Posts by Steve Todd 2645 publicly visible posts • joined 19 Sep 2007
The standard ARM licence lets customers bolt together their standard bits of IP, and combine them with third party designs (you want a PowerVR graphics core rather than a Mali, certainly sir). The architecture licence lets them roll their own cores, with changes to things like the execution units (for higher IPC) or the memory manager (for higher throughput).
Apple have an ARM architecture licence. They design their own ARM compatible chips. They run the same ARM instruction set as the official designs, but otherwise contain no ARM IP. They aren't the only ones to do this either (Qualcomm and nVidia do likewise).
That's part of the genius of the ARM strategy, they'll sell you pre-rolled designs for anything from an M0 micro controller through to a fully fledged A73 64 bit application processor, or you can buy off the shelf chips from their partner companies. If you're working in huge volumes though you can design your own from the ground up, save a little on the licensing costs and (possibly) create a more efficient design.
You're confusing design with manufacture, and the f-35 was designed using more traditional methods. The bulkheads were originally machined from titanium blocks for example. The modern approach would be to create them from laser sintered titanium powder, which is a very much faster process.
Your failure to understand how your home items were made is showing. Most were made from parts that were original cast or molded, then were machine finished with traditional subtractive methods. Your car for example has many cast parts that are then precision machined to their final shape. This saves money over machining them from a solid block, but costs rather more in tooling and setup costs and produces weaker parts.
[Harvard] isn't an ISP. They are not breaking net neutrality by providing a free service any more than Google or El Reg are. The local ISP would be breaking net neutrality if they only provided a subset of the net for free, rather than an alowance. What if, for example Oxford or the Sorbon decide to offer free courses also. The locals wouldn't be able to access them under the scheme you describe, even though they may be better for some if not all users.
Also just because something is held up to be good by TPTB doesn't mean that Internet users should be forced to take it. The point of neutrality is to let the users decide what they want rather than forcing particular systems on them.
So speaks someone who has completely failed to understand net neutrality. ISPs are allowed to manage traffic for quality of service reasons, but can't prioritise one service above another based on payment by the originator. They can't for example charge Netflix to give priority to their video over other video streaming services, but they can slow down all video steaming if the network is congested.
@bazza - you are joking about Itanium right? It was always going to be a technical disaster. It relied on static compile time organisation of the code rather than dynamic run time out of order execution, so it couldn't adapt to changes. It emulated the x86 (badly), and took a bunch of silicon to do that. It was very late, and never got the critical mass required to make a modern CPU profitable. All in all Intel had holes in their heads when they came up with the idea.
One could be mistaken for thinking that Mr Orlowski has something in for smart watches, rather than just hit whoring that he usually does. He personally can't see the point in them, but apparently there are millions of people who can. If you don't want something then don't buy it. Don't think yours is the only opinion though, and don't try to put down those who think otherwise. If the manufacturer makes money out of their product then enough people have voted that they disagree with you.
The problem is that cable companies aren't investing in expanding their networks to large areas of the US, and that their service levels are poor for the areas that they do serve. Some tame accademic claiming that the cable companies offer better value triple pay packages doesn't change that, and isn't useful if you only want broadband. Some service for less than the cost of an (expensive) triple pay package is better than no service at all.
That municipal broadband schemes are a reaction to the cable companies failures to provide the coverage and service levels that they promised when they were granted their licences. Without incentive to compete and improve their service they have rested on their laurels while most of the first world passed them by.
Put another way, they demonstrably weren't investing beforehand, so any argument that municipal broadband stops them investing is purest bullsh*t.
How about never? Branch prediction isn't something code sees. If the prediction is correct then code executes seamlessly. If it misses then there's a pipeline stall while the correct instructions are loaded and executed. Your code has no sight over or access to the predictor, and it is still subject to the memory management unit allowing it access the code in the first place.
@dwarf, if an orbiting rocket applies a burst of thrust in the direction of the orbit then the altitude of the orbit will increase and the velocity will decrease until there is once more a balance between gravity trying to pull the rocket down and Newtonian physics trying to keep it going in a straight line. Climbing out of the gravity well takes energy, and that energy is subtracted from the kinetic energy (mass x velocity) of the craft. It seems counter intuitive, but low orbits are fast orbits.
@h4rm0ny - it wasn't just the public who were saying that 10 years is too long, it was parliament and the independent review. I've nothing against harmonising the sentences, but the implication is that the sentence for physical rather than online breach of the law is too high.
The other objection is the sloppiness of the drafting. Make it clear what is considered a serious offence under this law or it WILL get used on folks sharing torrents.
@h4rm0ny - so you missed the stuff about parliament saying no, 10 years is too long, we need an independent review, and the review saying no, it's too long, and more than 2/3rds of all the responses for comment saying that it was too long also? What do we get after all of that? They try to push through 10 years again.
As with most badly framed and thought out laws, what starts out as something targeted at serious crime ends up being used against people committing low level indiscretions. Trying to create laws based on the cries of a special interest group, without taking in to account public opinion and without proper attention to its effects is bound to end in tears.
@AC - because semi trucks with failed brakes are something we have to contend with on a daily basis? People made that kind of stupid edge case when it came to seatbelt laws also. You stand a much better chance of survival in an accident with a belt than without. There is a small chance that the belt will hinder rather than help, but this is outweighed considerably by the chance of it helping. The same applies here.
The US has the highest per-capita prison population in the world, and you want to make it bigger? I'd suggest that as a policy it's not working, nor are US prisons cosy hotels that no one would want to leave.
Part of the objective of the system should be to rehabilitate offenders, and if you give them no incentive to behave and improve themselves then that isn't going to happen.
You need a good deal more skill, time, effort and space to use a hobbyist grade lathe or mill than you do a 3D printer, and you'll pay a lot more for your feedstock (there's a lot of wastage in the machining process).
What may make a difference is the expiration of the key patents for Selective Laser Sintering type systems. They allow the creation of parts from powdered materials including plastics and metals, and once out of patent the machine costs should fall.
So you're arguing over the semantics of the word "copy"? Simply put a copy is the same or close to the original. MacOS was radically different from Xerox STAR (not least of which because it wasn't a pre-emptive multi-tasking system), and was the fruit of a lot of research by Apple into human/computer interaction and GUI design. Apple literally wrote the book on how to do much of that.
The best you could argue is that SOME ideas were copied from STAR, and that Xerox were paid for their use. To use the hated car analogy, at most Apple bought a STAR chassis, then created their own engine and bodywork to produce something quite different.
Microsoft likewise copied STAR concepts for Windows 1.0, but for Windows 2.0 they copied ideas from Apple without payment or permission (and if you think Windows = Finder then you're an idiot). As I said previously, it was those parts that Microsoft lifted (like redrawing uncovered sections of a window) that caused Apple to sue them.
@Pompus git - Firstly "taking an idea and running with it" does not imply theft, it implies that it is built on and improved from the original. Secondly you implied that Xerox's ideas became unavailable to researchers once Apple used them. Untrue. The original ideas were implemented by Microsoft in Windows 1.0 and that wasn't a problem with Apple. Windows 2.0 used Apple's extended ideas, which is where lawsuits started (and were ultimately a failure because copyright wasn't strong enough protection).
"So when Apple were running away with PARC's windows, icons, menus and pointing device ideas, they became completely unavailable to the researchers any more because they weren't copies. How does that work? Curious minds and all that."
Firstly Xerox were actually paid for Apple seeing and making use of the ideas from PARC. Secondly Apple didn't claim ownership of the Xerox ideas. What they claimed were the ideas they had come up with themselves in the approximately 5 years between them seeing the PARC system and the Macintosh. Things like repainting uncovered portions of windows automatically.
See http://www.catb.org/esr/writings/taouu/html/ch02s05.html for some of the history.
@Pompus Git - Um, no. Take a look at the original PARC/STAR GUI and compare it to that of the Lisa/Mac. The Apple version is obviously not a copy. They took some of the ideas and ran with them. Others they forgot about, yet more were completely new to the Mac. They even back-ported some of the ideas in the Mac to the Lisa. The metaphors of a GUI settled down and standardised in large part due to the manuals that Apple wrote on how to design a good GUI.
The early Oric 1 was horribly bug ridden. Printer output was corrupted by the key scanner routine for example. The keyboard was also unpleasant to use. They fixed both of those things with the Oric Atmos (and if you knew what you were doing you could burn a copy of the Atmos ROMs and drop them in to the 1) but it was too little too late. At that point it was the lack of software support that killed it.
That seems to be the general problem. Getting them to nail down a requirement spec and then stick to it for at least as long as it takes to create a system which matches it. Politicians seem to feel a compulsive need to tinker with requirements while development is in progress.
The ratio of adverts to content has to be acceptable to the viewers, and they don't have to be forced to watch them.
Until web sites figure that out (are you listening El Reg) there are many ad blockers on the market, at least some of which block ads completely. If users chose to accept whitelisted adds then they have at least some assurance that the advertiser has agreed to conditions based on that whitelisting.
Nowhere near that simple.
The safe company has developed a safe that is guaranteed uncrackable, thus making it a valuable item to potential safe owners. The FBI is demanding that the safe company develop a completely new tool to open a safe of this type that was owned by a criminal (which may, but probably doesn't, contain information about his crimes). They've also managed to reset the combination to a random number they don't know. They pinky swear that the tool will be used only this once. Meanwhile they have a queue of other safes that they also want opening, as do the police and foreign governments. Due to the rules of evidence the safe company will also have to document how the tool is created and how it works. Result: no one with that type of safe can rest easy, and the FBI can apply the case law to other manufacturers also so there will be no such thing as a secure safe. The methods will leak out and burglars will soon be able to open them also.
If a company has spent a billion dollars bringing a drug to market they wouldn't have done so without there being a large target market. Having discovered a drug and got it through approval they've spent a lot of money. They want to make that back. Just sitting on it gets them no return. Worst case they can license the production to another company for specific uses and get SOMETHING back. At the end of 17 years it becomes generic and anyone can make it (quite how the FDA seems to have screwed up generics is beyond me). After the approvals process that's more like 10 years.
You seem to be confused over the law. The FBI are trying to extend the reach of the All Writs act in order to force Apple to do something they don't want to do. Apple don't want to do it because it's bad precedent (no matter what the FBI would have you believe, having done this once there'd be a queue of government agencies and police forces demanding that they do it again). It would also be next to impossible to refuse the same aid to foreign governments like the Chinese. Apple aren't asking to be exempted from the law, they are asking that there be limits placed on law to prevent this kind of request being made of ANY company.
That "nice guy" bought the rights for a drug that is out of patent. There is nothing stopping another company from making a generic except for the FDA.
There are many diseases around the world that aren't worth researching a cure for (last time I looked it costs about $1bn to bring a new drug to market). If governments think that it has social value to create a cure then they need to put their hands in their pockets, not expect the drug companies to spend that kind of cash out of charity.
People risking their lives to stop criminals have to follow the law in the same way that we do. The problem is that the FBI is trying to redefine the law to let them do whatever they want.
With enigma the problem was a combination of a mathematical weakness in the algorithm, combined with known text, that allowed the Bletchley Park code breakers to very much reduce the search space. AES on the other hand has been the product of multiple cryptographers and much analysis to look for weaknesses, none practical having been found.
The Register 
Biting the hand that feeds IT
Copyright. All rights reserved © 1998–2024
