Well, it's not "non standard" inasmuch as it's a documented part of the Windows API now.
Is it an incremental download protocol that can restart if it's interrupted? From Microsoft's web site:
Background Intelligent Transfer Service (BITS) asynchronously transfers files in the foreground or background, throttles the transfers to preserve the responsiveness of other network applications, and automatically resumes file transfers after network disconnects and machine restarts.
Can the protocol work through firewalls? Can the protocol work through proxies? Can the protocol work through VPNs?
It's just HTTP or HTTPS as far as the network is concerned. So the answer is, if you can browse, you can use BITS. Note that the this isn't just for Windows updates; the API can be used by any application that needs to move a lot of data around and would prefer not to interfere with normal operation. The API works like the print queue; you can change the priority or order of jobs, authenticate for particular jobs, etc.
Does the protocol try to work *around* VPNs?
No, the underlying protocol is just HTTP/HTTPS.
What's the MTU size of BITS packets? Bandwidth percentage limitations are fine, but once a packet gets its turn on the wire, it's a latency hit, especially at low bandwidth.
Now you're trying to be too smart, and you just look silly.
Is there some vague possibility of security?
BITS is no more or less secure than anything that just downloads over HTTP, like say wget. How secure it is wholly depends on what you do with what you've downloaded.
It's because of the computer elitist group (Hi Slashdot!) that computers "scare" people. They aren't interested, and would rather just have someone who is interested fix their problems. There is nothing wrong with that, and it doesn't make them stupid.
Let me explain the psychology at work here. Back in the day, computers were housed in dedicated buildings. Ordinary users never got near them, if they wanted anything done, they had to ask the techno-elite, and the techno-elite would, if they felt like, do it, like high priests making sacrifices in the inner temple on behalf of the masses outside.
This made the "high priests" feel special. They had all the power.
Along comes Bill Gates, with his idea of "a computer on every desktop". That's actually Microsoft's "mission statement", you know. Suddenly, the masses don't need the "high priests" any more. They want to do calculations, they've got spreadsheets, they want to store records, they've got databases. Some of them bought compilers and discovered that programming isn't a secret ritual after all, and they began to produce software for the ordinary people to use. Now, the "high priests" are relegated to the role of "support". No more are they help in awe, when once a user would have to beg for some computer time, now those users just expect the "priests" to refill the printer with paper.
That's why Slashbots hate Bill Gates and Microsoft with such fervour.
Like, if you put Deep Blue against Gnuchess... or even Sargon II on the Vic-20... would Deep Blue win *every single time* in a trouncing defeat?
It would depend on what conditions were imposed on timings. If you assume that each machine is allowed 5 minutes to think about each move, and that each game has (say) approximately 100 moves, then between the first and second moves, Deep Blue could model every single game that the VIC20 could possibly play and be ready for it.
But if you set the game up so each machine could look no more than 3 moves into the future, then it comes down to the skill of the algorithm designers. It could well be that the Sargon II programmer was smarter because he had such limited resources and the IBM programmers grew sloppy because they assumed that they would always have enough processor time to look further into the future than their opponents.
What will be the next challenge? Where is there a game that requires the uniqueness of human thought over the pure power of computer calculations?
You're missing the point of games, which is what chess is, a means of entertainment for the players and in some cases spectators. We have machines that can lift many orders of magnitude more weight than any human, yet weightlifting is still a sport. We have machines that travel at 20,000 mph, yet running at 20 mph is still a sport. You can use a 10-megapixel camera to take pictures, yet painting is still an art. And the finest technology in the world can't turn a mediocre musician into a good one.
Superseding the individual capabilities of humans is the whole point of machines. Humans, for the forseeable future, are the only entities that can decide what to do, and why, and when, bringing capabilities to bear on a problem. If you're looking for a challenge, try implementing "hard" AI:-)
Thing is, if you've chosen to study computer science to develop complex algorithm to solve a real-world problem, are you/have you also studied the art of developing software so that your implementation isn't hindered by common development faults. This means that you'll gain (or have) a deep understanding of the software development lifecycle and all the possible failure points in that lifecycle.
I've never met a recent graduate who has done any of that. When you hire a CS grad you have to spend 6 months to a year forcing them unlearn all that crap their professors, who haven't worked in industry for decades if ever, have taught them.
A fully developed software engineer has a much better chance of securing employment in the current climate here and around the world.
If your intent is to work in industry, 3 years real experience wipes the floor with 5 years of grad school.
It's conceivable, sure, but you're a lot more likely to get turned down for a job for a lack of education than too much education.
I'm not sure that's true. Certainly, my own experience of interviewing candidates is I'd rather hire a candidate with less education and more experience than one with more education and less experience. That comes from hiring people and seeing how they perform in the "real world". A PhD comes across as being too "theoretical", interested in abstracts and research, and not in day-to-day programming which might be just grinding out database code or fine-tuning GUIs. A Master's will stand you in good stead for "high level" job like a System Architect. But PhDs are too highly qualified for an entry-level coder (and are reluctant to take a "junior" position anyway), and not experienced enough for a senior position, so they're stuck in limbo.
Only do a PhD if you have a genuine interest in the research you want to do - for example, if you're deeply interested in AI anyway, a PhD will be a rewarding experience. But it is a big mistake to do a PhD purely as a way into the job market.
And all this while being massively underfunded by the Tories in an attempt to make it look bad because they hated state-run institutions for ideological reasons.
Bliar (not a typo) could blame "Tory underinvestment" for the first year or two, but he's had 6 years, massively increased tax and spending, yet he can't even maintain services at the same level as the Tories, let alone improve them! That's because his own ideology can't differentiate between effort and results.
Not to mention that the power needs in california were caused by witholding of power by Enron, not by deregulation.
Not quite. What happened was that the CA government deregulated the market between distributor and generator, but not between distributor and consumer. Therefore, as the price at which the distributor bought power fluctuated, it could not vary the price it charged. This made power artificially cheap for the consumer, and being CA, the locals use a lot of power to run aircon, swimming pools, etc. Demand outstripped supply. In a true market, the consumer price would have gone up and the locals would have thought of ways to be less wasteful of power. That didn't happen until the half-regulated market collapsed.
Why did the price of the generators go up? Simply because CA doesn't allow power stations to be built - there was a built-in shortage in the system. As soon as that was reached, the distributor had to buy from elsewhere on the open market.
The solution would have been for the CA government to a) permit private industry to construct sufficient capacity and b) fully deregulate. But they didn't; I suspect because they simply wanted to give deregulation per se a bad name.
The solution to every domestic energy issue must be to drill oil wells in Alaska. The problem to every foreign policy problem must be to invade a country in the gulf with large oil reserves. The problem to every economic problem must be to give stupendous tax cuts where at least 80% but hopefully as much as is possible goes to the richest of the rich, and in particular rich Texas oil-men.
It was all going so well before you went all rabid. Let me just point out:
The solution to every domestic energy issue is not to put up windmills
The solution to every foreign policy problem is not to blow up their pharmaceutical industry with cruise missiles, as Clinton did in Sudan
The solution to every economic problem is not to give huge raises and generous pensions to government workers thus bankrupting your state, as Davis did in California.
Incidentally, of course tax cuts are going to benefit those who pay the most tax. If people who don't pay Federal income tax anyway get the money, it's not a cut but a grant. No wonder lefties can't run economies, they even struggle with simple language!
The same for the british rail : it worked perfectly. deregulation came in and it went down in flames, late trains, dirty wagons, and dead peoples in accidents
I'm guessing you're too young to have ever actually used BR, or have clear memories of doing so. It sucked. Not quite as high-profile as modern failures, but then nothing was because the media lacked the ability it has now to actually be there as things went wrong. But it was late, and filthy, and unsafe.
The problem faced by the old BR was that it was forced to provide services where they weren't economically viable - running trains to tiny, out-of-the way villages where almost no-one used trains anyway, for example. This starved it of resources for the core infrastrucure - the busy intercity and commuter routes. And it also faced the same problem that modern train companies face, unions who refuse to link pay to performance.
If you noticed, there isn't much to the site... a few servlets, jsps, and some queries. 6 lines of code maybe...
The queries required to find how person A is linked to person B via C, D and E are quite processor-intensive, if there are hundreds of people involved. That'll be where they're spending the money on new hardware. If course, they may be using a custom DB and not an RDBMS, in which case you'd be right.
Who cares about "fake" members? Friendster, probably. And some journalist who can't find a better story.
I don't know what Friendster's infrastructure is, but on a relational database doing the sorts of queries they do (finding multiple paths through networks in near real-time) would be pretty heavy - even with a professional RDBMS like Oracle. Most real people have a few dozen friends linked, but the fakers often have hundreds. If serving pages belonging to fakers means Friendster needs more hardware, then it is costing them real money.
In the city it is significantly higher, but not enough to call the average high or rich.
You aren't comparing like with like. A salary of $10,000 makes you very well paid in India, and it's still a fraction of what would be a merely average salary in Silicon Valley.
Comparing dollars in India with dollars in the US is pointless because a) Indians aren't spending USD but INR and b) they aren't paying US prices but local prices. Anyone who quotes dollars like that is either ignorant of economics, or is trying to make an (unfounded) political point about exploitation.
In other news, it was discovered that everyone looks like an idiot when they require the services of a domain expert. What's next, neurosurgeons complaining that patients don't know as much as them? Of course end users don't know much about tech - that's what they're paying support workers for! Just like drivers pay auto mechanics, and anyone who has a bathroom pays a plumber.
Just because someone doesn't happen to have some specialized piece of knowledge you have, that doesn't make them "not so bright". I know plenty of PhDs who are extremely competent in their fields, which aren't computing, who need to call helpdesks from time to time. You see, and this will sound harsh to a Slashbot, most people have better things to do than learn the minutae of their PCs.
Most of the people who call for help don't even know what operating system they're using -- even though they've spent their money buying the machine.
How many drivers know what OS runs their engine control computer? Even tho' they spent their money buying the machine. You see, techies are into operating systems are care a lot about them. End users care about getting their jobs done, and the computer is just a tool. One version of Windows looks a lot like another - can you tell the difference between '95, '98 and ME with just a glance? You can? Can you tell the difference between Red Hat, Debian and SuSE at a glance? You think so? I didn't tell you they're all in console mode at a $ prompt.
Tech support needs to stop thinking of end users as the enemy and start thinking of them as what they really are, its bread and butter.
Related point: one of the sad things is that very few Unix applications can gracefully handle errors such as out of disk and I/O errors. A particularly popular response is to blindly continue execution after a failed system call and eventually segfault. In the case of X11 applications, this means the program simply disappears.
This is because of the way Unix handles an interrupt while in a system call. In user mode, it's just a matter of saving the registers and program counter, then you can start where you left off. In kernel mode, you might be in the middle of an operation that changes state outside of the kernel - for example, on the filesystem. It's much harder to preserve state and come back to it... so it doesn't. It just throws it away, and goes to do whatever the interrupt wanted.
That's why every system call, for example file I/O, has a return value. You, as the programmer, are supposed to check that return value and repeat the command if you get a failure. So to ensure correct behavior, you would have to wrap every single system call you make in:
int retval = FAIL; int tries = 0; while ((retval == FAIL) && (tries < MAXRETRIES)) {
retval =...;
tries ++; } if (tries == MAXRETRIES) { ...; }
Unsurprisingly, most programmers don't do that. And that's why Unix programs handle I/O errors badly - the OS simply makes it too difficult.
Have any national governments taken measures to subsidize open source projects?
Yes, who do you think paid for the earliest work on Linux? The Finnish government, of course! Like in many European countries, the taxpayer gives grants to students, and that's most likely what Linus lived on.
For most Americans, even those motivated to learn another language, there's little practical ways to get and stay proficient (ie, carry on colloquial conversations, read/write), since almost nobody speaks the 'other' languages, excluding Spanish.
I'm an Englishman and I concur with this. I've worked in Holland and France, with teams of people from several countries, and the default language at all times was English. That's what a Dutchman and a Frenchie will use when they need to speak to each other, or a Spaniard and a Pole.
English is the universal language, and like it or not, how well you can communicate internationally is directly related to how well you can speak it.
SIP users include AT&T, Avaya, Texas Instruments, Worldcom, Alcatel, Nortel and MCI. You know, the biggest telcos and telco equipment manufacturers in the world. SIP is hardly "doomed".
Re:It was only a matter of time...
on
HavenCo In Trouble?
·
· Score: 4, Insightful
The only problem is that Sealand's Law is whatever their "Crown Prince" says it is.
Sealand's law is whatever the British Government will let them get away with. A frigate's detachment of Marines could re-occupy the platform in minutes without breaking a sweat. The Crown Prince is tolerated because Britain has a tradition of tolerating eccentrics so long as they don't harm anyone. If Sealand were to declare that it was willing to break British laws wholesale, bearing in mind that it is strategically located, it would rapidly - and perhaps physically - cease to exist.
No: most software projects are far more complex than one of those, and have the additional disadvantage of never having been done before.
I don't believe that is the case. Nearly all applications written are about retrieving data from some form of database, displaying it in some way, optionally allowing the user to edit it, optionally performing some calculation and writing it back to some form of database. Yet still, software projects frequently end in disaster.
Also, look at the people involved. Professional engineers with decades of experience work on airliners and skyscrapers, whereas any kid who hasn't even completed college can write large amounts of code. That indicates that writing code is considerably easier than real engineering.
Software will only get better when we do away with processes. Software is complex, it requires us to think -- not follow procedures.
I'm guessing you have little or no real-world experience. Consider artifacts from other technical disciplines: airliners, nuclear power stations, skyscrapers, etc. Very few software projects in the world are even within an order of magnitude of the complexity of one of those. Yet all those things require a great deal of process to get right, and the proof, as they say, is in the pudding - they are all very reliable. Whereas software, much simpler, is renowned for being unreliable. And you think software needs less process??
Slashdot Mirror
Degree != real-world experience. I've got both, as I'm sure do many on slashdot. The two are symbiotic, not the same.
A good MBA programme won't take you without experience. Typical students have worked for 3 to 8 years before applying to B-school.
Well, it's not "non standard" inasmuch as it's a documented part of the Windows API now.
Is it an incremental download protocol that can restart if it's interrupted?
From Microsoft's web site:
Can the protocol work through firewalls?
Can the protocol work through proxies?
Can the protocol work through VPNs?
It's just HTTP or HTTPS as far as the network is concerned. So the answer is, if you can browse, you can use BITS. Note that the this isn't just for Windows updates; the API can be used by any application that needs to move a lot of data around and would prefer not to interfere with normal operation. The API works like the print queue; you can change the priority or order of jobs, authenticate for particular jobs, etc.
Does the protocol try to work *around* VPNs?
No, the underlying protocol is just HTTP/HTTPS.
What's the MTU size of BITS packets? Bandwidth percentage limitations are fine, but once a packet gets its turn on the wire, it's a latency hit, especially at low bandwidth.
Now you're trying to be too smart, and you just look silly.
Is there some vague possibility of security?
BITS is no more or less secure than anything that just downloads over HTTP, like say wget. How secure it is wholly depends on what you do with what you've downloaded.
It's because of the computer elitist group (Hi Slashdot!) that computers "scare" people. They aren't interested, and would rather just have someone who is interested fix their problems. There is nothing wrong with that, and it doesn't make them stupid.
Let me explain the psychology at work here. Back in the day, computers were housed in dedicated buildings. Ordinary users never got near them, if they wanted anything done, they had to ask the techno-elite, and the techno-elite would, if they felt like, do it, like high priests making sacrifices in the inner temple on behalf of the masses outside.
This made the "high priests" feel special. They had all the power.
Along comes Bill Gates, with his idea of "a computer on every desktop". That's actually Microsoft's "mission statement", you know. Suddenly, the masses don't need the "high priests" any more. They want to do calculations, they've got spreadsheets, they want to store records, they've got databases. Some of them bought compilers and discovered that programming isn't a secret ritual after all, and they began to produce software for the ordinary people to use. Now, the "high priests" are relegated to the role of "support". No more are they help in awe, when once a user would have to beg for some computer time, now those users just expect the "priests" to refill the printer with paper.
That's why Slashbots hate Bill Gates and Microsoft with such fervour.
Like, if you put Deep Blue against Gnuchess... or even Sargon II on the Vic-20... would Deep Blue win *every single time* in a trouncing defeat?
It would depend on what conditions were imposed on timings. If you assume that each machine is allowed 5 minutes to think about each move, and that each game has (say) approximately 100 moves, then between the first and second moves, Deep Blue could model every single game that the VIC20 could possibly play and be ready for it.
But if you set the game up so each machine could look no more than 3 moves into the future, then it comes down to the skill of the algorithm designers. It could well be that the Sargon II programmer was smarter because he had such limited resources and the IBM programmers grew sloppy because they assumed that they would always have enough processor time to look further into the future than their opponents.
What will be the next challenge? Where is there a game that requires the uniqueness of human thought over the pure power of computer calculations?
:-)
You're missing the point of games, which is what chess is, a means of entertainment for the players and in some cases spectators. We have machines that can lift many orders of magnitude more weight than any human, yet weightlifting is still a sport. We have machines that travel at 20,000 mph, yet running at 20 mph is still a sport. You can use a 10-megapixel camera to take pictures, yet painting is still an art. And the finest technology in the world can't turn a mediocre musician into a good one.
Superseding the individual capabilities of humans is the whole point of machines. Humans, for the forseeable future, are the only entities that can decide what to do, and why, and when, bringing capabilities to bear on a problem. If you're looking for a challenge, try implementing "hard" AI
Thing is, if you've chosen to study computer science to develop complex algorithm to solve a real-world problem, are you/have you also studied the art of developing software so that your implementation isn't hindered by common development faults. This means that you'll gain (or have) a deep understanding of the software development lifecycle and all the possible failure points in that lifecycle.
I've never met a recent graduate who has done any of that. When you hire a CS grad you have to spend 6 months to a year forcing them unlearn all that crap their professors, who haven't worked in industry for decades if ever, have taught them.
A fully developed software engineer has a much better chance of securing employment in the current climate here and around the world.
If your intent is to work in industry, 3 years real experience wipes the floor with 5 years of grad school.
. I think someone working in the field and getting a Master's degree is pretty valuable itself. (Because I'm doing that! :) )
:-)
I hope so - I have one too
It's conceivable, sure, but you're a lot more likely to get turned down for a job for a lack of education than too much education.
I'm not sure that's true. Certainly, my own experience of interviewing candidates is I'd rather hire a candidate with less education and more experience than one with more education and less experience. That comes from hiring people and seeing how they perform in the "real world". A PhD comes across as being too "theoretical", interested in abstracts and research, and not in day-to-day programming which might be just grinding out database code or fine-tuning GUIs. A Master's will stand you in good stead for "high level" job like a System Architect. But PhDs are too highly qualified for an entry-level coder (and are reluctant to take a "junior" position anyway), and not experienced enough for a senior position, so they're stuck in limbo.
Only do a PhD if you have a genuine interest in the research you want to do - for example, if you're deeply interested in AI anyway, a PhD will be a rewarding experience. But it is a big mistake to do a PhD purely as a way into the job market.
There's really no point explaining that it's because of MS that they have these problems.
Rubbish. I expect you blame Ford for the existance of car thieves? Damn Ford, they should have used brick-proof glass in the windows!
And it's not as if Linux has never been r00ted via sendmail or BIND, is it? MS Blaster is the same, it just propagates over DCOM.
And all this while being massively underfunded by the Tories in an attempt to make it look bad because they hated state-run institutions for ideological reasons.
Bliar (not a typo) could blame "Tory underinvestment" for the first year or two, but he's had 6 years, massively increased tax and spending, yet he can't even maintain services at the same level as the Tories, let alone improve them! That's because his own ideology can't differentiate between effort and results.
Not to mention that the power needs in california were caused by witholding of power by Enron, not by deregulation.
Not quite. What happened was that the CA government deregulated the market between distributor and generator, but not between distributor and consumer. Therefore, as the price at which the distributor bought power fluctuated, it could not vary the price it charged. This made power artificially cheap for the consumer, and being CA, the locals use a lot of power to run aircon, swimming pools, etc. Demand outstripped supply. In a true market, the consumer price would have gone up and the locals would have thought of ways to be less wasteful of power. That didn't happen until the half-regulated market collapsed.
Why did the price of the generators go up? Simply because CA doesn't allow power stations to be built - there was a built-in shortage in the system. As soon as that was reached, the distributor had to buy from elsewhere on the open market.
The solution would have been for the CA government to a) permit private industry to construct sufficient capacity and b) fully deregulate. But they didn't; I suspect because they simply wanted to give deregulation per se a bad name.
It was all going so well before you went all rabid. Let me just point out:
Incidentally, of course tax cuts are going to benefit those who pay the most tax. If people who don't pay Federal income tax anyway get the money, it's not a cut but a grant. No wonder lefties can't run economies, they even struggle with simple language!
The same for the british rail : it worked perfectly. deregulation came in and it went down in flames, late trains, dirty wagons, and dead peoples in accidents
I'm guessing you're too young to have ever actually used BR, or have clear memories of doing so. It sucked. Not quite as high-profile as modern failures, but then nothing was because the media lacked the ability it has now to actually be there as things went wrong. But it was late, and filthy, and unsafe.
The problem faced by the old BR was that it was forced to provide services where they weren't economically viable - running trains to tiny, out-of-the way villages where almost no-one used trains anyway, for example. This starved it of resources for the core infrastrucure - the busy intercity and commuter routes. And it also faced the same problem that modern train companies face, unions who refuse to link pay to performance.
If you noticed, there isn't much to the site... a few servlets, jsps, and some queries. 6 lines of code maybe...
The queries required to find how person A is linked to person B via C, D and E are quite processor-intensive, if there are hundreds of people involved. That'll be where they're spending the money on new hardware. If course, they may be using a custom DB and not an RDBMS, in which case you'd be right.
Who cares about "fake" members? Friendster, probably. And some journalist who can't find a better story.
I don't know what Friendster's infrastructure is, but on a relational database doing the sorts of queries they do (finding multiple paths through networks in near real-time) would be pretty heavy - even with a professional RDBMS like Oracle. Most real people have a few dozen friends linked, but the fakers often have hundreds. If serving pages belonging to fakers means Friendster needs more hardware, then it is costing them real money.
In the city it is significantly higher, but not enough to call the average high or rich.
You aren't comparing like with like. A salary of $10,000 makes you very well paid in India, and it's still a fraction of what would be a merely average salary in Silicon Valley.
Comparing dollars in India with dollars in the US is pointless because a) Indians aren't spending USD but INR and b) they aren't paying US prices but local prices. Anyone who quotes dollars like that is either ignorant of economics, or is trying to make an (unfounded) political point about exploitation.
As for linux at the CLI, a simple "cat /etc/issue" should reveal what distro the person is running
:-P
I did say, at a glance
In other news, it was discovered that everyone looks like an idiot when they require the services of a domain expert. What's next, neurosurgeons complaining that patients don't know as much as them? Of course end users don't know much about tech - that's what they're paying support workers for! Just like drivers pay auto mechanics, and anyone who has a bathroom pays a plumber.
Just because someone doesn't happen to have some specialized piece of knowledge you have, that doesn't make them "not so bright". I know plenty of PhDs who are extremely competent in their fields, which aren't computing, who need to call helpdesks from time to time. You see, and this will sound harsh to a Slashbot, most people have better things to do than learn the minutae of their PCs.
Most of the people who call for help don't even know what operating system they're using -- even though they've spent their money buying the machine.
How many drivers know what OS runs their engine control computer? Even tho' they spent their money buying the machine. You see, techies are into operating systems are care a lot about them. End users care about getting their jobs done, and the computer is just a tool. One version of Windows looks a lot like another - can you tell the difference between '95, '98 and ME with just a glance? You can? Can you tell the difference between Red Hat, Debian and SuSE at a glance? You think so? I didn't tell you they're all in console mode at a $ prompt.
Tech support needs to stop thinking of end users as the enemy and start thinking of them as what they really are, its bread and butter.
This is because of the way Unix handles an interrupt while in a system call. In user mode, it's just a matter of saving the registers and program counter, then you can start where you left off. In kernel mode, you might be in the middle of an operation that changes state outside of the kernel - for example, on the filesystem. It's much harder to preserve state and come back to it... so it doesn't. It just throws it away, and goes to do whatever the interrupt wanted.
That's why every system call, for example file I/O, has a return value. You, as the programmer, are supposed to check that return value and repeat the command if you get a failure. So to ensure correct behavior, you would have to wrap every single system call you make in:Unsurprisingly, most programmers don't do that. And that's why Unix programs handle I/O errors badly - the OS simply makes it too difficult.
Have any national governments taken measures to subsidize open source projects?
Yes, who do you think paid for the earliest work on Linux? The Finnish government, of course! Like in many European countries, the taxpayer gives grants to students, and that's most likely what Linus lived on.
For most Americans, even those motivated to learn another language, there's little practical ways to get and stay proficient (ie, carry on colloquial conversations, read/write), since almost nobody speaks the 'other' languages, excluding Spanish.
I'm an Englishman and I concur with this. I've worked in Holland and France, with teams of people from several countries, and the default language at all times was English. That's what a Dutchman and a Frenchie will use when they need to speak to each other, or a Spaniard and a Pole.
English is the universal language, and like it or not, how well you can communicate internationally is directly related to how well you can speak it.
...then it's doomed already.
SIP users include AT&T, Avaya, Texas Instruments, Worldcom, Alcatel, Nortel and MCI. You know, the biggest telcos and telco equipment manufacturers in the world. SIP is hardly "doomed".
The only problem is that Sealand's Law is whatever their "Crown Prince" says it is.
Sealand's law is whatever the British Government will let them get away with. A frigate's detachment of Marines could re-occupy the platform in minutes without breaking a sweat. The Crown Prince is tolerated because Britain has a tradition of tolerating eccentrics so long as they don't harm anyone. If Sealand were to declare that it was willing to break British laws wholesale, bearing in mind that it is strategically located, it would rapidly - and perhaps physically - cease to exist.
No: most software projects are far more complex than one of those, and have the additional disadvantage of never having been done before.
I don't believe that is the case. Nearly all applications written are about retrieving data from some form of database, displaying it in some way, optionally allowing the user to edit it, optionally performing some calculation and writing it back to some form of database. Yet still, software projects frequently end in disaster.
Also, look at the people involved. Professional engineers with decades of experience work on airliners and skyscrapers, whereas any kid who hasn't even completed college can write large amounts of code. That indicates that writing code is considerably easier than real engineering.
Software will only get better when we do away with processes. Software is complex, it requires us to think -- not follow procedures.
I'm guessing you have little or no real-world experience. Consider artifacts from other technical disciplines: airliners, nuclear power stations, skyscrapers, etc. Very few software projects in the world are even within an order of magnitude of the complexity of one of those. Yet all those things require a great deal of process to get right, and the proof, as they say, is in the pudding - they are all very reliable. Whereas software, much simpler, is renowned for being unreliable. And you think software needs less process??