The giant multi-hundred pin connector on a dock suffers wear and tear too.
I just don't see the big deal. In a business situation, it means two things to plug in instead of one (plug in the monitor and keyboard, Apple keyboards have USB ports, and the mouse connects to one of those). No external speakers - my computer should be seen but not heard in the office, and if I need to hear stuff, I use headphones rather than inflict it on my colleagues. (I have the sound permanently muted on my work desktop already, I want sound off by default because websites sometimes unexpectedly make annoying sounds - like an ad that Slashdot carried a while back - don't know whose it was but it featured the sound of a slamming door). The expense of a dock simply isn't worth it to save the effort of having to plug one extra cable.
In my observation, MS get it wrong on one extreme or the other, they can never get it just right. They either make it user hostile and inconsistent or they make it downright patronising (i.e. Clippy and the teletubby default XP theme). For 3d effects in Vista, instead of using subtle 3d effects and transparency to aid the eye, they make them massively over the top in a "HEY LOOK WOW I CAN DO 3D AND TRANSPARENCY" kind of way, making the UI cluttered.
I've made LED lighting units for my Dad's boat, so I'm passingly familiar with the subject.
No, it's nowhere near as difficult as you state. The Luxeon Rebel with the Lambertian pattern with no lens put over the LED has an illumination angle approaching 180 degrees. That is NOT narrow angle. To make an omnidirectional lamp you'd only need to put two of them back to back, or if you're fussy you can put three in a triangular arrangement. I'm using these LEDs as downlighters though, so I did use a lens to make them *more directional* because the beam was far too wide. They are plastic, and inexpensive, and contain a reflector. The lenses I used give the lighting pattern of a typical halogen downlighter.
The problem with LEDs isn't the illumination angle, simply, at the moment, it's cost. The best Luxeon Rebels are now about the same efficiency as a compact flourescent, but a 3W LED costs about two or three times the price of a 9W compact florescent - and that's without the lens or power supply. People normally can't see past the acquisition cost either, at the 50,000 hour lifespan (guaranteed to produce 80% of rated lumens at 50,000 hours at rated current). If you used the light for 8 hours a day, every day, that's 17 years - and they don't have the warm-up issues of a CF lamp.
The future is LEDs, just give it three or four years, and the costs will come down to a level that won't make most people wince.
You misunderstand quite completely. A capacitor (or any power source for that matter) will only release energy as fast as its demanded. If you had the ideal 12V voltage source (i.e. a 12V voltage source that could give infinite current given the ideal short circuit), and put a 12 watt bulb across it, only 1 amp will flow at 12 volts.
Read about Ohm's Law here to get a deeper understanding: http://www.ibiblio.org/kuphaldt/electricCircuits/ (and if you want further understanding still, read about the Thevenin Equivalent Circuit).
Building a SCSI interface is something easily within the reach of the electronics hobbyist, as is an IDE interface (in fact, an IDE interface is trivially easy to make for anyone with some practical digital electronics experience). There are USB FIFO chips on the market that should make it relatively straightforward to make a USB - SCSI adapter for old drives. While 8 inch disc drives are as common as rocking horse shit, 5.25 and 3.5 inch disc drives were made in huge numbers so there's no real shortage of those. (My 25 year old 5.25 inch discs all read perfectly), and 5.25 inch drives aren't difficult to repair. Floppy disc controllers are still pretty easy to get hold of - any self respecting electronics junk shop should have some WD177x types in stock still, and they come up on ebay quite frequently. Even when they get hard to find, it would be perfectly possible to roll-your-own disc controller with a CPLD and a handful of discrete components.
Actually, I do. Someone at work has the same make and model of car as mine in the same colour, so I have to look so as to not try and drive off in the wrong car.
Generally, in aviation, if it looks good it flies good, and if it looks bad it also flies bad. There are some exceptions to this, for example the Fairey Gannet, the ugliest aircraft ever built, actually flies quite nicely, and most helicopters which are so ugly the earth repels them.
These machines were in 35 different locations, not all in the same place...and other machines didn't have a problem. It was only the HP desktop machines. This is actually the second "bad caps" issue we've had with these machines, all the motherboards were replaced under warranty, as they started failing (with bulging caps) within 6 months of being installed. The power supplies started failing 2.5 to 3 years after the machines were put into service.
The capacitors were (according to the rating printed on the can) adequately sized, they were just crap.
Don't forget embedded. Everyone forgets embedded, despite you having probably more embedded devices in their house than desktop computers!
Lots of assembly language used there. When you're trying to fit your washing machine code in an Atmel AVR with 1K of flash and 64 bytes of RAM, optimisation is important and throwing more hardware at it is expensive - because each instance of the software comes with an instance of the hardware, too. Anyone who likes getting down and dirty with the bare iron, try doing embedded. Or if it's just some fun programming you want to do, get a retrocomputer like a Sinclair Spectrum or Commodore 64.
The story is cute, but it also must be remembered that all the redundant code may also have unknown bugs and security holes. What would the cost of being pwned be? Would being pwned cost the company more than having good quality code without redundant and possibly exploitable internet-facing code that probably never gets tested?
It's worse than that, in around 1999 or so, I got a $10 charge on my landline bill (then GTE, now Verizon) for a "third party call" from Florida to New York when I wasn't even in the country. Apparently you can call the operator and have your call put on someone else's bill unless they have 3rd party call blocking...which costs money.
I was really tempted to cancel tone dialing service (GTE charged a few cents a month to let you tone dial!!!!) just to spite them.
The real WTF is running a consumer OS on a submarine, with consumer hardware - most of which, if you look at the data sheets for the ICs, say "not to be used in life critical systems". I think pretty much any system on a sub could be considered life critical.
Base 10? The world would be a lot better if we switched to base 16. It'd save lots of CPU cycles in the embedded world, making conversions from binary to what's displayed on the LCD much easier.
We've had the opposite experience of HP power supplies, we just had to replace 70 HP supplies. When machines started failing in the field, I found that there was massive amounts of ripple on the 12v and 5v lines. When I disassembled the PSU it wasn't hard to tell why - bulging and leaking capacitors.
A computer doesn't just "draft up a circuit diagram" and connect chips like Lego. PCB design and layout is a real engineering discipline that takes real engineers to make a real working product. The computer tools are just that, they provide tools to make the engineer's job easier, but they don't magic up a circuit diagram automatically. It takes a real engineer with real experience to make a high speed digital design work properly. It is much more tricky than writing software because the real world (which includes things like parasitic inductance and capacitance, trace impedances, RF interference) heavily impinges on high speed digital design. There are almost books written on the subject of using decoupling capacitors alone and tomes of information on innocuous subjects that to the untrained eye look simple, like power and ground planes. In a high speed digital design, a PCB trace isn't simply a wire linking A to B to the engineer who is deciding where it should go on the PCB.
It takes a significant amount of time and effort, and a significant amount of knowledge to put together even relatively straightforward high speed digital designs. With the iPhone you also have to cram it into a very constrained amount of space, too. You can bet the engineers who laid out the PCB spent a great deal of time making sure that not only it would fit, but the resultant device would work and pass FCC testing.
Touch computing may be mainstream for handheld devices, but it will be a long time before the mouse is replaced on a desktop PC. What these prognosticators always seem to forget about PC displays is the display is vertical and in front of you. It gets tiring if you have to hold your arm up to touch stuff on the screen all day. Your arm does at least get to rest on the table if you're using a mouse or trackpad.
For a PC, the prognosticators also seem to forget that the mouse is good enough, and it's tremendously difficult to replace "good enough" in three years. Touch interfaces on a desktop system don't offer any benefits over a mouse (unlike on handheld devices, where a touch interface is obviously very very much better than any other kind of pointing device). For laptops, again, the vertical screen problem and arm-tiredness/screen smudging issues persist, and people find trackpads good enough with a touch screen not really offering any worthwhile benefit on a full size laptop.
On my last project for an extremely large customer (with an equally huge project), the good line managers were our (the developers) advocates, they took the bullshit so we didn't have to, and determined the "big picture". They didn't manage who was doing which particular piece of code - that was down to the developers to organize themselves. Managing the development team was less about managing the people in it (the developers could organize themselves) but being an advocate for the team, and making sure that the people who knew how to do the stuff were fully involved in decisions affecting them. Developers were not merely involved but critical to things such as sizing parts of the project, so that unrealistic schedules were not set. The line manager's job was in this case often to tell upper management "this is why it's going to take this long" in terms they could understand, and persuade upper management not to cause a disaster by compressing schedules or adding more work.
It resulted in a productive development team which did not have to do unpaid overtime, and delivered a quality product to the customer - earning a very large sum of money for the company.
What's wrong with just a normal <A HREF=... for navigation? Even JavaScript would be better for navigation than a proprietary, Microsoft-only language. If you're using proprietary languages, you don't have a web site, you have an Internet Explorer site.
I think it's the syllabus. Not necessarily just bad teachers, but otherwise good teachers having to teach to the exam, and the method in teaching a particular subject being wrong in the first place.
I have always considered myself bad at (human) languages. After something like 7 years of school French, I couldn't even reserve a hotel room in French, let alone carry out a very basic conversation. However, I discovered that wasn't really the case; it wasn't that I was bad at languages, it's just that languages are taught fundamentally in the wrong way in this country. It can't just be me considering how many British people are monolingual despite being compelled to do a language GCSE.
In May this year, I needed to learn some Spanish so I wouldn't feel like a complete Imperialist when visiting Palma, and did a beginner's course on the BBC website. The course was fun to do, and I found myself wanting to learn more, and since May, I have learned more Spanish in 7 months than I did French in 7 years of school French, supposedly when the mind is so much more able to pick up another language. Already, I can understand most stories on Spanish news websites like El Pais or BBC Mundo, and I've had meaningful conversations online with Spanish people. I've even managed to laugh at a joke in Spanish.
So I don't think British people are inherently bad at languages, I just think the method languages are taught is bad, and those who do learn another language at school do so in spite of their education, not because of it. Interestingly, on Radio 4's PM programme a couple of months back, there was a report on language learning in the UK, and how so many young people want to work in France or Spain for a while, but never get to grips with the language. They had some girl on there, who had been doing French for 5 years and got a good grade at GCSE, but when the reporter asked her to say something simple ("What did you do this morning?") in French, she couldn't.
In summary, I agree with your observation - the content is wrong, and often so is the teaching method.
If you live in the United States, you need a degree.
If you live in Europe, for the most part, you don't for an IT job, but you will for a software development job. "IT" is really a trade like plumbing. You don't need a degree to pull cat5 cables, fix broken printers, or write shell scripts to automate various repetitive jobs, just intelligence and experience.
Slashdot Mirror
I'd say buggery, more than sexual reproduction.
The giant multi-hundred pin connector on a dock suffers wear and tear too.
I just don't see the big deal. In a business situation, it means two things to plug in instead of one (plug in the monitor and keyboard, Apple keyboards have USB ports, and the mouse connects to one of those). No external speakers - my computer should be seen but not heard in the office, and if I need to hear stuff, I use headphones rather than inflict it on my colleagues. (I have the sound permanently muted on my work desktop already, I want sound off by default because websites sometimes unexpectedly make annoying sounds - like an ad that Slashdot carried a while back - don't know whose it was but it featured the sound of a slamming door). The expense of a dock simply isn't worth it to save the effort of having to plug one extra cable.
Matt Smith is hardly beautiful. I watched the clip on the BBC News website, and he looks older than his years.
On the other hand he did write Jet Set Willy. Oh, wrong Matt Smith!
In my observation, MS get it wrong on one extreme or the other, they can never get it just right. They either make it user hostile and inconsistent or they make it downright patronising (i.e. Clippy and the teletubby default XP theme). For 3d effects in Vista, instead of using subtle 3d effects and transparency to aid the eye, they make them massively over the top in a "HEY LOOK WOW I CAN DO 3D AND TRANSPARENCY" kind of way, making the UI cluttered.
I've made LED lighting units for my Dad's boat, so I'm passingly familiar with the subject.
No, it's nowhere near as difficult as you state. The Luxeon Rebel with the Lambertian pattern with no lens put over the LED has an illumination angle approaching 180 degrees. That is NOT narrow angle. To make an omnidirectional lamp you'd only need to put two of them back to back, or if you're fussy you can put three in a triangular arrangement. I'm using these LEDs as downlighters though, so I did use a lens to make them *more directional* because the beam was far too wide. They are plastic, and inexpensive, and contain a reflector. The lenses I used give the lighting pattern of a typical halogen downlighter.
The problem with LEDs isn't the illumination angle, simply, at the moment, it's cost. The best Luxeon Rebels are now about the same efficiency as a compact flourescent, but a 3W LED costs about two or three times the price of a 9W compact florescent - and that's without the lens or power supply. People normally can't see past the acquisition cost either, at the 50,000 hour lifespan (guaranteed to produce 80% of rated lumens at 50,000 hours at rated current). If you used the light for 8 hours a day, every day, that's 17 years - and they don't have the warm-up issues of a CF lamp.
The future is LEDs, just give it three or four years, and the costs will come down to a level that won't make most people wince.
You misunderstand quite completely. A capacitor (or any power source for that matter) will only release energy as fast as its demanded. If you had the ideal 12V voltage source (i.e. a 12V voltage source that could give infinite current given the ideal short circuit), and put a 12 watt bulb across it, only 1 amp will flow at 12 volts.
Read about Ohm's Law here to get a deeper understanding: http://www.ibiblio.org/kuphaldt/electricCircuits/ (and if you want further understanding still, read about the Thevenin Equivalent Circuit).
Our security cameras have been hard disc based for years.
Building a SCSI interface is something easily within the reach of the electronics hobbyist, as is an IDE interface (in fact, an IDE interface is trivially easy to make for anyone with some practical digital electronics experience). There are USB FIFO chips on the market that should make it relatively straightforward to make a USB - SCSI adapter for old drives. While 8 inch disc drives are as common as rocking horse shit, 5.25 and 3.5 inch disc drives were made in huge numbers so there's no real shortage of those. (My 25 year old 5.25 inch discs all read perfectly), and 5.25 inch drives aren't difficult to repair. Floppy disc controllers are still pretty easy to get hold of - any self respecting electronics junk shop should have some WD177x types in stock still, and they come up on ebay quite frequently. Even when they get hard to find, it would be perfectly possible to roll-your-own disc controller with a CPLD and a handful of discrete components.
Actually, I do. Someone at work has the same make and model of car as mine in the same colour, so I have to look so as to not try and drive off in the wrong car.
Because if it were up to the insurance companies, there would be no intersections at all :-)
Generally, in aviation, if it looks good it flies good, and if it looks bad it also flies bad. There are some exceptions to this, for example the Fairey Gannet, the ugliest aircraft ever built, actually flies quite nicely, and most helicopters which are so ugly the earth repels them.
These machines were in 35 different locations, not all in the same place...and other machines didn't have a problem. It was only the HP desktop machines. This is actually the second "bad caps" issue we've had with these machines, all the motherboards were replaced under warranty, as they started failing (with bulging caps) within 6 months of being installed. The power supplies started failing 2.5 to 3 years after the machines were put into service.
The capacitors were (according to the rating printed on the can) adequately sized, they were just crap.
Don't forget embedded. Everyone forgets embedded, despite you having probably more embedded devices in their house than desktop computers!
Lots of assembly language used there. When you're trying to fit your washing machine code in an Atmel AVR with 1K of flash and 64 bytes of RAM, optimisation is important and throwing more hardware at it is expensive - because each instance of the software comes with an instance of the hardware, too. Anyone who likes getting down and dirty with the bare iron, try doing embedded. Or if it's just some fun programming you want to do, get a retrocomputer like a Sinclair Spectrum or Commodore 64.
The story is cute, but it also must be remembered that all the redundant code may also have unknown bugs and security holes. What would the cost of being pwned be? Would being pwned cost the company more than having good quality code without redundant and possibly exploitable internet-facing code that probably never gets tested?
It's worse than that, in around 1999 or so, I got a $10 charge on my landline bill (then GTE, now Verizon) for a "third party call" from Florida to New York when I wasn't even in the country. Apparently you can call the operator and have your call put on someone else's bill unless they have 3rd party call blocking...which costs money.
I was really tempted to cancel tone dialing service (GTE charged a few cents a month to let you tone dial!!!!) just to spite them.
The real WTF is running a consumer OS on a submarine, with consumer hardware - most of which, if you look at the data sheets for the ICs, say "not to be used in life critical systems". I think pretty much any system on a sub could be considered life critical.
Base 10? The world would be a lot better if we switched to base 16. It'd save lots of CPU cycles in the embedded world, making conversions from binary to what's displayed on the LCD much easier.
We've had the opposite experience of HP power supplies, we just had to replace 70 HP supplies. When machines started failing in the field, I found that there was massive amounts of ripple on the 12v and 5v lines. When I disassembled the PSU it wasn't hard to tell why - bulging and leaking capacitors.
A computer doesn't just "draft up a circuit diagram" and connect chips like Lego. PCB design and layout is a real engineering discipline that takes real engineers to make a real working product. The computer tools are just that, they provide tools to make the engineer's job easier, but they don't magic up a circuit diagram automatically. It takes a real engineer with real experience to make a high speed digital design work properly. It is much more tricky than writing software because the real world (which includes things like parasitic inductance and capacitance, trace impedances, RF interference) heavily impinges on high speed digital design. There are almost books written on the subject of using decoupling capacitors alone and tomes of information on innocuous subjects that to the untrained eye look simple, like power and ground planes. In a high speed digital design, a PCB trace isn't simply a wire linking A to B to the engineer who is deciding where it should go on the PCB.
It takes a significant amount of time and effort, and a significant amount of knowledge to put together even relatively straightforward high speed digital designs. With the iPhone you also have to cram it into a very constrained amount of space, too. You can bet the engineers who laid out the PCB spent a great deal of time making sure that not only it would fit, but the resultant device would work and pass FCC testing.
Touch computing may be mainstream for handheld devices, but it will be a long time before the mouse is replaced on a desktop PC. What these prognosticators always seem to forget about PC displays is the display is vertical and in front of you. It gets tiring if you have to hold your arm up to touch stuff on the screen all day. Your arm does at least get to rest on the table if you're using a mouse or trackpad.
For a PC, the prognosticators also seem to forget that the mouse is good enough, and it's tremendously difficult to replace "good enough" in three years. Touch interfaces on a desktop system don't offer any benefits over a mouse (unlike on handheld devices, where a touch interface is obviously very very much better than any other kind of pointing device). For laptops, again, the vertical screen problem and arm-tiredness/screen smudging issues persist, and people find trackpads good enough with a touch screen not really offering any worthwhile benefit on a full size laptop.
On my last project for an extremely large customer (with an equally huge project), the good line managers were our (the developers) advocates, they took the bullshit so we didn't have to, and determined the "big picture". They didn't manage who was doing which particular piece of code - that was down to the developers to organize themselves. Managing the development team was less about managing the people in it (the developers could organize themselves) but being an advocate for the team, and making sure that the people who knew how to do the stuff were fully involved in decisions affecting them. Developers were not merely involved but critical to things such as sizing parts of the project, so that unrealistic schedules were not set. The line manager's job was in this case often to tell upper management "this is why it's going to take this long" in terms they could understand, and persuade upper management not to cause a disaster by compressing schedules or adding more work.
It resulted in a productive development team which did not have to do unpaid overtime, and delivered a quality product to the customer - earning a very large sum of money for the company.
What's wrong with just a normal <A HREF=... for navigation? Even JavaScript would be better for navigation than a proprietary, Microsoft-only language. If you're using proprietary languages, you don't have a web site, you have an Internet Explorer site.
For loss of confidential data, you can already do full disc encryption using TrueCrypt (including the system disc).
I think it's the syllabus. Not necessarily just bad teachers, but otherwise good teachers having to teach to the exam, and the method in teaching a particular subject being wrong in the first place.
I have always considered myself bad at (human) languages. After something like 7 years of school French, I couldn't even reserve a hotel room in French, let alone carry out a very basic conversation. However, I discovered that wasn't really the case; it wasn't that I was bad at languages, it's just that languages are taught fundamentally in the wrong way in this country. It can't just be me considering how many British people are monolingual despite being compelled to do a language GCSE.
In May this year, I needed to learn some Spanish so I wouldn't feel like a complete Imperialist when visiting Palma, and did a beginner's course on the BBC website. The course was fun to do, and I found myself wanting to learn more, and since May, I have learned more Spanish in 7 months than I did French in 7 years of school French, supposedly when the mind is so much more able to pick up another language. Already, I can understand most stories on Spanish news websites like El Pais or BBC Mundo, and I've had meaningful conversations online with Spanish people. I've even managed to laugh at a joke in Spanish.
So I don't think British people are inherently bad at languages, I just think the method languages are taught is bad, and those who do learn another language at school do so in spite of their education, not because of it. Interestingly, on Radio 4's PM programme a couple of months back, there was a report on language learning in the UK, and how so many young people want to work in France or Spain for a while, but never get to grips with the language. They had some girl on there, who had been doing French for 5 years and got a good grade at GCSE, but when the reporter asked her to say something simple ("What did you do this morning?") in French, she couldn't.
In summary, I agree with your observation - the content is wrong, and often so is the teaching method.
If you live in the United States, you need a degree.
If you live in Europe, for the most part, you don't for an IT job, but you will for a software development job. "IT" is really a trade like plumbing. You don't need a degree to pull cat5 cables, fix broken printers, or write shell scripts to automate various repetitive jobs, just intelligence and experience.