Given the apparently limitless supply of candidates for positions in government and the lack of demand from citizens, administration members have a replacement value of approximately zero. It doesn't really seem worth worrying whether you have to restock them wholesale or retail...
Most of the mobile (phone) ISPs do this routinely in the UK - they proxy HTTP connections, compress images and insert Javascript into the web pages which detect some key combination to restore the original page contents. You could perhaps understand this for GPRS, but the same applies to 3G and HSDPA: the ISP's caching proxy diligently compresses a bunch of JPEGs and modifies your web pages just so you can have a worse user experience on your 3.6MBit/sec connection...
Although the Ad industry might think otherwise, advertising isn't an end in itself, its purpose is to extend people's horizons beyond what they simply need (and will therefore seek out) to goods and services that they might want, or can be persuaded to want for as long as it takes the cash to leave their fingers.
We've always had "advertising" - markets existed to get people together not only for trading convenience but also to provide an audience for retail opportunities; shop windows have always been, well, "shop windows" to display goods to the passing masses. Itinerant peddlers would bring the foreign and unfamiliar to small local communities. Newspapers have featured advertising since their inception - indeed, the front page of The Times used to be where the announcements could be read - the news was on the inside.
The character of advertising opportunities do influence the kind of goods and services that can be developed, though. Not control, but influence - you'd be hard pressed to develop a mass-market brand without mass-market advertising. So if you can't put out a TV ad that hits 20 million people, there may be a change in the kind of businesses that exist. You might not have so many new global brands, no more international quasi-monopolies like Microsoft, for example. It might be bad news for the purveyors of overpriced, environmentally-harmful fizzy drinks and great for sales of lemonade from the front lawn. On the other hand, the cost of insurance has been driven down by direct insurers dealing with the public over the phone or internet and eliminating tiers of costly intermediaries: without mass advertising, people would use the insurance office in their local high street but pay more for the "privilege".
So "the end of advertising as we know it" isn't going to be the end of advertising, but it might well have an effect on the kind of new businesses that can be created - you can only have a business if you have a market and if the market can't be reached en masse then you might have to serve people individually as well as advertise to them individually. Which sounds great in principle, but doesn't come cheap.
One of the features of this little spat is that iPlayer is designed (actually, bought in from Verisign...) as a P2P application, so most of the data shifted will not originate from the BBC in the first place. So this is really an attempt by ISPs to charge the BBC for data that is actually moving between ISP's end users.
Interestingly, both Tiscali and BT (cited as participants in this by El Reg) have their own video-over-IP services (Homechoice http://www.homechoice.co.uk/ rebranded as Tiscali TV and BT Vision http://www.bt.com/vision) and it's probably not a surprise they'd like to disadvantage the BBC in that marketplace.
I tend to the view that iPlayer is a broken means of distributing TV-on-demand, but it's broken partly because the ISPs don't seem interested in developing better mechanisms. It's certainly not going to get fixed by ISPs whingeing that consumers want what they've been sold at the cost they were promised. But since a recent survey (http://www.theregister.co.uk/2007/08/08/line_rent al_cost_discouraging/) revealed that consumers are very resistant to broadband price hikes, it does seem like the ISPs have dug themselves a hole that can only get deeper.
Consider a real example outside the software domain - the PAL TV system. It's a simple, but significant improvement over NTSC in that the phase of the colour subcarrier is reversed every line so that hue errors caused by phase shifts in the transmission path appear merely as desaturation. It's a clever idea, but not particularly revolutionary: the main innovation was in fact being able to produce a consumer-grade device (piezo-acoustic delay line) to make it work in practice. The system was patented and used as the basis of analogue colour TV across most of Europe. The patents were then used to prevent non-European manufacturers enter the European market (France, of course, wanted to protect its market even from other European manufacturers and adopted SECAM). Once the PAL patents expired, the formerly-protected TV manufacturers were exposed to competition from the Far East and mostly closed their European factories.
So, who benefited from these patents? Not consumers, who were forced to pay over the odds for TVs for years. Not, ultimately, the manufacturers who'd failed to improve the productivity of their manufacturing plants because they were not subject to price competition. Not European innovation, since manufacturers did not have to continue to innovate, simply rest on their laurels.
Interestingly, the PAL patents are often used as an example of the SUCCESS of the patent system...
I guess it may just be my age, but playing other people's records seems like a rather vicarious talent. Is it really a "right" to make a bit of pocket money/find small-pond fame/be cool-by-association on the back of someone else's creativity?
Most people want to hear songs that are familiar to them and those songs becomes familiar largely by being pushed by an industry that has honed its art since the days of sheet music. If you want to be associated with/exploit the the public image of a band or its music, then you have to buy into the system that delivered it and pay the cost.
Of course, these broadcasters could spend their day off-air commissioning or creating some new and original music, but that would cost money or time and probably no-one would listen to it. But at least it would be a genuinely creative activity.
This is not just a PVR phenomenon - people who would have watched TV in the past have other media to divert them. This in turn drives down advertising revenues, reduces programming quality and further reduces audiences. It's just starker if the audience you have isn't watching the advertising anyway.
So guess what's coming...
Content Protection and Copy Management (documents, EFF critique) a sort of super-DRM that applies not only to a single TV receiver, but pervades every device to which the protected content might be copied. Although there are reassuring words about this regime only applying to "premium" content, all the mechanisms are there to disable recording, restrict the number of devices having access to the content simultaneously and cause the content to evaporate after a certain period of time. So the broadcasters are clearly thinking about how to preserve their income stream.
Of course, we shouldn't be surprised, even public broadcasters are getting addicted to rights-management. Although you can make a perfectly good permanent copy of an off-air MPEG programme stream from any BBC broadcast, if you're part of the BBC's iPlayer pilot you donate your Internent bandwidth to their P2P service and in return receive a Windows Media file of the same programme at one quarter of the resolution which self-destructs 7 days after you first play it. It's not quite clear who this is protecting now, but it's not a great leap to suggest that unencumbered recording is now seen as an historic error by the controlling suits.
Of course, if you want TV programmes in their traditional sense, they have to be paid for somehow. The BBC, despite their current DRM frenzy, are guaranteed an income from the TV licence fee (or at least until the government decides otherwise). Advertising revenue is, though, inexorably dropping. In the UK the rules for commercial broadcasters were relaxed to permit sponsorship and, in future, product placement, but that's not going to make a huge difference to lower-profile content. There's also been a major scandal over the use of premium-rate phone lines which have been used to supplement the income stream of a wide range of programmes under the flimsy pretext of "interactivity". So the advertising model may well be doomed.
There are payment models which continue to work: pay per view (the traditional cinema model), subscription (eg cable, satellite) and the reviled but suprisingly resilient TV licence. If advertising-supported TV no longer makes economic sense, it might mean the end of broadcast TV as it's know in the USA, but it's not necessarily the end of broadcast TV in countries which have other ways of funding free-to-air television.
I suspect that applying DRM to try to shore up a declining industry is more likely to kill it off quickly, though!
I wouldn't rely on FoI - the BBC seem to be extraordinarily reticent about anything to do with iPlayer. People who are signed up to the current "closed trial" are reminded that "all information regarding BBC iPlayer Beta is confidential". The T&Cs for its use also include an injunction "to not subject the BBC iPlayer and/or the BBC Content to any derogatory treatment or use them in such a way that would bring the BBC into disrepute".
So never mind the content, you can't even talk about the delivery mechanism, still less make fun of it!
Actually, the specific example of transcription-by-ear is one of the reasons there was great pressure to develop international copyright law.
In the 19th Century, authors and producers of popular musical entertainment (such as, at the time, Gilbert & Sullivan) had to go to extraordinary lengths to prevent "entrepreneurial" productions of their works appearing on the American stage within weeks of the London production as a result of transcribers busily noting down the entire work in the audience - the Victorian equivalent of a "screener".
You can reasonably debate whether copyright is a "good thing" and how long it should last, but if there is to be copyright, then being able to transcribe by ear is just as much a copyright infringement as photocopying the sheet music.
There's an old joke about Bell having to follow up his invention of the telephone with the invention of another telephone so that he had someone to talk to.
The IP equivalent is that the first people to have IPv6 addresses and no IPv4 addresses essentially have access only to each other. If you want to cut yourself off from the Internet, it would be much cheaper to cancel your broadband subscription.
Of course, if you use some sort of gateway/proxy/NAT to restore connectivity, you might has well have done so using IPv4 gateway/proxy/NAT solutions. The only way to get from IPv4 to IPv6 that doesn't require IPv4 to survive indefinitely, is to move everyone before they need to - even if you could communicate that to the entire population of the globe, how would you encourage a change or know when it had occurred?
This isn't a new dilemma - the same thing happened to DECnet 20 years ago and look how successful that has become!
IPv6 is also based on the assumption that universal end-to-end connectivity is a "good thing". Although CIDR has made some contribution to reducing the exhaustion rate of the IPv4 address space, a big factor has been that large users put their entire networks in private address space behind firewalls and proxies. Their primary motivation for that is not the conservation of address space, it's to keep universal end-to-end connectivity as far from their desktops as possible.
There will be a new IP at some point which may fix addressing problems, but only as a consequence of fixing something more pressing and tangible to the end user.
Although there's been a relatively high take-up of Digital TV in the UK (about 19 million homes roughly split 45%/45%/10% DVB-S/DVB-T/DVB-C depending on whose figures you believe) there have been two distinct driving factors for this - one is consumer demand for a wider choice of channels and the other is government determination to shut down analogue terrestrial TV in order to make money from the spectrum space by auctioning it.
To a first approximation, the technologies are each presently operated by a separate single supplier: DVB-S by Murdoch's Sky TV, DVB-C by Virgin Media (formerly NTL/Telewest) and DVB-T by Freeview (a consortium in which the BBC is a main player).
The principle advantage of "Freeview" is that it provides a very simple marketing message: you buy a £30 box which plugs into your existing aerial (mostly) and TV and that's it. In exchange you get both a wider range of channels and protection against the "analogue switch-off".
That message is getting progressively less simple, though. Freeview has subscription channels (entertainment and sport) as a result of the legacy of the collapsed On-Digital/ITV Digital service that preceded it (needs box with a card slot). There is also some limited PPV (mainly porn) which interestingly requires no card but an access code obtained by telephone.
Furthermore, there is a significant demand from broadcasters for additional Freeview slots - recently there have been auctions for broadcast rights which have reached millions of pounds per channel. In order to prevent interference to analogue signals, the number and power of DVB-T channels is artificially limited at present leading to a seriously-constrained supply of channels. Partly this has been addressed by improving the encoding and stat-muxing process, partly by compromising on quality (some services average below 2MBit/sec video rate) but basically the available spectrum is full. Sky TV (which is also a content provider as well as operating a DVB-S platform) has 3 channels on Freeview which it has contemplated pulling and replacing with 4 MPEG-4 channels which existing Freeview hardware cannot receive in order to increase the number of channels it can provide on DVB-T.
So, despite the current constraints, DVB-T has been a success. The danger is that the capacity constraints will cause the platform to fragment as different content-providers try to deal with this and the consequent increasing costs by invoking a range of incompatible technical solutions and payment mechanisms.
It was originally believed that the closure of analogue TV would enable a significant increase in the coverage and capacity of Freeview, but the government has since made it very clear that it intends to auction off the spectrum space rather than simply re-allocate it to Freeview as it becomes available. Consequently, the future expansion of Freeview is in doubt and the BBC in particular is concerned that it might only be able to provide HD programmes on DVB-T by sacrificing other channels.
This technical research by the BBC is very much a desperate bid to retain the future viability of DVB-T, in which it now has a significant stake, in the face of current government policy. So don't assume it's the BBC's preferred option: the recent DVB-T HD trials used rather more mundane and easily-deployable technology.
At the same time, there are consultations on switching off FM radio (about which there is a serious outcry as the DAB alternative is also seriously constrained by limited spectrum space meaning that the audio quality of FM is actually better; one possible consequence of which is that FM is switched off and DAB is enhanced incompatibly with current receivers) and trials replacing AM radio with DRM (that's Digital Radio Mondiale in this context!).
In almost 90 years of public broadcasting there has historically been only one major technology shift in Europe that has obsoleted consumer equipment - the move to colour TV (which led, after a decent interval, to the shutting down
I've been in a similar situation. I'd done some development work for a public (but non-government) body without any such NDA stipulations; the body was then taken over by a government agency which then tried to impose these kind of conditions retrospectively to complete the small remaining part of the work. As soon as I consulted a lawyer they freaked out because their procedures didn't allow for anyone to question their standard terms and conditions, even when they had not applied at the start of the contract.
They certainly had no intention of negotiating and ultimately attempted to redevelop the entire system for themselves (3 people for 2 years) before deciding that it wasn't the job of a government agency to be a software house and so scrapped their development effort. They then went out to buy a different package from a 3rd party over which they could assert no rights as it was entirely developed prior to their purchase of it. As I write they'are at least a year behind in the rollout of this new package.
So don't underestimate the stupidity of your potential clients. They may well be prepared to spend vast sums of time & money and have no ultimate stake in the IPR of their solution provided that they can demonstrate they have stuck to the standard terms and conditions their lawyers worked dilligently to produce.
From my point of view, the fact they walked away was a good deal in the short term: I got more revenue from supporting the original system than I ever expected to get out of the project simply because they couldn't afford to be without it while they figured out how to get rid of me. Bad news in the long term, though, because this is a part of the world where most work is government-funded in one way or another and you mysteriously stop being asked to tender for work if you seem to be "difficult".
How much of a business are you likely to have if you make it a policy not to accept contract of this kind? If you've plenty of other work, then fine; if not, it comes down to your need to feed versus your self-esteem.
One of the things too few people allow for in their business plan is the "fuck you" factor. Businesses tend to be highly geared to secure growth and it can be difficult to walk away from potential clients because you need the money to service your borrowing and pay for your staff. I've always made sure there was a reserve account to make it easier to say "no" now and again. However, there's always some stuff you just have to swallow if you want an income: having a business is worse than being employed in that respect because at least an employee has rights...
Great, I thought - an interview with one of the brightest people I've ever worked with: must be full of insight and wisdom.
Don't even waste your time reading it. Just a couple of dull, out-of-context remarks about P2P that the interviewer picked out of what I hope was a rather more interesting conversation. Who is Richard MacManus - and why?
I regularly use my PC for recording (digital) TV and radio, but I find myself creating DVDs to watch the stuff "offline". It's mainly down to the (lack of) ease of use of the software and its ridiculous resource requirements (which effectively mean I can't do anything else with the PC while I'm recording or watching TV).
The idea that you should have your PC desktop on the TV is kind of daft, but using your PC with a media adaptor as an adjunct to a TV isn't a bad idea at all. It ought to be straightforward and cheap - a digital TV stream (SD) doesn't get much above 5Mbit/sec and an adaptor to display it back on a TV is pretty cheap (eg Hauppauge MediaMVP) so it should be a fairly easy task to get a modern PC to churn away in the background organising and retrieving media.
Not so on my 2.5GHz PC - I can't effectively use the PC for anything else whilst recording TV and due to some rather bizarre design choices it appears that I can't effectively use it when playing back TV using the Hauppauge software. And, of course, none of the software integrates together.
Unfortunately, other media-server products seem to be just as bad. MCE has the desktop-on-TV paradigm which kind of rules it out of being used as a proper PC (though I acknowledge that you can use it with a networked media device which may overcome this problem).
If I can't use my PC as a PC while it's doing its media thing, and need access to the PC screen to launch the various bits of random software, then integration has failed - I'd be better off with separate boxes.
The CPU requirements of a media box are decidely modest so it might make sense to turn the model around: have a cheap media server with a network interface a host PC can use to access its EPG, recordings, etc, but which can also function standalone. Some of the latest DVB hard-disk recorders come close to this (remote access to recordings but no remote control from the PC), so it wouldn't be a big step.
My father was an architect for many years and he has many examples of going to inspect sites and finding that the construction crew had misunderstood or wilfully ignored the specifications for critical structural components: columns incorrectly constructed for the projected load, or a pile of roof components rusting in a corner of the site and clearly not installed in the almost-completed roof. Building workers are fairly hazy on the niceties of engineering and are on the kind of contracts that make it attractive to get as much money for doing as little work as possible; they do, generally, though, have a "comfort zone" of familiarity with traditional construction techniques which is why most regular construction projects don't fall down.
Anyone specifying a new or unusual process has to be aware of the fact that the typical construction worker won't believe it's important to follow the rules exactly, won't understand which parts of the process are most vital and won't be around at the end of the project to take any responsbility. If you have a design that depends on technology unfamiliar to the people who're responsible for implementing it, then you need tight supervision during the build and tight inspection afterwards. You often don't get either - the foremen are on bonuses to accelerate the construction phase and the people most qualified to inspect afterwards are the people who designed the structure in the first place.
Of course there are many projects which are simply not feasible using traditional construction, but for those that are, any apparent savings from using new technology can be negated by the costs of ensuring it's correctly applied.
Indeed. It's noteworthy that a lot of the work being done on "The Semantic Web" is by academics. They come from an environment in which there are peer-review mechanisms and established publishing channels which ensure that "trust" is the norm. Outside that world, information is generally less trustworthy but it may still be relevant. The research challenge is to make use of "trust" where it can be proved to exist but not assume it elsewhere. In commercial terms, though, it may not be worth even trying to do this for the generality of information on the web, only a tiny proportion of which will ever fall within a "web of trust".
TBL also originally intended that hyperlinks be bi-directional: the target of the link would also link back to the origin. Clearly a useful idea but only practical in a small community with controlled access to documents. It's not unreasonable that there should be more feature-rich versions of the web for use in certain community groups, but it doesn't mean they would necessarily suit the Information Highway's Wacky Racers...
A long time ago, until the idea of the stored program came along, computers were programmed with patch leads, so programs were little more than knitting patterns. The instruction sets of early stored program computers were developed around the hardware and were typically pretty chaotic. Then computers came along with consistent instructions sets that were specifically designed to assist compiler-writers (eg the VAX), but they no longer mapped terribly efficiently to hardware operations (as I recall, a single VAX instruction could cause over 50 page faults before completing!), so the "machine code" was a kind of intermediate language which was then interpreted by the instruction unit of the particular piece of hardware the program ran on. We're now kind of going back to the hardware-led instruction set (out-of-order execution, vector instructions, etc) simply because the balance of performance in the various parts of the computer (execution unit, memory, etc) has shifted again.
The purpose of human-oriented programming languages (and I guess C just about qualifies...) is partly to make it easier to express the problem (and perhaps this is where we stop talking about C) and partly to provide a level of abstraction so that you don't have to rewrite the program for every piece of hardware. You can, of course, get the same effective abstraction from an intermediate language (such as the Java or.net virtual machine languages).
But notice the number of abstractions we may have - high level language is compiled to an intermediate language that is then compiled to a CPU instruction set that is then interpreted by microcode. And don't forget the abstractions taking place in the CPU hardware - the addresses are virtualised, memory is cached, instructions are re-ordered.
If you could control every aspect of the placement of your data, the use of the address-translation caches, the individual gates of the CPU, then your program would no doubt be very efficient. It would also run on only one piece of hardware.
There's always going to be a loss of potential performance if you don't code down to the gate level, but there are compensating benefits in productivity and the longevity of the solution, providing you choose a level of abstraction that is a natural fit for the problem you are trying to solve. If you try to use C to write a program for which it isn't particularly well suited (and that's a long list), then it's unlikely the compiler or anything else can help you!
If there is a shortage of "skills" (as seen by employers) it's because the "skills" people have are regularly invalidated by changes in technology which serve only the narrow economic interests of the technology manufacturers. In most forms of engineering, technology develops in such a way as to make it possible to build bigger/better/lighter/stronger/faster. Indeed, in the field of computing, the development of silicon has pretty well followed this model. In software engineering, though, each new "innovation" (.net, C#, Java, Windows Vista...) makes only a marginal difference (if any) to what could be achieved with a previous generation of technology - whilst imposing a huge cost in terms of training and redundancy of previous knowledge. It's as if every three years there was a decree such as "OK, you can't use steel for building bridges any more: the new material is antelope carcases".
Which is why "skill" in IT is not usefully defined as a knowledge of PHP or ASP or C++ or how to write a Linux device driver. Expecting a recruiter to understand this is as futile as expecting him to return your calls, though.
I know a number of (UK) mailorder businesses that routinely store the card number, expiry date and CVV of all transactions. It's either done for convenience (if a refund is required later you don't have to phone the customer to get the card number) or because of operational issues (for example, there is a batch process that extracts the payment details from one system and passes it to another to actually debit the card and it has to be repeatable in case one part of the process fails: the lazy solution is to store everything indefinitely).
The need to retain customer confidence in the card-processing system means that the interesting question of who would be liable in the case of a mass theft is unlikely to be tested in court - even if it were useful to do so (a lot of mailorder businesses are not cash rich and neither are the software companies that supply them).
This risk will persist until there is some sort of two-factor authentication on all card transactions.
The main problem I've had is that if the database becomes corrupted for whatever reason, it can become essentially irreparable. One example: a "text" field in one record became corrupt (the "toast" indices didn't match up) and it then became impossible to vacuum, dump or otherwise maintain the database: each operation would fail when in encountered the corrupt record. Reverting to the last "good" backup isn't really an option in a database that gets a lot of updates during the day and the documentation on alternative recovery strategies (transaction logs et al) seems mainly to come from the realms of folklore. I've also had problems with broken system indexes where re-indexing has simply (and silently) made a whole bunch of records inaccessible. There's a lot I like about PostgreSQL (particularly the ability to have derived tables which inherit columns from their parents), but it needs work on the administration and deployment side.
It could be very useful to have an environment (hardware, or a virtual environment on a PC) which replicated an (encrypted) private data store with a central server and did the same (in the other direction) for a public application repository. Documents available everywhere, continue to work if network down, no software installation (it just arrives as required). However, it would require a *standard* environment which delivered a decent, interactive user experience for all application types. So not AJAX. Or Java (with its current set of UIs).
Arguably, of course, this is where Microsoft is heading with.NET... though you'll have to watch your credit card as that software lands in your local cache.
... Sony Corporation denies it has been using the knowledge to develop a "cerebral root kit" in an attempt to "attract" new customers to its music division..
>You have huge amounts of data in that file that is not text, and you can't guarantee that the text is stored as actual text
True, you can't. But if I open a random selection of.DOC files with a text editor I can see the contents of most of them. Archaeology doesn't necessarily depend on being able to decode every word of every document: most historical documents haven't survived or have survived only in a damaged form. That's partly why archeologists have jobs.
>contracts are still typically printed, signed, and stored, all on wood pulp paper
A significant number of contracts which I have seen or been a party to are deemed on signature to incorporate "standard terms and conditions" which are supplied electronically. Typically, there's no real mechanism for ensuring that both parties to the contract actually are in possession of the same electronic document, never mind a process for identifying specific revisions or being able to recover them in the event of a liability claim in 20 years. Of course it would be wise to print a copy, but this wouldn't prevent someone arguing later that the document had been amended without authorisation before printing.
Well, it's not quite that bad from an archaeological point of view: many word processor files contain text encoded in a standard character set in roughly the order it should be read - the text just has extraneous mark-up. I'm sure future civilisations can do without knowing exactly which bits of the text were originally rendered in Comic Sans.
From a legal and administrative point of view, though, it's worse - you'll need an accurate and reproducible record of the contracts you've made for the lifetime of those contracts and any liability arising from them. I wouldn't want to find myself in court arguing about who had to pay for the consequences of a collapsed bridge on the basis of "the most likely reconstruction of the original contract we can come up with"... Or indeed the blueprints, which are another aspect of the same problem.
Slashdot Mirror
Given the apparently limitless supply of candidates for positions in government and the lack of demand from citizens, administration members have a replacement value of approximately zero. It doesn't really seem worth worrying whether you have to restock them wholesale or retail...
Most of the mobile (phone) ISPs do this routinely in the UK - they proxy HTTP connections, compress images and insert Javascript into the web pages which detect some key combination to restore the original page contents. You could perhaps understand this for GPRS, but the same applies to 3G and HSDPA: the ISP's caching proxy diligently compresses a bunch of JPEGs and modifies your web pages just so you can have a worse user experience on your 3.6MBit/sec connection...
Although the Ad industry might think otherwise, advertising isn't an end in itself, its purpose is to extend people's horizons beyond what they simply need (and will therefore seek out) to goods and services that they might want, or can be persuaded to want for as long as it takes the cash to leave their fingers.
We've always had "advertising" - markets existed to get people together not only for trading convenience but also to provide an audience for retail opportunities; shop windows have always been, well, "shop windows" to display goods to the passing masses. Itinerant peddlers would bring the foreign and unfamiliar to small local communities. Newspapers have featured advertising since their inception - indeed, the front page of The Times used to be where the announcements could be read - the news was on the inside.
The character of advertising opportunities do influence the kind of goods and services that can be developed, though. Not control, but influence - you'd be hard pressed to develop a mass-market brand without mass-market advertising. So if you can't put out a TV ad that hits 20 million people, there may be a change in the kind of businesses that exist. You might not have so many new global brands, no more international quasi-monopolies like Microsoft, for example. It might be bad news for the purveyors of overpriced, environmentally-harmful fizzy drinks and great for sales of lemonade from the front lawn. On the other hand, the cost of insurance has been driven down by direct insurers dealing with the public over the phone or internet and eliminating tiers of costly intermediaries: without mass advertising, people would use the insurance office in their local high street but pay more for the "privilege".
So "the end of advertising as we know it" isn't going to be the end of advertising, but it might well have an effect on the kind of new businesses that can be created - you can only have a business if you have a market and if the market can't be reached en masse then you might have to serve people individually as well as advertise to them individually. Which sounds great in principle, but doesn't come cheap.
One of the features of this little spat is that iPlayer is designed (actually, bought in from Verisign...) as a P2P application, so most of the data shifted will not originate from the BBC in the first place. So this is really an attempt by ISPs to charge the BBC for data that is actually moving between ISP's end users.
t al_cost_discouraging/) revealed that consumers are very resistant to broadband price hikes, it does seem like the ISPs have dug themselves a hole that can only get deeper.
Interestingly, both Tiscali and BT (cited as participants in this by El Reg) have their own video-over-IP services (Homechoice http://www.homechoice.co.uk/ rebranded as Tiscali TV and BT Vision http://www.bt.com/vision) and it's probably not a surprise they'd like to disadvantage the BBC in that marketplace.
I tend to the view that iPlayer is a broken means of distributing TV-on-demand, but it's broken partly because the ISPs don't seem interested in developing better mechanisms. It's certainly not going to get fixed by ISPs whingeing that consumers want what they've been sold at the cost they were promised. But since a recent survey (http://www.theregister.co.uk/2007/08/08/line_ren
Consider a real example outside the software domain - the PAL TV system. It's a simple, but significant improvement over NTSC in that the phase of the colour subcarrier is reversed every line so that hue errors caused by phase shifts in the transmission path appear merely as desaturation. It's a clever idea, but not particularly revolutionary: the main innovation was in fact being able to produce a consumer-grade device (piezo-acoustic delay line) to make it work in practice. The system was patented and used as the basis of analogue colour TV across most of Europe. The patents were then used to prevent non-European manufacturers enter the European market (France, of course, wanted to protect its market even from other European manufacturers and adopted SECAM). Once the PAL patents expired, the formerly-protected TV manufacturers were exposed to competition from the Far East and mostly closed their European factories.
So, who benefited from these patents? Not consumers, who were forced to pay over the odds for TVs for years. Not, ultimately, the manufacturers who'd failed to improve the productivity of their manufacturing plants because they were not subject to price competition. Not European innovation, since manufacturers did not have to continue to innovate, simply rest on their laurels.
Interestingly, the PAL patents are often used as an example of the SUCCESS of the patent system...
I guess it may just be my age, but playing other people's records seems like a rather vicarious talent. Is it really a "right" to make a bit of pocket money/find small-pond fame/be cool-by-association on the back of someone else's creativity?
Most people want to hear songs that are familiar to them and those songs becomes familiar largely by being pushed by an industry that has honed its art since the days of sheet music. If you want to be associated with/exploit the the public image of a band or its music, then you have to buy into the system that delivered it and pay the cost.
Of course, these broadcasters could spend their day off-air commissioning or creating some new and original music, but that would cost money or time and probably no-one would listen to it. But at least it would be a genuinely creative activity.
So guess what's coming...
Content Protection and Copy Management (documents, EFF critique) a sort of super-DRM that applies not only to a single TV receiver, but pervades every device to which the protected content might be copied. Although there are reassuring words about this regime only applying to "premium" content, all the mechanisms are there to disable recording, restrict the number of devices having access to the content simultaneously and cause the content to evaporate after a certain period of time. So the broadcasters are clearly thinking about how to preserve their income stream.
Of course, we shouldn't be surprised, even public broadcasters are getting addicted to rights-management. Although you can make a perfectly good permanent copy of an off-air MPEG programme stream from any BBC broadcast, if you're part of the BBC's iPlayer pilot you donate your Internent bandwidth to their P2P service and in return receive a Windows Media file of the same programme at one quarter of the resolution which self-destructs 7 days after you first play it. It's not quite clear who this is protecting now, but it's not a great leap to suggest that unencumbered recording is now seen as an historic error by the controlling suits.
Of course, if you want TV programmes in their traditional sense, they have to be paid for somehow. The BBC, despite their current DRM frenzy, are guaranteed an income from the TV licence fee (or at least until the government decides otherwise). Advertising revenue is, though, inexorably dropping. In the UK the rules for commercial broadcasters were relaxed to permit sponsorship and, in future, product placement, but that's not going to make a huge difference to lower-profile content. There's also been a major scandal over the use of premium-rate phone lines which have been used to supplement the income stream of a wide range of programmes under the flimsy pretext of "interactivity". So the advertising model may well be doomed.
There are payment models which continue to work: pay per view (the traditional cinema model), subscription (eg cable, satellite) and the reviled but suprisingly resilient TV licence. If advertising-supported TV no longer makes economic sense, it might mean the end of broadcast TV as it's know in the USA, but it's not necessarily the end of broadcast TV in countries which have other ways of funding free-to-air television.
I suspect that applying DRM to try to shore up a declining industry is more likely to kill it off quickly, though!
I wouldn't rely on FoI - the BBC seem to be extraordinarily reticent about anything to do with iPlayer. People who are signed up to the current "closed trial" are reminded that "all information regarding BBC iPlayer Beta is confidential". The T&Cs for its use also include an injunction "to not subject the BBC iPlayer and/or the BBC Content to any derogatory treatment or use them in such a way that would bring the BBC into disrepute".
So never mind the content, you can't even talk about the delivery mechanism, still less make fun of it!
Actually, the specific example of transcription-by-ear is one of the reasons there was great pressure to develop international copyright law.
In the 19th Century, authors and producers of popular musical entertainment (such as, at the time, Gilbert & Sullivan) had to go to extraordinary lengths to prevent "entrepreneurial" productions of their works appearing on the American stage within weeks of the London production as a result of transcribers busily noting down the entire work in the audience - the Victorian equivalent of a "screener".
You can reasonably debate whether copyright is a "good thing" and how long it should last, but if there is to be copyright, then being able to transcribe by ear is just as much a copyright infringement as photocopying the sheet music.
There's an old joke about Bell having to follow up his invention of the telephone with the invention of another telephone so that he had someone to talk to.
The IP equivalent is that the first people to have IPv6 addresses and no IPv4 addresses essentially have access only to each other. If you want to cut yourself off from the Internet, it would be much cheaper to cancel your broadband subscription.
Of course, if you use some sort of gateway/proxy/NAT to restore connectivity, you might has well have done so using IPv4 gateway/proxy/NAT solutions. The only way to get from IPv4 to IPv6 that doesn't require IPv4 to survive indefinitely, is to move everyone before they need to - even if you could communicate that to the entire population of the globe, how would you encourage a change or know when it had occurred?
This isn't a new dilemma - the same thing happened to DECnet 20 years ago and look how successful that has become!
IPv6 is also based on the assumption that universal end-to-end connectivity is a "good thing". Although CIDR has made some contribution to reducing the exhaustion rate of the IPv4 address space, a big factor has been that large users put their entire networks in private address space behind firewalls and proxies. Their primary motivation for that is not the conservation of address space, it's to keep universal end-to-end connectivity as far from their desktops as possible.
There will be a new IP at some point which may fix addressing problems, but only as a consequence of fixing something more pressing and tangible to the end user.
Although there's been a relatively high take-up of Digital TV in the UK (about 19 million homes roughly split 45%/45%/10% DVB-S/DVB-T/DVB-C depending on whose figures you believe) there have been two distinct driving factors for this - one is consumer demand for a wider choice of channels and the other is government determination to shut down analogue terrestrial TV in order to make money from the spectrum space by auctioning it.
To a first approximation, the technologies are each presently operated by a separate single supplier: DVB-S by Murdoch's Sky TV, DVB-C by Virgin Media (formerly NTL/Telewest) and DVB-T by Freeview (a consortium in which the BBC is a main player).
The principle advantage of "Freeview" is that it provides a very simple marketing message: you buy a £30 box which plugs into your existing aerial (mostly) and TV and that's it. In exchange you get both a wider range of channels and protection against the "analogue switch-off".
That message is getting progressively less simple, though. Freeview has subscription channels (entertainment and sport) as a result of the legacy of the collapsed On-Digital/ITV Digital service that preceded it (needs box with a card slot). There is also some limited PPV (mainly porn) which interestingly requires no card but an access code obtained by telephone.
Furthermore, there is a significant demand from broadcasters for additional Freeview slots - recently there have been auctions for broadcast rights which have reached millions of pounds per channel. In order to prevent interference to analogue signals, the number and power of DVB-T channels is artificially limited at present leading to a seriously-constrained supply of channels. Partly this has been addressed by improving the encoding and stat-muxing process, partly by compromising on quality (some services average below 2MBit/sec video rate) but basically the available spectrum is full. Sky TV (which is also a content provider as well as operating a DVB-S platform) has 3 channels on Freeview which it has contemplated pulling and replacing with 4 MPEG-4 channels which existing Freeview hardware cannot receive in order to increase the number of channels it can provide on DVB-T.
So, despite the current constraints, DVB-T has been a success. The danger is that the capacity constraints will cause the platform to fragment as different content-providers try to deal with this and the consequent increasing costs by invoking a range of incompatible technical solutions and payment mechanisms.
It was originally believed that the closure of analogue TV would enable a significant increase in the coverage and capacity of Freeview, but the government has since made it very clear that it intends to auction off the spectrum space rather than simply re-allocate it to Freeview as it becomes available. Consequently, the future expansion of Freeview is in doubt and the BBC in particular is concerned that it might only be able to provide HD programmes on DVB-T by sacrificing other channels.
This technical research by the BBC is very much a desperate bid to retain the future viability of DVB-T, in which it now has a significant stake, in the face of current government policy. So don't assume it's the BBC's preferred option: the recent DVB-T HD trials used rather more mundane and easily-deployable technology.
At the same time, there are consultations on switching off FM radio (about which there is a serious outcry as the DAB alternative is also seriously constrained by limited spectrum space meaning that the audio quality of FM is actually better; one possible consequence of which is that FM is switched off and DAB is enhanced incompatibly with current receivers) and trials replacing AM radio with DRM (that's Digital Radio Mondiale in this context!).
In almost 90 years of public broadcasting there has historically been only one major technology shift in Europe that has obsoleted consumer equipment - the move to colour TV (which led, after a decent interval, to the shutting down
I've been in a similar situation. I'd done some development work for a public (but non-government) body without any such NDA stipulations; the body was then taken over by a government agency which then tried to impose these kind of conditions retrospectively to complete the small remaining part of the work. As soon as I consulted a lawyer they freaked out because their procedures didn't allow for anyone to question their standard terms and conditions, even when they had not applied at the start of the contract.
They certainly had no intention of negotiating and ultimately attempted to redevelop the entire system for themselves (3 people for 2 years) before deciding that it wasn't the job of a government agency to be a software house and so scrapped their development effort. They then went out to buy a different package from a 3rd party over which they could assert no rights as it was entirely developed prior to their purchase of it. As I write they'are at least a year behind in the rollout of this new package.
So don't underestimate the stupidity of your potential clients. They may well be prepared to spend vast sums of time & money and have no ultimate stake in the IPR of their solution provided that they can demonstrate they have stuck to the standard terms and conditions their lawyers worked dilligently to produce.
From my point of view, the fact they walked away was a good deal in the short term: I got more revenue from supporting the original system than I ever expected to get out of the project simply because they couldn't afford to be without it while they figured out how to get rid of me. Bad news in the long term, though, because this is a part of the world where most work is government-funded in one way or another and you mysteriously stop being asked to tender for work if you seem to be "difficult".
How much of a business are you likely to have if you make it a policy not to accept contract of this kind? If you've plenty of other work, then fine; if not, it comes down to your need to feed versus your self-esteem.
One of the things too few people allow for in their business plan is the "fuck you" factor. Businesses tend to be highly geared to secure growth and it can be difficult to walk away from potential clients because you need the money to service your borrowing and pay for your staff. I've always made sure there was a reserve account to make it easier to say "no" now and again. However, there's always some stuff you just have to swallow if you want an income: having a business is worse than being employed in that respect because at least an employee has rights...
Great, I thought - an interview with one of the brightest people I've ever worked with: must be full of insight and wisdom.
Don't even waste your time reading it. Just a couple of dull, out-of-context remarks about P2P that the interviewer picked out of what I hope was a rather more interesting conversation. Who is Richard MacManus - and why?
I regularly use my PC for recording (digital) TV and radio, but I find myself creating DVDs to watch the stuff "offline". It's mainly down to the (lack of) ease of use of the software and its ridiculous resource requirements (which effectively mean I can't do anything else with the PC while I'm recording or watching TV).
The idea that you should have your PC desktop on the TV is kind of daft, but using your PC with a media adaptor as an adjunct to a TV isn't a bad idea at all. It ought to be straightforward and cheap - a digital TV stream (SD) doesn't get much above 5Mbit/sec and an adaptor to display it back on a TV is pretty cheap (eg Hauppauge MediaMVP) so it should be a fairly easy task to get a modern PC to churn away in the background organising and retrieving media.
Not so on my 2.5GHz PC - I can't effectively use the PC for anything else whilst recording TV and due to some rather bizarre design choices it appears that I can't effectively use it when playing back TV using the Hauppauge software. And, of course, none of the software integrates together.
Unfortunately, other media-server products seem to be just as bad. MCE has the desktop-on-TV paradigm which kind of rules it out of being used as a proper PC (though I acknowledge that you can use it with a networked media device which may overcome this problem).
If I can't use my PC as a PC while it's doing its media thing, and need access to the PC screen to launch the various bits of random software, then integration has failed - I'd be better off with separate boxes.
The CPU requirements of a media box are decidely modest so it might make sense to turn the model around: have a cheap media server with a network interface a host PC can use to access its EPG, recordings, etc, but which can also function standalone. Some of the latest DVB hard-disk recorders come close to this (remote access to recordings but no remote control from the PC), so it wouldn't be a big step.
My father was an architect for many years and he has many examples of going to inspect sites and finding that the construction crew had misunderstood or wilfully ignored the specifications for critical structural components: columns incorrectly constructed for the projected load, or a pile of roof components rusting in a corner of the site and clearly not installed in the almost-completed roof. Building workers are fairly hazy on the niceties of engineering and are on the kind of contracts that make it attractive to get as much money for doing as little work as possible; they do, generally, though, have a "comfort zone" of familiarity with traditional construction techniques which is why most regular construction projects don't fall down.
Anyone specifying a new or unusual process has to be aware of the fact that the typical construction worker won't believe it's important to follow the rules exactly, won't understand which parts of the process are most vital and won't be around at the end of the project to take any responsbility. If you have a design that depends on technology unfamiliar to the people who're responsible for implementing it, then you need tight supervision during the build and tight inspection afterwards. You often don't get either - the foremen are on bonuses to accelerate the construction phase and the people most qualified to inspect afterwards are the people who designed the structure in the first place.
Of course there are many projects which are simply not feasible using traditional construction, but for those that are, any apparent savings from using new technology can be negated by the costs of ensuring it's correctly applied.
Indeed. It's noteworthy that a lot of the work being done on "The Semantic Web" is by academics. They come from an environment in which there are peer-review mechanisms and established publishing channels which ensure that "trust" is the norm. Outside that world, information is generally less trustworthy but it may still be relevant. The research challenge is to make use of "trust" where it can be proved to exist but not assume it elsewhere. In commercial terms, though, it may not be worth even trying to do this for the generality of information on the web, only a tiny proportion of which will ever fall within a "web of trust".
TBL also originally intended that hyperlinks be bi-directional: the target of the link would also link back to the origin. Clearly a useful idea but only practical in a small community with controlled access to documents. It's not unreasonable that there should be more feature-rich versions of the web for use in certain community groups, but it doesn't mean they would necessarily suit the Information Highway's Wacky Racers...
A long time ago, until the idea of the stored program came along, computers were programmed with patch leads, so programs were little more than knitting patterns. The instruction sets of early stored program computers were developed around the hardware and were typically pretty chaotic. Then computers came along with consistent instructions sets that were specifically designed to assist compiler-writers (eg the VAX), but they no longer mapped terribly efficiently to hardware operations (as I recall, a single VAX instruction could cause over 50 page faults before completing!), so the "machine code" was a kind of intermediate language which was then interpreted by the instruction unit of the particular piece of hardware the program ran on. We're now kind of going back to the hardware-led instruction set (out-of-order execution, vector instructions, etc) simply because the balance of performance in the various parts of the computer (execution unit, memory, etc) has shifted again.
.net virtual machine languages).
The purpose of human-oriented programming languages (and I guess C just about qualifies...) is partly to make it easier to express the problem (and perhaps this is where we stop talking about C) and partly to provide a level of abstraction so that you don't have to rewrite the program for every piece of hardware. You can, of course, get the same effective abstraction from an intermediate language (such as the Java or
But notice the number of abstractions we may have - high level language is compiled to an intermediate language that is then compiled to a CPU instruction set that is then interpreted by microcode. And don't forget the abstractions taking place in the CPU hardware - the addresses are virtualised, memory is cached, instructions are re-ordered.
If you could control every aspect of the placement of your data, the use of the address-translation caches, the individual gates of the CPU, then your program would no doubt be very efficient. It would also run on only one piece of hardware.
There's always going to be a loss of potential performance if you don't code down to the gate level, but there are compensating benefits in productivity and the longevity of the solution, providing you choose a level of abstraction that is a natural fit for the problem you are trying to solve. If you try to use C to write a program for which it isn't particularly well suited (and that's a long list), then it's unlikely the compiler or anything else can help you!
... what Bruce Schneier is to security. Always worth listening to and usually right on the button.
If there is a shortage of "skills" (as seen by employers) it's because the "skills" people have are regularly invalidated by changes in technology which serve only the narrow economic interests of the technology manufacturers. In most forms of engineering, technology develops in such a way as to make it possible to build bigger/better/lighter/stronger/faster. Indeed, in the field of computing, the development of silicon has pretty well followed this model. In software engineering, though, each new "innovation" (.net, C#, Java, Windows Vista...) makes only a marginal difference (if any) to what could be achieved with a previous generation of technology - whilst imposing a huge cost in terms of training and redundancy of previous knowledge. It's as if every three years there was a decree such as "OK, you can't use steel for building bridges any more: the new material is antelope carcases".
Which is why "skill" in IT is not usefully defined as a knowledge of PHP or ASP or C++ or how to write a Linux device driver. Expecting a recruiter to understand this is as futile as expecting him to return your calls, though.
I know a number of (UK) mailorder businesses that routinely store the card number, expiry date and CVV of all transactions. It's either done for convenience (if a refund is required later you don't have to phone the customer to get the card number) or because of operational issues (for example, there is a batch process that extracts the payment details from one system and passes it to another to actually debit the card and it has to be repeatable in case one part of the process fails: the lazy solution is to store everything indefinitely).
The need to retain customer confidence in the card-processing system means that the interesting question of who would be liable in the case of a mass theft is unlikely to be tested in court - even if it were useful to do so (a lot of mailorder businesses are not cash rich and neither are the software companies that supply them).
This risk will persist until there is some sort of two-factor authentication on all card transactions.
The main problem I've had is that if the database becomes corrupted for whatever reason, it can become essentially irreparable. One example: a "text" field in one record became corrupt (the "toast" indices didn't match up) and it then became impossible to vacuum, dump or otherwise maintain the database: each operation would fail when in encountered the corrupt record. Reverting to the last "good" backup isn't really an option in a database that gets a lot of updates during the day and the documentation on alternative recovery strategies (transaction logs et al) seems mainly to come from the realms of folklore. I've also had problems with broken system indexes where re-indexing has simply (and silently) made a whole bunch of records inaccessible. There's a lot I like about PostgreSQL (particularly the ability to have derived tables which inherit columns from their parents), but it needs work on the administration and deployment side.
Perhaps not.
.NET ... though you'll have to watch your credit card as that software lands in your local cache.
It could be very useful to have an environment (hardware, or a virtual environment on a PC) which replicated an (encrypted) private data store with a central server and did the same (in the other direction) for a public application repository. Documents available everywhere, continue to work if network down, no software installation (it just arrives as required). However, it would require a *standard* environment which delivered a decent, interactive user experience for all application types. So not AJAX. Or Java (with its current set of UIs).
Arguably, of course, this is where Microsoft is heading with
... Sony Corporation denies it has been using the knowledge to develop a "cerebral root kit" in an attempt to "attract" new customers to its music division..
>You have huge amounts of data in that file that is not text, and you can't guarantee that the text is stored as actual text True, you can't. But if I open a random selection of .DOC files with a text editor I can see the contents of most of them. Archaeology doesn't necessarily depend on being able to decode every word of every document: most historical documents haven't survived or have survived only in a damaged form. That's partly why archeologists have jobs.
>contracts are still typically printed, signed, and stored, all on wood pulp paper
A significant number of contracts which I have seen or been a party to are deemed on signature to incorporate "standard terms and conditions" which are supplied electronically. Typically, there's no real mechanism for ensuring that both parties to the contract actually are in possession of the same electronic document, never mind a process for identifying specific revisions or being able to recover them in the event of a liability claim in 20 years. Of course it would be wise to print a copy, but this wouldn't prevent someone arguing later that the document had been amended without authorisation before printing.
Well, it's not quite that bad from an archaeological point of view: many word processor files contain text encoded in a standard character set in roughly the order it should be read - the text just has extraneous mark-up. I'm sure future civilisations can do without knowing exactly which bits of the text were originally rendered in Comic Sans. From a legal and administrative point of view, though, it's worse - you'll need an accurate and reproducible record of the contracts you've made for the lifetime of those contracts and any liability arising from them. I wouldn't want to find myself in court arguing about who had to pay for the consequences of a collapsed bridge on the basis of "the most likely reconstruction of the original contract we can come up with"... Or indeed the blueprints, which are another aspect of the same problem.