Actually lots of these Slashdot readers have Videolan Client, mplayer and other multicast tools installed already. So its you that's "just making shit up". Some of them may even have multicast conferencing and multicast file distribution software too. It's just that all this stuff doesn't work on today's Internet for most people because their ISPs don't want to deploy it.
Right now, if the ISPs actually made their end of the system work properly, you could have 1000 people listening to a 128kbit Ogg Vorbis stream from your home ADSL, each of them using 128kbits at their end, and a total of 128kbits at your end, no huge bills for fat pipes, everyone walks away happy, except the ISPs. From the point of view of an ISP it's far better to have some poor clown paying $$$$ for the ability to send 128Mbit/s in order to connect 1000 people at 128 kbit/s unicast. Wasteful, but it all generates more revenue, and revenue feeds ISPs.
Many manufacturers already operate a "no dead pixels" policy, BUT you need to read the small print to discover that they don't regard a pixel in which one of the three color components is faulty as "dead". Only one of the subpixels is dead.
So if you buy one of their LCDs and it has a red subpixel that has failed, leaving a permanently red-tinted pixel, that's not considered a serious enough fault for a DOA replacement.
Faults which cause a whole pixel to die are relatively rare and certainly more obvious, so that's why they replace LCDs with this problem. You're still going to pay a premium if you want to buy a batch of LCDs and return all the ones with display faults. At least for a while yet. If you (as a dealer or reseller) want to buy such zero defect screens, you need ISO class 1, instead of the consumer class 2.
If you're a consumer, buying a single LCD, and you're having sleepless nights over dead (sub)pixels, go to a real world store, and insist on either opening & testing the LCD or buying a display model so that you can see exactly what you're getting. You will undoubtedly pay more than you would buying online, but you'll be certain to avoid dead (sub)pixels this way. Display faults are practically always a manufacturing problem, not something that develops over time.
Paying for megabits sent makes no sense. The only reason they charge you per kilojoule for the electric bill is because it costs money to make electricity. Do you think packets cost money? Can your five year old bankrupt you by running too many packets through::1 (127.0.0.1 to the old-schoolers) ?
Access charges are how you charge for networks. Sometimes idiots let other people charge them use charges for the network (eg. consumer POTS service long distance or international charges) but that's basically a license to print money for whoever does the charging.
[ Consider for a moment a five minute period in two parallel universes. In one you call Cape Town from New York over POTS for 5 minutes. In the other you just sit there meditating. Why did your POTS company get $2 or so of income in one universe but not in the other? What were their additional expenses incurred ? ]
What you get to charge for access depends on the quality of service, which means typical and guaranteed bitrate, availability, service limitations (e.g. no outbound SMTP packets) and so on. You contract with other access providers to exchange packets in order to increase the network value for everyone (routing policy).
Re:Down with BBCi - keep Ceefax!
on
Ceefax Turns 30
·
· Score: 1
There are three separate problems here, all of which can be fixed, and the result could be markedly better than Ceefax, although I grant that it isn't better yet.
1. Content. The digital subtitles are good, the rest of the content is weak. This would be fixed by spending more on content provision, but the BBC is reluctant to do so while digital takeup is still relatively slow.
2. Bandwidth. With the very restricted piece of radio spectrum assigned for digital TV it's hard to justify using more for rarely viewed digital interactive content, when it could be used to upgrade picture quality for millions of viewers. The easy way to fix this would be to kill Channel 5's analog slots and re-assign the frequencies, freeing up more space for BBCi and other low priority services. If C5 was appropriately compensated even their shareholders might support this idea.
3. Low quality devices. Have you used a non-fasttext TV to view Ceefax lately? Not so great is it. Right now the relatively high cost & low demand means that you can't buy a freeview or cable box that delivers the intended experience. They don't keep enough data (especially images) in cache, and they crash too easily. Time will heal this one, just like today's web browsers are a lot better than those from a decade ago. For a while you'll be able to pay more for a better experience, and then, like fasttext it will just become a standard feature on a chip somewhere.
An out-of-box Mac, or a recent Linux distro (e.g. Fedora Core 2) will automatically use IPv6 for most things if placed on an IPv6 connected network.
If you have IPv4 Internet on an FC2 machine, five minutes of work provides IPv6 to that machine and to every computer on the same LAN. If those computers run recent Linux or OS X they too will then be fully IPv6 enabled with no further work. Improved performance is enabled by the usual route: complain to your ISP about the lousy performance or switch to a better one.
On Windows you need to type in a command (I know, most people can't do that) or wait for Longhorn. Yet again Microsoft holds up the traffic.
Of course because it Just Works(TM) there aren't a lot of people shouting about this in the streets.
Intel's trick only works if there is no alternative supplier. Only one company made 486DXs, there are dozens just in my area competing to provide Internet service over DSL. If one says "We charge extra for $arbitrarything" then the others seize on it, and advertise their Free ArbitraryThing, especially if $arbitrarything is a fixed cost or very low per-user cost item.
Remember that every SINGLE IPv4 address is automatically granted & routed the associated/48 of IPv6 space under 6to4. So if your ISP takes away your one IPv4 routable address in exchange for even a/64 you're being ripped off to the tune of 2^16 already in address space terms.
No, the age of charging $5 per month for an IP address is over, whether or not IPv6 takes off in the next few years. In 1995 most people didn't know that they wanted the Internet. In 2000 most of them couldn't see why they'd need it in more than one room. Now some of the same people can't stand to be without it while on a train or in a pub. You're not going to convince them that a single address is enough.
And what about the people who needed human blood products in order to stay alive, who became infected and will eventually die because there is not only no vaccine, but no effective cure?
Still, so long as there's SOME way to avoid being infected there's no need for a vaccine right? Even the common cold is vanquished by your amazing discovery, simply never breathe the same air as other people without reprocessing.
"Look at the math - if the world population stabilizes at 9 billion, those resources relinquished by the West will simply be consumed by the growing third-world population."
So what?
A lot of crazy people have been running around saying that we're consuming too much of everything, as a species. They've been saying it for about a century now, and it isn't getting any more true. They point at charts (which show hilarious things like an assumption of no net energy input to the planet, ah hahahah) and they hold committee meetings and occasionally they write miserable books. They anthropomorphise an entire planet (Hello Gaia theory) in order to support their demand for arbitrary lifestyle changes which would make them feel comfortable.
I'm not interested in preventing hippies from drinking rainwater, living in cardboard houses, recycling plastic bottles or whatever else they want to do. I am interested in preserving my rather comfortable and wholely sustainable lifestyle despite the hippies, and thus I don't like to see them press their agenda as the "only sustainable way forward" regardless of whether it's printed on recycled paper. This planet is quite capable of feeding 10 billion people take-out Indian food in aluminium containers for longer than all recorded history, and even letting them all drive home with it, and eat it in air conditioned comfort.
That makes no sense. If you would normally receive a packet (e.g. because you provide web service, or have an IM port open or whatever) then the NAT router will rewrite the packets so that you still receive the trojan.
OTOH if you wouldn't normally receive something (e.g. it's an HTTP attack and you don't run a web server) then the NAT makes no difference, you still won't receive it. Big deal.
NATs are not magical protective charms. They're just a desperate hack to get around running out of IP addresses. If you want a firewall, install a firewall, not a NAT.
You can write testcases, but this is equivalent to the task undertaken by your attackers, with the unfortunate distinction that you must find ALL the bugs, and they only need ONE that you missed.
I've written testcases for N:M ratio software where many different applications make use of many different interoperable components (e.g. SANE's frontends and backends, or LADSPA's plugins and hosts). In this case it's important to preserve interop by detecting components that expect or provide something not agreed by the interface specification.
It was quite common to write SANE scanner backends which only handled large, even-sized buffers properly. With most frontends (typically requesting 64 kbytes at a time) this works fine, and so the authors had considered their code to be of good quality. But network users (who request Ethernet packet sized buffers) had a lot of problems. These days the SANE testbed application includes a test routine to check for this and report it as a bug in the backend. Unlike the network frontend (which not everyone can really test) this test is built as part of the SDK, so no driver author has an excuse for not using it to QA their code.
We never read about blackhats or greyhats who spent six months attacking a product with nothing to show for it, just like we never read about terrorists who survey a target and call off the operation because it's too well guarded. Only our failures get headlines:(
We like to call it "The United Kingdom" (which it isn't) or "Great Britain" (which is arguable) but it's certainly not bizarro world.
Remember, for a small once per year tax all UK residents have ten radio channels and half a dozen TV stations with no adverts, relatively unobtusive DOGs and decent programming. From One Xtra to Radio Four, CBBBC to The Parliament Channel, our tax money delivers something almost everyone can enjoy (yes, even those who can't see or hear, thanks to news.bbc.co.uk etc.)
Thanks to the unique way that the BBC is funded, virtually all UK slashdot readers contribute to the development of Internet multicast AV distribution which will one day allow everyone to host TV shows watched by millions (just add talent). This is stuff that won't ever be paid for by for-profit corporations because it cuts out the middle man - you'll be able to go straight to your favourite production teams and buy their material without watching useless adverts. Good for the BBC (which isn't trying to make a profit anyway), but bad for Fox/ Sky, and for local "value added" distributors that don't add value.
We already know the answers to that, the Joint Photographic Expert Group and Motion Picture Expert Group both created a number of standards some of which are partially encumbered.
On the one hand with JPEG we all use the unencumbered simple encoding, not the IBM patented arithmetic encoding (better compression) which is also in the same standards document. But with MPEG more or less all the essential character of the standards is patented by one or more contributors. So people just implement it without permission.
So the determining factor is whether you can get something "nearly as good" without accepting the patent encumbered stuff. If you can, people will drop the patented bits like a hot potato. Unfortunately it's not good enough to deliver something "nearly as good" later, because now it's not nearly as good after all since it doesn't interoperate with the existing content. This explains the limited popularity of Vorbis, for example.
ITV's shows are dire, as atested not only by critical failure (not winning many awards these days are you, ITV?) but also by poor audience figures. Some ITV regions are supported by the taxpayer indirectly, but it's true that the large part of programming and broadcasting is funded through the obnoxious advertising.
Channel 4 is partly government funded, and seeks grants for its, uh, unconventional programming from European projects which are themselves... government funded. Whether it means sending film crews to Italian beaches to film topless women, or showing 30 year old obscure Dutch movies about bicycling in 16:9 with subtitles, C4 reads the latest funding trends from Brussels and incorporates their needs into its schedule.
Channel 5 is entirely pointless and should never have been launched on analog. The government (the one you think shouldn't be interfering) forced them to add the movies and news bulletins which break up their otherwise relentless schedule of old material bought from other networks. In some cases the BBC (which you don't like) paid for this material (which you apparently DO like) to be made more than 20 years ago. Didn't you notice how the average C5 program seems kinda... retro?
In general I'm not in favour of government interference, but it's the reality we face. The technology for everyone and their dog to try to run a TV station doesn't exist yet, and might not for another decade. In the absence of that situation the invisible hand of market forces cannot operate properly, so the government inevitably must REGULATE broadcasting activity or we'll experience the spiral of reduced expectations. Once the government actively regulates the activity you're going to pay those taxes, and you might as well get something useful out of it. I think the BBC is fairly good value for money, and would support direct taxation rather than the "license fee" to support it until better means are available, despite the fact that this would inevitably mean that I personally wind up paying more for the same service.
Theora is a conventional (block, motion, color transform, throw away bits, then ordinary compression) 2nd generation video codec, it is alive and well, and it reached bitstream freeze just a couple of months ago. Presumably beta and then final releases of the software & associated documentation will follow in good time.
Tarkin is the Ogg wavelet codec. You're correct that work on Tarkin has more or less stalled, but wavelet codecs are a legal quagmire today, in part because so many people have conflicting patents in this area and are just waiting for the chance to litigate. Are any of the images on your website JPEG2000 instead of regular JFIF? Thought not.
Intel has been co-operating by producing GPL'd but semi-undocumented drivers for their Centrino laptop wirless chipsets. The concept seems to be that GPL'd drivers with poor/no documentation can be maintained by 3rd parties but don't make it any easier to hack in illegal features (too much power, wrong frequencies) than it already is in Windows.
The drivers will eventually merge into 2.6.x unless some catastrophe prevents that. They're already suitable for RPM / DEB packaging if someone else using your distro cares to set that up.
ALSA works much better than OSS/Free for most people, but of course there's always a minority group in any such change who suffer. Did you report your problems upstream both to the Fedora bugzill and to the ALSA developers?
Neither of my banks owns any public buildings. One exists only on the Internet, the other was originally spun out for banking by telephone and now also does its business on the Internet. So...
-- Opening new accounts: Bank #1, telephone call "I'm Tialaramex, and I would like to open an account with your bank" "That's great sir, tell us your postal address & telephone number and we'll send the government-required paperwork to you by 1st class post"
Bank #2, web page "Fill out details, choose type of account, read long contract terms, click OK" "That's great sir, the government-required paperwork will arrive on your doorstep soon."
In each case I had a bank account within 10 days, and I could have lived anywhere so long as I was able to receive regular mail.
-- Information Both banks have a web form labelled "Enquiries" or similar and I've filled it out once or twice. A human reads what you type and you get an answer the next banking day, just as quickly as if you'd waited for a branch to open, and without skipping lunch.
The telephone bank has polite and efficient operators who typically answer within 5-10 seconds. I've called from various places around the world where net access was expensive or unavailable, and they even have the capability to order a local bank to hand you a pile of cash if your cards are eaten or lost, so that you don't waste time waiting for new cards. "I've come to collect $500, my bank sent me." "Um, OK, let me go ask the manager... do you have a passphrase?" "Yes, it's Purple Jaguar", "Here's your $500 sir"
In both cases the staff work with comprehensive knowledge bases, and have more experienced support that they can forward problem cases to. They can do this because they've centralised the knowledge, rather than trying to put one experienced manager in every major town. I've never yet had a wrong or unhelpful answer despite asking awkward questions (e.g. "Where's the nearest ATM to the Eurostar station in Brussels?" it turns out that there's one inside, near one of the exits).
--Mortgages I've yet to borrow more than a few thousand pounds from either of my banks, but I'm assured by endless inserts in my paper statements (I opted to have one bank send me those before I had ADSL) that it's quite possible to arrange the mortgage entirely without travelling to a bank. As with opening an account the paperwork is filled out for you, and then sent by snail mail to be signed.
--Supplemental financial services I've used the Internet-only bank to create savings bonds and ISAs, and to consider but ultimately reject various other investments.
I've also used a related financial company to sell shares, once again entirely online. This happens in more or less realtime, and I can't fathom how anyone would attempt such a thing through a conventional bank. Most likely you'd lose a substantial proportion of the gain in overheads.
-- Major transactions Sorry to disappoint you, but unless you make prior arrangements the bank is never under obligation to hand you big stacks of cash. Your original agreement upon opening an account probably says that you agree to give them say 2-3 business days notice before attempting to withdraw more than EUR 5000 or closing the account. They MAY choose to waive this requirement, but at a small outlying branch it's quite possible that they simply don't have EUR 5000 to spare after the rest of the day's transactions. The days when a bank robber could expect to find dozens of people's life savings in a vault at the back of the building are long gone.
To buy something as expensive as a new car with cash has never been easy. Might I suggest that this is hardly a common enough sort of activity to justify the continued expense of a bank branch in every town?
BTW What town is large enough to support a car dealership but doesn't have a bank? Of course, if you buy your car online they can accept the money online too, killing two birds with one stone.
We use C-like languages to make things go very, very fast, immediately. Sometimes a high level language _could_ deliver this if we were willing to wait for the hardware architecture to be re-designed in its favour (which we're not), and sometimes it's just not possible because the C-like language lets you do things which cannot be proven by the machine to be safe, yet nevertheless are correct. Even scary old gets() can be quite harmless, under certain carefully controlled circumstances.
Now, some people use C (or even more stupidly, C++) to make petty database GUIs, or configure their font preferences, and here I agree that it's worthless, and a modern high-level language would be more appropriate. But note that even when Red Hat use Python throughout their management applets, they still crash, just with a run-time error (a list is too short, or a string is found when a number was expected) instead of SIGSEGV. Some of those mistakes might even have security implications... Programmers make mistakes in any and every language.
You can implement that security model on any PAM based system. Actually I proposed this setup for a (Linux) system I co-administrate, but the other administrators didn't like it. Maybe your favourite Linux vendor can be persuaded to support it officially in some future distribution, in the same way that most offer to use modern password crypto, setuid-helpers which remember your authentication etc.
There's no such thing as "too dependent on technology". This is one of the Big Lies of ludditism, that some arbitrary level of technology is OK, but anything exceeding that in some way diminishes us as human beings.
The "massive, permanent electric power outage" falls into the same category as Day of the Triffids style events. They're all incredibly unlikely and we have no way of planning for them in bulk. Learning to grow carrots is no more useful against this general class of disaster than buying a shortwave radio (maybe it gives advance warning of Triffid attack) or buying a pocket translator for an obscure tribal language (maybe Babel was real and will happen again).
Remember, to actually trigger this disaster all known ways of generating significant electrical power must fail, presumably simultaneously in a way that can't be repaired in days or weeks. Might as well ask what we're going to do if gravity stops working.
"But if I'm stuck with a lame system that automatically copies my text to the clipboard when I highlight it, I'm screwed."
The only significant systems to do this were early versions of Qt/KDE, due to a misunderstanding by Trolltech. I'm not aware of any serious X app that still has this problem in current versions.
X has always provided both a current 'PRIMARY' selection (to be updated whenever you explicitly select something, and can usually be pasted with the 3rd mouse button if you have one) and a 'CLIPBOARD' (to be updated when you choose to "Cut" or "Copy" and pasted when you choose "Paste").
Most of the time PRIMARY is faster (if you know about it) but it's not a substitute for the clipboard 100% of the time, and people who pretend it is are mostly just ignorant.
In both Windows and Linux the clipboard functionality is provided by the windowing system, and the _interface_ to that functionality is provided by the toolkit. So in both cases "almost all" programs will do what you want, but there will be exceptions because when a programmer creates their own widgets, or does something unusual with an existing widget they have to re-implement the functionality.
That means when you see a generic entry widget 99.99% of the time (on Linux or Windows) Ctrl-C will do what you expect. But when you see an app-specific widget like an audio waveform preview, or a 3D wireframe model then you can't be so sure... Worse, even in big name apps like Excel you may find that someone decided the expected behaviour wasn't so expected after all. Hit Ctrl-X while a region is selected in Excel. What happened? That's right, Excel's developers knew "better" and decided to delay the effect of the operation.
There _is_ a misunderstanding in Roblimo's rant here, because Ctrl-C/ Ctrl-V do the same thing in Linux as in Windows. The middle mouse paste trick is a power user shortcut, it doesn't support more complicated operations like "replace" or "cut" and so while I do miss it on Windows it's not as though Ctrl-C should be unfamiliar.
"hardware raid is much superior, especially with hotswap and auto-rebuild capabilities."
Are you actually familiar with Linux software RAID?
We have a disk failure (own a lot of disks, bound to have some fail), the system automatically rebuilds the array on the fly using the hot spare. If the affected disk is hot swap then the manufacturer's replacement is inserted when it arrives and the machine has no downtime. If it's not hot swap then the new spare is added during the next scheduled downtime (typically once per month). If a 2nd failure occurs the array will run in degraded mode, and at this point management can make the call on whether to have an unscheduled maintenance event.
We _could_ do a lot of things in hardware, but in general it is better not to unless price/ performance is on the hardware's side. That's not the case with RAID, so IMNSHO people should choose software RAID when given the option, and spend the extra money on better disks and/or more CPU throughput.
It's not "select to copy", and it's frustrating that so many people who think they "know Unix" not only believe this, but tell other people to believe it as well.
When you select some text, you're just selecting some text. You're not copying it anywhere, although yes, you can 'paste' the current selection using the middle mouse button.
When you use Edit->Copy or Ctrl-C or whatever in your app to actually COPY something, that gets put on the virtual clipboard, and you can Edit->Paste or Ctrl-V it later, even if you idly select something else in between.
Yes, that means you can replace a selection with something you copied earlier, just like on Windows or MacOS. Select the thing to be copied, hit Ctrl-C, select the thing to be replaced, and hit Ctrl-V.
It's actually been working how you wanted all this time, and you've been deliberately crippling yourself because/someone/ told you an untruth and you believe them.
X has supported this for longer than Windows existed BUT (and this is the huge BUT) no-one is demanding support for this stuff from the apps. No amount of architectural support and documentation makes one single X application actually paste images.
In X11 (and presumably Windows) there is a negotiation step during the Paste phase of the Cut/Copy/Paste clipboard system, where the conversation goes like this:
App1: Hi App2, I hear you have the clipboard contents. What's in there? App2: Hi App1, I have ASCII, UTF8, HTML, SLASHDOT-RANT or INSANE-NONSENSE App1: Gee, most of that went straight over my head, I'll take some UTF8 please. App2: Here you go: UTF8 text follows
All you need to do is convince X app authors who might have some use for it to add XPIXMAP or whatever to their send/receive acceptable types list and then write the appropriate encoding and decoding stuff for their app.
Once one or two popular apps do this AT ALL, it would be worth going to FreeDesktop and getting a simple standard written which says e.g. what format the clipboard image should be, and how to encode/decode it.
Slashdot Mirror
Actually lots of these Slashdot readers have Videolan Client, mplayer and other multicast tools installed already. So its you that's "just making shit up". Some of them may even have multicast conferencing and multicast file distribution software too. It's just that all this stuff doesn't work on today's Internet for most people because their ISPs don't want to deploy it.
Right now, if the ISPs actually made their end of the system work properly, you could have 1000 people listening to a 128kbit Ogg Vorbis stream from your home ADSL, each of them using 128kbits at their end, and a total of 128kbits at your end, no huge bills for fat pipes, everyone walks away happy, except the ISPs. From the point of view of an ISP it's far better to have some poor clown paying $$$$ for the ability to send 128Mbit/s in order to connect 1000 people at 128 kbit/s unicast. Wasteful, but it all generates more revenue, and revenue feeds ISPs.
Many manufacturers already operate a "no dead pixels" policy, BUT you need to read the small print to discover that they don't regard a pixel in which one of the three color components is faulty as "dead". Only one of the subpixels is dead.
So if you buy one of their LCDs and it has a red subpixel that has failed, leaving a permanently red-tinted pixel, that's not considered a serious enough fault for a DOA replacement.
Faults which cause a whole pixel to die are relatively rare and certainly more obvious, so that's why they replace LCDs with this problem. You're still going to pay a premium if you want to buy a batch of LCDs and return all the ones with display faults. At least for a while yet. If you (as a dealer or reseller) want to buy such zero defect screens, you need ISO class 1, instead of the consumer class 2.
If you're a consumer, buying a single LCD, and you're having sleepless nights over dead (sub)pixels, go to a real world store, and insist on either opening & testing the LCD or buying a display model so that you can see exactly what you're getting. You will undoubtedly pay more than you would buying online, but you'll be certain to avoid dead (sub)pixels this way. Display faults are practically always a manufacturing problem, not something that develops over time.
Paying for megabits sent makes no sense. The only reason they charge you per kilojoule for the electric bill is because it costs money to make electricity. Do you think packets cost money? Can your five year old bankrupt you by running too many packets through ::1 (127.0.0.1 to the old-schoolers) ?
Access charges are how you charge for networks. Sometimes idiots let other people charge them use charges for the network (eg. consumer POTS service long distance or international charges) but that's basically a license to print money for whoever does the charging.
[ Consider for a moment a five minute period in two parallel universes. In one you call Cape Town from New York over POTS for 5 minutes. In the other you just sit there meditating. Why did your POTS company get $2 or so of income in one universe but not in the other? What were their additional expenses incurred ? ]
What you get to charge for access depends on the quality of service, which means typical and guaranteed bitrate, availability, service limitations (e.g. no outbound SMTP packets) and so on. You contract with other access providers to exchange packets in order to increase the network value for everyone (routing policy).
There are three separate problems here, all of which can be fixed, and the result could be markedly better than Ceefax, although I grant that it isn't better yet.
1. Content. The digital subtitles are good, the rest of the content is weak. This would be fixed by spending more on content provision, but the BBC is reluctant to do so while digital takeup is still relatively slow.
2. Bandwidth. With the very restricted piece of radio spectrum assigned for digital TV it's hard to justify using more for rarely viewed digital interactive content, when it could be used to upgrade picture quality for millions of viewers. The easy way to fix this would be to kill Channel 5's analog slots and re-assign the frequencies, freeing up more space for BBCi and other low priority services. If C5 was appropriately compensated even their shareholders might support this idea.
3. Low quality devices. Have you used a non-fasttext TV to view Ceefax lately? Not so great is it. Right now the relatively high cost & low demand means that you can't buy a freeview or cable box that delivers the intended experience. They don't keep enough data (especially images) in cache, and they crash too easily. Time will heal this one, just like today's web browsers are a lot better than those from a decade ago. For a while you'll be able to pay more for a better experience, and then, like fasttext it will just become a standard feature on a chip somewhere.
An out-of-box Mac, or a recent Linux distro (e.g. Fedora Core 2) will automatically use IPv6 for most things if placed on an IPv6 connected network.
If you have IPv4 Internet on an FC2 machine, five minutes of work provides IPv6 to that machine and to every computer on the same LAN. If those computers run recent Linux or OS X they too will then be fully IPv6 enabled with no further work. Improved performance is enabled by the usual route: complain to your ISP about the lousy performance or switch to a better one.
On Windows you need to type in a command (I know, most people can't do that) or wait for Longhorn. Yet again Microsoft holds up the traffic.
Of course because it Just Works(TM) there aren't a lot of people shouting about this in the streets.
Intel's trick only works if there is no alternative supplier. Only one company made 486DXs, there are dozens just in my area competing to provide Internet service over DSL. If one says "We charge extra for $arbitrarything" then the others seize on it, and advertise their Free ArbitraryThing, especially if $arbitrarything is a fixed cost or very low per-user cost item.
/48 of IPv6 space under 6to4. So if your ISP takes away your one IPv4 routable address in exchange for even a /64 you're being ripped off to the tune of 2^16 already in address space terms.
Remember that every SINGLE IPv4 address is automatically granted & routed the associated
No, the age of charging $5 per month for an IP address is over, whether or not IPv6 takes off in the next few years. In 1995 most people didn't know that they wanted the Internet. In 2000 most of them couldn't see why they'd need it in more than one room. Now some of the same people can't stand to be without it while on a train or in a pub. You're not going to convince them that a single address is enough.
And what about the people who needed human blood products in order to stay alive, who became infected and will eventually die because there is not only no vaccine, but no effective cure?
Still, so long as there's SOME way to avoid being infected there's no need for a vaccine right? Even the common cold is vanquished by your amazing discovery, simply never breathe the same air as other people without reprocessing.
"Look at the math - if the world population stabilizes at 9 billion, those resources relinquished by the West will simply be consumed by the growing third-world population."
So what?
A lot of crazy people have been running around saying that we're consuming too much of everything, as a species. They've been saying it for about a century now, and it isn't getting any more true. They point at charts (which show hilarious things like an assumption of no net energy input to the planet, ah hahahah) and they hold committee meetings and occasionally they write miserable books. They anthropomorphise an entire planet (Hello Gaia theory) in order to support their demand for arbitrary lifestyle changes which would make them feel comfortable.
I'm not interested in preventing hippies from drinking rainwater, living in cardboard houses, recycling plastic bottles or whatever else they want to do. I am interested in preserving my rather comfortable and wholely sustainable lifestyle despite the hippies, and thus I don't like to see them press their agenda as the "only sustainable way forward" regardless of whether it's printed on recycled paper. This planet is quite capable of feeding 10 billion people take-out Indian food in aluminium containers for longer than all recorded history, and even letting them all drive home with it, and eat it in air conditioned comfort.
That makes no sense. If you would normally receive a packet (e.g. because you provide web service, or have an IM port open or whatever) then the NAT router will rewrite the packets so that you still receive the trojan.
OTOH if you wouldn't normally receive something (e.g. it's an HTTP attack and you don't run a web server) then the NAT makes no difference, you still won't receive it. Big deal.
NATs are not magical protective charms. They're just a desperate hack to get around running out of IP addresses. If you want a firewall, install a firewall, not a NAT.
You can write testcases, but this is equivalent to the task undertaken by your attackers, with the unfortunate distinction that you must find ALL the bugs, and they only need ONE that you missed.
:(
I've written testcases for N:M ratio software where many different applications make use of many different interoperable components (e.g. SANE's frontends and backends, or LADSPA's plugins and hosts). In this case it's important to preserve interop by detecting components that expect or provide something not agreed by the interface specification.
It was quite common to write SANE scanner backends which only handled large, even-sized buffers properly. With most frontends (typically requesting 64 kbytes at a time) this works fine, and so the authors had considered their code to be of good quality. But network users (who request Ethernet packet sized buffers) had a lot of problems. These days the SANE testbed application includes a test routine to check for this and report it as a bug in the backend. Unlike the network frontend (which not everyone can really test) this test is built as part of the SDK, so no driver author has an excuse for not using it to QA their code.
We never read about blackhats or greyhats who spent six months attacking a product with nothing to show for it, just like we never read about terrorists who survey a target and call off the operation because it's too well guarded. Only our failures get headlines
We like to call it "The United Kingdom" (which it isn't) or "Great Britain" (which is arguable) but it's certainly not bizarro world.
Remember, for a small once per year tax all UK residents have ten radio channels and half a dozen TV stations with no adverts, relatively unobtusive DOGs and decent programming. From One Xtra to Radio Four, CBBBC to The Parliament Channel, our tax money delivers something almost everyone can enjoy (yes, even those who can't see or hear, thanks to news.bbc.co.uk etc.)
Thanks to the unique way that the BBC is funded, virtually all UK slashdot readers contribute to the development of Internet multicast AV distribution which will one day allow everyone to host TV shows watched by millions (just add talent). This is stuff that won't ever be paid for by for-profit corporations because it cuts out the middle man - you'll be able to go straight to your favourite production teams and buy their material without watching useless adverts. Good for the BBC (which isn't trying to make a profit anyway), but bad for Fox/ Sky, and for local "value added" distributors that don't add value.
We already know the answers to that, the Joint Photographic Expert Group and Motion Picture Expert Group both created a number of standards some of which are partially encumbered.
On the one hand with JPEG we all use the unencumbered simple encoding, not the IBM patented arithmetic encoding (better compression) which is also in the same standards document. But with MPEG more or less all the essential character of the standards is patented by one or more contributors. So people just implement it without permission.
So the determining factor is whether you can get something "nearly as good" without accepting the patent encumbered stuff. If you can, people will drop the patented bits like a hot potato. Unfortunately it's not good enough to deliver something "nearly as good" later, because now it's not nearly as good after all since it doesn't interoperate with the existing content. This explains the limited popularity of Vorbis, for example.
ITV's shows are dire, as atested not only by critical failure (not winning many awards these days are you, ITV?) but also by poor audience figures. Some ITV regions are supported by the taxpayer indirectly, but it's true that the large part of programming and broadcasting is funded through the obnoxious advertising.
Channel 4 is partly government funded, and seeks grants for its, uh, unconventional programming from European projects which are themselves... government funded. Whether it means sending film crews to Italian beaches to film topless women, or showing 30 year old obscure Dutch movies about bicycling in 16:9 with subtitles, C4 reads the latest funding trends from Brussels and incorporates their needs into its schedule.
Channel 5 is entirely pointless and should never have been launched on analog. The government (the one you think shouldn't be interfering) forced them to add the movies and news bulletins which break up their otherwise relentless schedule of old material bought from other networks. In some cases the BBC (which you don't like) paid for this material (which you apparently DO like) to be made more than 20 years ago. Didn't you notice how the average C5 program seems kinda... retro?
In general I'm not in favour of government interference, but it's the reality we face. The technology for everyone and their dog to try to run a TV station doesn't exist yet, and might not for another decade. In the absence of that situation the invisible hand of market forces cannot operate properly, so the government inevitably must REGULATE broadcasting activity or we'll experience the spiral of reduced expectations. Once the government actively regulates the activity you're going to pay those taxes, and you might as well get something useful out of it. I think the BBC is fairly good value for money, and would support direct taxation rather than the "license fee" to support it until better means are available, despite the fact that this would inevitably mean that I personally wind up paying more for the same service.
Theora is a conventional (block, motion, color transform, throw away bits, then ordinary compression) 2nd generation video codec, it is alive and well, and it reached bitstream freeze just a couple of months ago. Presumably beta and then final releases of the software & associated documentation will follow in good time.
Tarkin is the Ogg wavelet codec. You're correct that work on Tarkin has more or less stalled, but wavelet codecs are a legal quagmire today, in part because so many people have conflicting patents in this area and are just waiting for the chance to litigate. Are any of the images on your website JPEG2000 instead of regular JFIF? Thought not.
Intel has been co-operating by producing GPL'd but semi-undocumented drivers for their Centrino laptop wirless chipsets. The concept seems to be that GPL'd drivers with poor/no documentation can be maintained by 3rd parties but don't make it any easier to hack in illegal features (too much power, wrong frequencies) than it already is in Windows.
The drivers will eventually merge into 2.6.x unless some catastrophe prevents that. They're already suitable for RPM / DEB packaging if someone else using your distro cares to set that up.
ALSA works much better than OSS/Free for most people, but of course there's always a minority group in any such change who suffer. Did you report your problems upstream both to the Fedora bugzill and to the ALSA developers?
Neither of my banks owns any public buildings. One exists only on the Internet, the other was originally spun out for banking by telephone and now also does its business on the Internet. So...
-- Opening new accounts:
Bank #1, telephone call "I'm Tialaramex, and I would like to open an account with your bank" "That's great sir, tell us your postal address & telephone number and we'll send the government-required paperwork to you by 1st class post"
Bank #2, web page "Fill out details, choose type of account, read long contract terms, click OK" "That's great sir, the government-required paperwork will arrive on your doorstep soon."
In each case I had a bank account within 10 days, and I could have lived anywhere so long as I was able to receive regular mail.
-- Information
Both banks have a web form labelled "Enquiries" or similar and I've filled it out once or twice. A human reads what you type and you get an answer the next banking day, just as quickly as if you'd waited for a branch to open, and without skipping lunch.
The telephone bank has polite and efficient operators who typically answer within 5-10 seconds. I've called from various places around the world where net access was expensive or unavailable, and they even have the capability to order a local bank to hand you a pile of cash if your cards are eaten or lost, so that you don't waste time waiting for new cards. "I've come to collect $500, my bank sent me." "Um, OK, let me go ask the manager... do you have a passphrase?" "Yes, it's Purple Jaguar", "Here's your $500 sir"
In both cases the staff work with comprehensive knowledge bases, and have more experienced support that they can forward problem cases to. They can do this because they've centralised the knowledge, rather than trying to put one experienced manager in every major town. I've never yet had a wrong or unhelpful answer despite asking awkward questions (e.g. "Where's the nearest ATM to the Eurostar station in Brussels?" it turns out that there's one inside, near one of the exits).
--Mortgages
I've yet to borrow more than a few thousand pounds from either of my banks, but I'm assured by endless inserts in my paper statements (I opted to have one bank send me those before I had ADSL) that it's quite possible to arrange the mortgage entirely without travelling to a bank. As with opening an account the paperwork is filled out for you, and then sent by snail mail to be signed.
--Supplemental financial services
I've used the Internet-only bank to create savings bonds and ISAs, and to consider but ultimately reject various other investments.
I've also used a related financial company to sell shares, once again entirely online. This happens in more or less realtime, and I can't fathom how anyone would attempt such a thing through a conventional bank. Most likely you'd lose a substantial proportion of the gain in overheads.
-- Major transactions
Sorry to disappoint you, but unless you make prior arrangements the bank is never under obligation to hand you big stacks of cash. Your original agreement upon opening an account probably says that you agree to give them say 2-3 business days notice before attempting to withdraw more than EUR 5000 or closing the account. They MAY choose to waive this requirement, but at a small outlying branch it's quite possible that they simply don't have EUR 5000 to spare after the rest of the day's transactions. The days when a bank robber could expect to find dozens of people's life savings in a vault at the back of the building are long gone.
To buy something as expensive as a new car with cash has never been easy. Might I suggest that this is hardly a common enough sort of activity to justify the continued expense of a bank branch in every town?
BTW What town is large enough to support a car dealership but doesn't have a bank? Of course, if you buy your car online they can accept the money online too, killing two birds with one stone.
We use C-like languages to make things go very, very fast, immediately. Sometimes a high level language _could_ deliver this if we were willing to wait for the hardware architecture to be re-designed in its favour (which we're not), and sometimes it's just not possible because the C-like language lets you do things which cannot be proven by the machine to be safe, yet nevertheless are correct. Even scary old gets() can be quite harmless, under certain carefully controlled circumstances.
Now, some people use C (or even more stupidly, C++) to make petty database GUIs, or configure their font preferences, and here I agree that it's worthless, and a modern high-level language would be more appropriate. But note that even when Red Hat use Python throughout their management applets, they still crash, just with a run-time error (a list is too short, or a string is found when a number was expected) instead of SIGSEGV. Some of those mistakes might even have security implications... Programmers make mistakes in any and every language.
You can implement that security model on any PAM based system. Actually I proposed this setup for a (Linux) system I co-administrate, but the other administrators didn't like it. Maybe your favourite Linux vendor can be persuaded to support it officially in some future distribution, in the same way that most offer to use modern password crypto, setuid-helpers which remember your authentication etc.
There's no such thing as "too dependent on technology". This is one of the Big Lies of ludditism, that some arbitrary level of technology is OK, but anything exceeding that in some way diminishes us as human beings.
The "massive, permanent electric power outage" falls into the same category as Day of the Triffids style events. They're all incredibly unlikely and we have no way of planning for them in bulk. Learning to grow carrots is no more useful against this general class of disaster than buying a shortwave radio (maybe it gives advance warning of Triffid attack) or buying a pocket translator for an obscure tribal language (maybe Babel was real and will happen again).
Remember, to actually trigger this disaster all known ways of generating significant electrical power must fail, presumably simultaneously in a way that can't be repaired in days or weeks. Might as well ask what we're going to do if gravity stops working.
"But if I'm stuck with a lame system that automatically copies my text to the clipboard when I highlight it, I'm screwed."
The only significant systems to do this were early versions of Qt/KDE, due to a misunderstanding by Trolltech. I'm not aware of any serious X app that still has this problem in current versions.
X has always provided both a current 'PRIMARY' selection (to be updated whenever you explicitly select something, and can usually be pasted with the 3rd mouse button if you have one) and a 'CLIPBOARD' (to be updated when you choose to "Cut" or "Copy" and pasted when you choose "Paste").
Most of the time PRIMARY is faster (if you know about it) but it's not a substitute for the clipboard 100% of the time, and people who pretend it is are mostly just ignorant.
In both Windows and Linux the clipboard functionality is provided by the windowing system, and the _interface_ to that functionality is provided by the toolkit. So in both cases "almost all" programs will do what you want, but there will be exceptions because when a programmer creates their own widgets, or does something unusual with an existing widget they have to re-implement the functionality.
That means when you see a generic entry widget 99.99% of the time (on Linux or Windows) Ctrl-C will do what you expect. But when you see an app-specific widget like an audio waveform preview, or a 3D wireframe model then you can't be so sure... Worse, even in big name apps like Excel you may find that someone decided the expected behaviour wasn't so expected after all. Hit Ctrl-X while a region is selected in Excel. What happened? That's right, Excel's developers knew "better" and decided to delay the effect of the operation.
There _is_ a misunderstanding in Roblimo's rant here, because Ctrl-C/ Ctrl-V do the same thing in Linux as in Windows. The middle mouse paste trick is a power user shortcut, it doesn't support more complicated operations like "replace" or "cut" and so while I do miss it on Windows it's not as though Ctrl-C should be unfamiliar.
"hardware raid is much superior, especially with hotswap and auto-rebuild capabilities."
Are you actually familiar with Linux software RAID?
We have a disk failure (own a lot of disks, bound to have some fail), the system automatically rebuilds the array on the fly using the hot spare. If the affected disk is hot swap then the manufacturer's replacement is inserted when it arrives and the machine has no downtime. If it's not hot swap then the new spare is added during the next scheduled downtime (typically once per month). If a 2nd failure occurs the array will run in degraded mode, and at this point management can make the call on whether to have an unscheduled maintenance event.
We _could_ do a lot of things in hardware, but in general it is better not to unless price/ performance is on the hardware's side. That's not the case with RAID, so IMNSHO people should choose software RAID when given the option, and spend the extra money on better disks and/or more CPU throughput.
"Most /. readers won't understand this because they think Linux is easy to use (it's not)."
/. users called jamezilla won't understand that because they think their opinions and beliefs are shared by everyone else (they're not).
Linux is easy for me to use.
Most
It's not "select to copy", and it's frustrating that so many people who think they "know Unix" not only believe this, but tell other people to believe it as well.
/someone/ told you an untruth and you believe them.
When you select some text, you're just selecting some text. You're not copying it anywhere, although yes, you can 'paste' the current selection using the middle mouse button.
When you use Edit->Copy or Ctrl-C or whatever in your app to actually COPY something, that gets put on the virtual clipboard, and you can Edit->Paste or Ctrl-V it later, even if you idly select something else in between.
Yes, that means you can replace a selection with something you copied earlier, just like on Windows or MacOS. Select the thing to be copied, hit Ctrl-C, select the thing to be replaced, and hit Ctrl-V.
It's actually been working how you wanted all this time, and you've been deliberately crippling yourself because
X has supported this for longer than Windows existed BUT (and this is the huge BUT) no-one is demanding support for this stuff from the apps. No amount of architectural support and documentation makes one single X application actually paste images.
In X11 (and presumably Windows) there is a negotiation step during the Paste phase of the Cut/Copy/Paste clipboard system, where the conversation goes like this:
App1: Hi App2, I hear you have the clipboard contents. What's in there?
App2: Hi App1, I have ASCII, UTF8, HTML, SLASHDOT-RANT or INSANE-NONSENSE
App1: Gee, most of that went straight over my head, I'll take some UTF8 please.
App2: Here you go: UTF8 text follows
All you need to do is convince X app authors who might have some use for it to add XPIXMAP or whatever to their send/receive acceptable types list and then write the appropriate encoding and decoding stuff for their app.
Once one or two popular apps do this AT ALL, it would be worth going to FreeDesktop and getting a simple standard written which says e.g. what format the clipboard image should be, and how to encode/decode it.