Where I work, I am pushing the boss to hire an X Windows programmer, because we are using that in a new project that is going to be around for a while. We could solve our immediate need by getting a contractor, but then we would not have the person around in a year when I need some more work done. The only way a contractor would make sense in this context would be if I teamed him up with one of my people to help my guy learn X.
The biggest single reason most shops contract rather than hire is that the contractor doesn't count against the department's headcount, even though he costs the department twice what a regular employee costs (can you say beancounter? I knew you could).
It sounds to me like your boss has fallen into the Dilbert trap: "Anybody who works here is a moron, anybody who comes from outside must be a genius".
Hire the skills you will need long-term. Pay them what they are worth, let them develop their skills, and above all else listen to what they tell you! After all, if you cannot trust their opinions, why are you paying their salary?
#include "std_disclaimer.h" The views expressed in this post are mine, not my employers.
The only problem with this is that is assumes the vast majority of people are sapient.
Most people won't understand the message, and it will just hack off ISPs.
Besides, most trolls are AC's, and therefor you can only filter them via IP, which is bad, because most AC's are on dynamic IP of some sort, and therefor you could hurt somebody else.
I think that Rob is choosing the best of a bad set of paths. The only real solution is to crank up your comment level to 1 or more, and let the trolls yell into a barrel until they lose their voices. The same solution as on UselessNet.
Don't answer these sub-sapients. Ignore them, and get on with life. Don't waste your points on moderating them down, just leave them at 0 where they belong, and don't read 0 level posts.
You guys need to use a low-pass filter before your sound card. Since you sample at 24kSamples/sec, you need to block anything 12kHz and over, or else it aliases back into the spectrum and makes the signal sound terrible!
Either sample at a high rate and use sox to downsample it before encoding, or get a lowpass filter in front of the sound card. I'd recommend a 3kHz corner frequency and about 4 poles (a Sallen-Key filter would do nicely).
Other than that, I look forward to each episode. You need to do a JPEG or PNG of a movie-theater with the three of you in silhouette (a la MST3K).
Windows NT in the public safety field
on
911 Calls Linux
·
· Score: 2
I went to the Association of Public Safety Communications Officials (APCO) exposition in Minniapolis, MN a few weeks ago. This is the show for the folks who make the radios, dispatch equipement, mobile computers, etc. for the police and fire departments nationwide.
I was dismayed at the number of dispatch systems, mobile computers, and radio setups that used Windows NT (or worse yet, 95/98) as the core. I saw about 3 vendors selling Unix (Solaris) solutions, and if there was a Linux box there (other than my laptop) I didn't see it.
The reason is a matter of demand: Most shops want stuff that works under Windows, and anything else does not compute. This is due to ignorance.
The solution is the same for this as anywhere else: ask for Linux/Unix versions of the code. If the radio programming software only runs under Windows, ask the radio manufacturer to port the code to Linux. If the test equipment only speaks to Windows, ask the manufacturer to port to Linux. If the dispatch software runs under Windows, ask for a port. If the mobile data unit runs Windows, ask if it will run Linux.
If enough shops ask for Linux support, it will be granted.
#include The views expressed in this post are my own, not my employers. I am employed by a company that makes test equipment for police radios. Personally, I'd love to have customers asking for Linux support! It'd show the marketing people I know what I'm talking about!
The pressure at the bottom of the sea would crush a sealed object containing surface level pressure. Remember, the pressure delta from sea level to space is 15 pounds per square inch, but the pressure difference from sea level to the bottom of the sea is hundreds of thousands of pounds per square inch.
First: Yes, if they did manage to squeeze some mass into a quantum black hole, the hole would radiate away it's entire mass as Hawking radiation in fraction of a femtosecond. However, that mass would only be the mass of two ions (plus whatever relativistic mass they had put on). I expend more energy that that each second keeping my skin warm. Second, if I remember my quantum chromodynamics correctly, a single quark cannot exist as a real particle. It must either be bound up to its antiparticle, or to enough other quarks to balance out to "colorless". Therefor, there is no way a single quark of any flavor could come out of the reaction. I think the place that wrote that article seized upon the name "strange" as somehow menecing, and is trying to stir up controversy (which, given the traffic on/., was a successful strategium). Go read the actual science site, and go get some books (Brief History of Time and the like) and then make up your mind.
If possible, could you get them to put the Real Video files themselves (the.RAM, not the.RM) files on an HTTP/FTP server? I have access to a T1 link at work, but they don't like us running Real Audio/Real Video. Besides, I prefer to download stuff FTP, then listen to it, so that I don't get blarked if I have a bandwidth crunch.
Slashdot Mirror
Furthur, this only deals with demodulating the signal once it has been received. You still need the RF front-end to do:
and that ain't easy.
Then, you have to decode the protocol (that's where the part comes in) plus any higher level signalling (usually, you'd used another microcontroller for this).
This isn't a one-chip phone by any means. A multiprotocol phone is going to be a mother of a beast to code.
Now, I work in the comms field, designing test equipment that uses the Motorola 56300 family of DSPs, and I am less than impressed. I hope this new chip fixes the bugs in the chip, the stupid DMA controller (MOTO DSP GUYS: ever hear of scatter/gather?), the brain-dead instruction set (who ever heard of a DSP without a saturating add instruction!) and oddball word length (let's move from 24 bits to 32, shall we?), and small stack size. Not to mention that this DSP is a fixed point only DSP, no floating point on board.
If anyone out there has a good link to this chip's data sheet (I looked on Moto's site, no dice), please post it.
To what extent is the Royal Family involved with the site (e.g. content creation)
OK, this helps us how? Without the Amiga chipset, much of what made the Amiga interesting is not possible. So, to what end can I spend my free time on this code?
Now, don't get me wrong. I love the idea of open sourcing the Amiga OS, just as I loved the idea of Caldera open sourcing GEM and Atari open sourcing TOS. However, how much activity has the latter generated?
All I am saying is that this is interesting, but not earth shattering. Perhaps, if the rumors are true, and Amiga and Transmeta release an ubercomputer, this will be a seminal moment in personal computer history, but unless that happens, I fear that this won't even be a footnote. The hacker community won't take up the code and make it wonderful, it will just sit around taking up FTP server space.
Unfortunately.
Perhaps it is time to unleash the mighty power of the /. effect upon Apple: Why doesn't everybody who has bitched about not having Sorenson support under Xanim e-mail Apple, and ask them nicely to allow Sorenson to allow Mr. Podlipec to make a codec plug-in for Xanim. Since Apple is claiming to support Open Source, lets give them a chance to walk the walk, and not just talk the talk.
No, that's not how it works. The wires your phone uses can handle signals from DC to about 1MHz. The voice line uses 300Hz - 3kHz. Then there is a gap to about 10kHz, then DSL uses from about 20kHz to 1MHz. There's a splitter at your house, and the high frequency signals go to the DSL modem, and the low freqs to your phone. There's the same at the telco office: the low freqs go to the phone switch, and the high freqs to another DSL modem. So, your voice is still analog (at least until it hits the switch), and your DSL data never touches the phone network. That's why the phone companies love DSL: your switch usage goes back to the "old model" of usage (lots of short calls), and your data usage is handled seperately.
Now, look at what MS is doing:
That has about as much chance of being "open" as the Windows API does: Sure, you may be able to read & write data (more or less, given how much of the extended HTTP standard you get), and you may be able to parse some of the tags, but MS can simple define new tags that only they understand.
I'll beleive in the promise of XML when good, well defined standards exists for defining tags, and standard tags are defined for standard things.
I'll beleive in SOAP when entropy runs backwards.
This is an example of Microsoft's long range planning: gut the department now, then start acting up again.
More "astroturf" style grassroots marketing.
Well, all I can say is write your Congressdrones and complain.
What I hate about this stuff is that there is no option to just freaking download the damn video!
.RA file, so that we can just download it.
Look, I live in the country. I have about as much chance of getting broadband access this millenium (i.e. before 2001 because that is the millenium dammit) as Bill Gates using Linux for his house. (less, actually). All I want to do is start a download at the paultry 28.8 kbps I get, and walk away. I'll watch the video when it is finished, whenever that is.
Some things they stream to keep you from saving a copy, but something by maddog? Where's the "open source" in that? There ought to be an FTP site with the actual
This is one of the many areas that Geeks In Space gets right: I can just download the file, and listen when I get it!
How is this any different from the printfilters package Red Hat has been shipping for some time now?
Given your stand on OSS and the 2nd Amendment, have you ever considered the ultimate in PHBdom: Running for President of the US?
And after you came down from the Nitrosolv fumes, what is your favorite handgun?
I found the cluelessness of the article repeatedly saying the hotel would be "intergalactic" quite offensive.
It probably wouldn't even be interplanetary, being well within the volume where the Earth's gravitational field is domimant.
I do wish reporters would pay attention to the words they use!
Somebody (if I ever get the time I'll look into it myself) needs to modify autofs so that a mountpoint can be specified, and will show up in the automount directory, but won't be mounted until you actually try to enter that directory and retrieve the file list. That way, under [Gnome|KDE|Enlightenment|*], there will be an icon for drooling user boy to click on. This would go a long way to making Linux easier to use.
Now, re: Loki's install. The biggest problem with the install script that I had was that it was not tagged as executable on the CD. That means that:
Now, I agree that autostart on programs is a BAD IDEA. I turn it off on my W9bluescreen systems, but it does make it easier for luserboy to run his stuff. I have a friend who has a five year old. She can play her games on the computer, since they all autostart. She finds the CD, puts it in the drive, and there is her game. There is no way my friend will leave that system in Linux and have to deal with a cranky kid! But, autostart should be an option (perhaps a daemon that is launched (or not) on a per user basis when you log in AT THE CONSOLE.
Lastly, the LSB needs to push some sort of standardized installer, be it RPM, DEB, or supercilfragoulishexpealidous. But it needs to be standardized, it needs to be able to set up any and all desktop managers (i.e. the various GUIs need to settle on a standard, common way to set file associations, icons, etc.), and it needs to work both as a GUI and CLI app.
Make no mistake: if we do not make Linux desktop ready, Bill will bide his time, gather his strength, and when he is ready, destroy us. In his world view, "There can be only one."
What YOU don't realize is:
Z80 - $2
RAM - $2
Glue logic - $2
ROM - $2
Board space - priceless, if your project must be small (hand-held), light (space), or both.
Furthur, if I already HAVE to have the FPGA for other reasons, incremental cost of the additional gates for the core is only a few dollars.
How do they make money?
Don't go there, don't buy that, get a real computer or get nothing at all, you will be happier in the long run.
I cannot remember the show, other than it was on PBS, but they were interviewing a FED economist and he said something that cause me to have an epiphany. What he said was something like:
Now, I'm an electrical engineer, and feedback is something I understand far better than any economist. One of the things you learn in control systems theory is that a system with lots of phase lag (i.e. it takes a long time from control input to response output) will be unstable if the loop feedback isn't lowpass filtered. The economy has a great deal of lag to it, but the FED's reactions to the vagarities of the market is almost instantanious. Also, you don't hit a system with step functions (sudden changes from one value to another) unless you want it to ring. The FED always makes step changes in the prime rate. Result: instablity.
My solution: The FED should be required to make all changes over a long period (say, six months), with a small increment each week. This would do two things:
Of course, since the FED would lose power under my plan, that is exactly why it will never happen.
What many of you don't realize is the importants of the statement "Synthesizable core". That means that I can get the VHDL description of the device, and drop it into an FPGA or ASIC (field programmable gate array and application specific integratted circuit) and make a single-chip solution to some embedded problem. Who would buy the actual silicon from Zilog? But give me a small, simple core in my ASIC and I'll be a happy man.
BTW, I too learned my assembly on a Z80. I learned real quick what happens when an ISR takes too long by writing code to blink the screen on a TRS-80 from interrupt, a lesson that I have carried forward into my years as a professional embedded software designer.
No, one yells "BLUESCREEN!"
I've been an EE and embedded systems designer for twelve years now. I'm one rung from the top of the ladder where I work. If a resumee comes through for someone applying in my department, I'm one of the people who reviews it, and interviews the person when they come in.
.5V, the switching regulator will blow", say "I'm confused, what happens to the regulator if Vcc drops?"
Having established my credentials, here's my $2.00 worth (because I think my opinion is worth more than $0.02):
1) If you are going to be a bit banger or electron pusher (for anything other than life support or power distribution), a PE is a luxury, not a necessity. This is not to say luxuries are bad, but:
2) Of more importance (and quicker payoff) are the other certifications you can get in other areas. Get certified on Real time OS's, CAD packages, VHDL, SDL, etc. (things a potential employer would otherwise have to train you in) Things like CNE, MSCE, etc. are useful only if you want to go into Information Services (i.e. be a network admin.)
3) Consider getting an amatuer radio license. Especially if you plan on going into any company that does anything remotely pertaining to radio, this is a very good (and cheap) way to break the ice. Plus, you will get hands on experience in making electronic things.
4) Whatever certs you get, whatever you put on your resumee, be damn sure you can actually "walk the walk." Yes, you have to make yourself look good in the resumee, but I cannot count the number of times I've seen "Experience in C++" on a resumee and the person couldn't even tell me what a virtual function was! "Bzzzt! Thank you for playing"
5) If you want to make the most money, get a lobotomy, errr, MBA (Masters of Business Administration). Sad to say, but being a creator doesn't pay as well in the long run as being a Pointy Haired Boss.
6) Learn your people skills. Learn to speak clearly in presentations, learn to manage people well, learn to write clearly. Even if you stay in a technical track, you WILL eventually be a project lead. Learn to take criticism well, learn to critisize well. Learn the "leading question": instead of saying "Your design is crap because if Vcc drops
7) Learn personal integrity (this is part of what the PE covers): You boss pays you for your opinion: if he cannot trust you, why is he paying you. Never lie when giving your professional opionon. If you don't know, say so (the three wisest words in English are "I don't know"). Take responsiblity for your actions: If you screwed up, say "I screwed up, and this is how I plan on fixing the problem."
8) If you do get a PE, you had better carry malpractice insurance. Just like a doctor.
The last piece of advise I can give any graduating EE: You may thing you are Gods give to engineering, but remember this: That technician, the guy who has a six-week certificate from DeVry, has been doing this for years longer than you. He remembers more about the mechanics of electrical design after his first cup of coffee in the morning than you know. You are a first leutenant, he is a seargent with hash marks down his sleeve. Respect him, listen to him, and learn!
However, the basic issue is the "rent vs. buy" issue: do you rent a skill (contractor) or buy it (employee). You use the same metric for this as you do for capital equipment:
If the need is short term or the cost is high, you rent.
If the need is long term or the cost is low, you buy.
For example, you buy a car in your home town, since you are going to need it for a long time. However, you rent a car if you are on a business trip, since you need it only for a few days.
With skills, you ask yourself "Am I going to need this skill in a year?" If the answer is yes, you hire an employee. If you can honestly say "No", then contracting makes sense. Where I work, I am pushing the boss to hire an X Windows programmer, because we are using that in a new project that is going to be around for a while. We could solve our immediate need by getting a contractor, but then we would not have the person around in a year when I need some more work done. The only way a contractor would make sense in this context would be if I teamed him up with one of my people to help my guy learn X.
The biggest single reason most shops contract rather than hire is that the contractor doesn't count against the department's headcount, even though he costs the department twice what a regular employee costs (can you say beancounter? I knew you could).
It sounds to me like your boss has fallen into the Dilbert trap: "Anybody who works here is a moron, anybody who comes from outside must be a genius".
Hire the skills you will need long-term. Pay them what they are worth, let them develop their skills, and above all else listen to what they tell you! After all, if you cannot trust their opinions, why are you paying their salary?
#include "std_disclaimer.h"
The views expressed in this post are mine, not my employers.
The only problem with this is that is assumes the vast majority of people are sapient.
Most people won't understand the message, and it will just hack off ISPs.
Besides, most trolls are AC's, and therefor you can only filter them via IP, which is bad, because most AC's are on dynamic IP of some sort, and therefor you could hurt somebody else.
I think that Rob is choosing the best of a bad set of paths. The only real solution is to crank up your comment level to 1 or more, and let the trolls yell into a barrel until they lose their voices. The same solution as on UselessNet.
Don't answer these sub-sapients. Ignore them, and get on with life. Don't waste your points on moderating them down, just leave them at 0 where they belong, and don't read 0 level posts.
You guys need to use a low-pass filter before your sound card. Since you sample at 24kSamples/sec, you need to block anything 12kHz and over, or else it aliases back into the spectrum and makes the signal sound terrible!
Either sample at a high rate and use sox to downsample it before encoding, or get a lowpass filter in front of the sound card. I'd recommend a 3kHz corner frequency and about 4 poles (a Sallen-Key filter would do nicely).
Other than that, I look forward to each episode. You need to do a JPEG or PNG of a movie-theater with the three of you in silhouette (a la MST3K).
I went to the Association of Public Safety Communications Officials (APCO) exposition in Minniapolis, MN a few weeks ago. This is the show for the folks who make the radios, dispatch equipement, mobile computers, etc. for the police and fire departments nationwide.
I was dismayed at the number of dispatch systems, mobile computers, and radio setups that used Windows NT (or worse yet, 95/98) as the core. I saw about 3 vendors selling Unix (Solaris) solutions, and if there was a Linux box there (other than my laptop) I didn't see it.
The reason is a matter of demand: Most shops want stuff that works under Windows, and anything else does not compute. This is due to ignorance.
The solution is the same for this as anywhere else: ask for Linux/Unix versions of the code. If the radio programming software only runs under Windows, ask the radio manufacturer to port the code to Linux. If the test equipment only speaks to Windows, ask the manufacturer to port to Linux. If the dispatch software runs under Windows, ask for a port. If the mobile data unit runs Windows, ask if it will run Linux.
If enough shops ask for Linux support, it will be granted.
#include
The views expressed in this post are my own, not my employers. I am employed by a company that makes test equipment for police radios. Personally, I'd love to have customers asking for Linux support! It'd show the marketing people I know what I'm talking about!
(B.T.W. the APCO homepage is http://www.apcointl.org)
The pressure at the bottom of the sea would crush a sealed object containing surface level pressure. Remember, the pressure delta from sea level to space is 15 pounds per square inch, but the pressure difference from sea level to the bottom of the sea is hundreds of thousands of pounds per square inch.
First: Yes, if they did manage to squeeze some mass into a quantum black hole, the hole would radiate away it's entire mass as Hawking radiation in fraction of a femtosecond. However, that mass would only be the mass of two ions (plus whatever relativistic mass they had put on). I expend more energy that that each second keeping my skin warm. /., was a successful strategium).
Second, if I remember my quantum chromodynamics correctly, a single quark cannot exist as a real particle. It must either be bound up to its antiparticle, or to enough other quarks to balance out to "colorless". Therefor, there is no way a single quark of any flavor could come out of the reaction.
I think the place that wrote that article seized upon the name "strange" as somehow menecing, and is trying to stir up controversy (which, given the traffic on
Go read the actual science site, and go get some books (Brief History of Time and the like) and then make up your mind.
If possible, could you get them to put the Real Video files themselves (the .RAM, not the .RM) files on an HTTP/FTP server? I have access to a T1 link at work, but they don't like us running Real Audio/Real Video. Besides, I prefer to download stuff FTP, then listen to it, so that I don't get blarked if I have a bandwidth crunch.