Ironically, the main benefit of DTT in the UK (since the collapse of the subscription-based model and the transfer of DTT to the BBC) is that you can now receive a wider range of digital radio stations.
Maybe we should simply close down TV altogether and use the spectrum for something more interesting. It's not like there's a public outcry for QVC 2.
Having, at least a decade and a half ago, been peripherally involved with IPv6 at the spec stage, I'm not convinced it's really the answer to anything.
There are loads of things in the Internet that are "broken" in the sense that they were never designed for a global and hostile network environment and some things that are just broken (such as TCP).
IPv6 doesn't really fix any of the broken things, apart from the lack of address space. The only reason to have a larger address space is to construct a (potentially) universally connected network layer. However, without the ability to control network-layer traffic more effectively at subnetwork boundaries - and more easily-managed and stable routing - the universally-connected network isn't going to happen.
IPv6 is a solution to a problem of 20 years ago and really has no bearing on today's problems.
I guess the numbers have some interest, but I'd be far more interested in what they're doing with their web servers. On the assumption that serving flat HTML is a minority interest, what, more significantly, are they using for their application development? Perl? PHP? Java? C?
One of the main problems with IIS is that its single-process, multi-threaded operation makes it very vulnerable to threadlocks and memory leakage by various ancillary software components (database drivers, Active X stuff, etc). Debugging these problems is next-to-impossible, particularly for someone who's chosen to use IIS largely because of a familiarity with Visual Basic.
I would not *a priori* expect threading in Apache 2.0 to work any better than IIS if it's working with, say, PHP into which you can build a myriad of library functions many of which have a single-threaded heritage.
So, if users are moving to Apache in droves because they've found a reliable rapid development environment for multi-threaded web applications, then I'd be interested to know what (apart from Apache) was involved.
After all, Apache (like IIS) is fundamentally no more than a dispatcher for HTTP requests. It's producing the responses that causes the trouble!
Suppose I was buying a new computer on which I wished to run a wide variety of Windows applications. I could buy an OS-free PC, install Linux and then work out how to install Crossover Office in order to be able to run "most" of the functionality of a small number of applications. For only slightly more than the cost of a Crossover license and at considerably less trouble, I could buy an OEM edition of Windows and have access to a complete range of fully-supported Windows applications on my new PC and some reasonable level of commitment to keep them working.
Or, perhaps I buy a PC with Windows and an OEM version of Office pre-installed and save some time and money.
There are some niche applications for this product, but if you want to run a lot of Windows applications, running Windows seems like a more obvious way of doing it...
Someone has no doubt suggested this before, but I think we need a new word here. How about "blatent" - a statement of the bleeding obvious used as an anti-competitive weapon.
And now to rush that one down to the trademark office....
A great feature of the Commodore ROMs was that to save space some of the subroutines had two entry points that were implemented by jumping either to the start of an instruction or into the middle of the same instruction (which just happened to have the correct byte code to be a different instruction). The different entry point would affect the flags or registers in such a way as to provide two different outcomes without wasting space on tests and branches. Try doing that in C++!
I did manage to come up with a nice high-level assembler though which started life as a basic program and eventually became a machine code program that would assemble itself. Got to make sure you don't lose the interim versions, though, when you're bootstrapping your software like that....
Of course, they don't, or at least they don't adequately when drawn at a scale which is small compared with the pixel resolution of a raster display device.
Which is why TrueType has (proprietary) font hinting: fonts are just scalable vector graphics and they don't scale small and stay readable.
Don't put away your bitmap icon editor just yet...
The energy supply system here in the UK used to be owned by the government (post-war socialism) and was privatised by the Thatcher government partly as a matter of dogma and partly because it was seen as "over-engineered" and bloated.
It was replaced by a single grid company, local distribution companies and generator companies. Power was bought on a complex "pool" system in which generators bid to provide power in various timeslots and the lowest bids were accepted. This was gradually changed and now consumers choose which company they buy their power from and the "pool" has become considerably more complex as a result.
One of the effects of this is that a large number of power plants have been mothballed or dismantled as they are too inefficient to compete with newer gas-powered stations during the summer months and cannot consequently provide an economic return. Unfortunately, this means that there is a severe danger that there will be insufficient generating capacity in the winter: the National Grid has already issued a warning for 2003-4.
You can have cheap power or reliable power, but not both: to have the latter to need slack, or "bloat" in the system. If you want reliable power, you have to have some sort of central authority exercised either through ownership or through regulation. Either solution is likely to be resisted by politicians (who wants to be responsible for the power going out?) until there's a major calamity.
Minor stuff - repeated "number withheld" calls that get hung up on answer around the time of a complaint. The occasional timewasting letter from companies wanting to gather "additional information" so that they can justify their actions to the TPS (one from a company that refused to give their address over the phone but obviously had no trouble in tracking down mine!).
Well, that could partly becase the TPS list is widely ignored: true, the "reputable" (large and worried about the publicity) companies take note. However, smaller (and downright dodgy) companies take the view that the remote possibility of a fine for non-compliance is considerably cheaper than renting the list (which apparently costs several K per year). I still get around a call a month despite being on the list and occasionally get harassed by companies when I report them to the TPS...
IBM is the only computer company I can think of that has successfully managed technological change and survived. HP would have been the other one, but I'm no longer too convinced about that!
As IT companies grow they are faced with the inveitable fact that the majority of their customers have singificant investment in outmoded IT and a vested interest in keeping that investment going. They can't abandon their customers, but they also have to deal with massive shifts in technology and - let's face it - technological fashion.
It's very, very difficult to deal with a huge and increasing legacy of support for the "old" technology and simultaneously have the resources to innovate. Most companies come to a point at which they can no longer do either effectively and the next technology "paradigm shift" (mainframe to mini; mini to PC; Unix to Linux...) makes them irrelevant.
Slashdot Mirror
Ironically, the main benefit of DTT in the UK (since the collapse of the subscription-based model and the transfer of DTT to the BBC) is that you can now receive a wider range of digital radio stations.
Maybe we should simply close down TV altogether and use the spectrum for something more interesting. It's not like there's a public outcry for QVC 2.
Having, at least a decade and a half ago, been peripherally involved with IPv6 at the spec stage, I'm not convinced it's really the answer to anything.
There are loads of things in the Internet that are "broken" in the sense that they were never designed for a global and hostile network environment and some things that are just broken (such as TCP).
IPv6 doesn't really fix any of the broken things, apart from the lack of address space. The only reason to have a larger address space is to construct a (potentially) universally connected network layer. However, without the ability to control network-layer traffic more effectively at subnetwork boundaries - and more easily-managed and stable routing - the universally-connected network isn't going to happen.
IPv6 is a solution to a problem of 20 years ago and really has no bearing on today's problems.
I guess the numbers have some interest, but I'd be far more interested in what they're doing with their web servers. On the assumption that serving flat HTML is a minority interest, what, more significantly, are they using for their application development? Perl? PHP? Java? C?
One of the main problems with IIS is that its single-process, multi-threaded operation makes it very vulnerable to threadlocks and memory leakage by various ancillary software components (database drivers, Active X stuff, etc). Debugging these problems is next-to-impossible, particularly for someone who's chosen to use IIS largely because of a familiarity with Visual Basic.
I would not *a priori* expect threading in Apache 2.0 to work any better than IIS if it's working with, say, PHP into which you can build a myriad of library functions many of which have a single-threaded heritage.
So, if users are moving to Apache in droves because they've found a reliable rapid development environment for multi-threaded web applications, then I'd be interested to know what (apart from Apache) was involved.
After all, Apache (like IIS) is fundamentally no more than a dispatcher for HTTP requests. It's producing the responses that causes the trouble!
Suppose I was buying a new computer on which I wished to run a wide variety of Windows applications. I could buy an OS-free PC, install Linux and then work out how to install Crossover Office in order to be able to run "most" of the functionality of a small number of applications. For only slightly more than the cost of a Crossover license and at considerably less trouble, I could buy an OEM edition of Windows and have access to a complete range of fully-supported Windows applications on my new PC and some reasonable level of commitment to keep them working.
Or, perhaps I buy a PC with Windows and an OEM version of Office pre-installed and save some time and money.
There are some niche applications for this product, but if you want to run a lot of Windows applications, running Windows seems like a more obvious way of doing it...
Someone has no doubt suggested this before, but I think we need a new word here. How about "blatent" - a statement of the bleeding obvious used as an anti-competitive weapon.
And now to rush that one down to the trademark office....
Replacing all those buffer-overflow errors with errors caused by incorrect amounts of whitespace is sure to be the root of commercial recovery....
I did manage to come up with a nice high-level assembler though which started life as a basic program and eventually became a machine code program that would assemble itself. Got to make sure you don't lose the interim versions, though, when you're bootstrapping your software like that....
Which is why TrueType has (proprietary) font hinting: fonts are just scalable vector graphics and they don't scale small and stay readable.
Don't put away your bitmap icon editor just yet...
It was replaced by a single grid company, local distribution companies and generator companies. Power was bought on a complex "pool" system in which generators bid to provide power in various timeslots and the lowest bids were accepted. This was gradually changed and now consumers choose which company they buy their power from and the "pool" has become considerably more complex as a result.
One of the effects of this is that a large number of power plants have been mothballed or dismantled as they are too inefficient to compete with newer gas-powered stations during the summer months and cannot consequently provide an economic return. Unfortunately, this means that there is a severe danger that there will be insufficient generating capacity in the winter: the National Grid has already issued a warning for 2003-4.
You can have cheap power or reliable power, but not both: to have the latter to need slack, or "bloat" in the system. If you want reliable power, you have to have some sort of central authority exercised either through ownership or through regulation. Either solution is likely to be resisted by politicians (who wants to be responsible for the power going out?) until there's a major calamity.
Minor stuff - repeated "number withheld" calls that get hung up on answer around the time of a complaint. The occasional timewasting letter from companies wanting to gather "additional information" so that they can justify their actions to the TPS (one from a company that refused to give their address over the phone but obviously had no trouble in tracking down mine!).
Well, that could partly becase the TPS list is widely ignored: true, the "reputable" (large and worried about the publicity) companies take note. However, smaller (and downright dodgy) companies take the view that the remote possibility of a fine for non-compliance is considerably cheaper than renting the list (which apparently costs several K per year). I still get around a call a month despite being on the list and occasionally get harassed by companies when I report them to the TPS...
As IT companies grow they are faced with the inveitable fact that the majority of their customers have singificant investment in outmoded IT and a vested interest in keeping that investment going. They can't abandon their customers, but they also have to deal with massive shifts in technology and - let's face it - technological fashion.
It's very, very difficult to deal with a huge and increasing legacy of support for the "old" technology and simultaneously have the resources to innovate. Most companies come to a point at which they can no longer do either effectively and the next technology "paradigm shift" (mainframe to mini; mini to PC; Unix to Linux...) makes them irrelevant.
Why should Sun be any different?