Piracy tools are a legal right

SUBJECT: 'PIRACY Tools'

"As I said in the original answer, professional thieves will still find tools to copy stuff but that isn't a reason to let piracy tools be sold in your local supermarket. Unless maybe you think hacking tools and spamming kits should be sold there too?"

What amuses me is that you calling them 'piracy tools' is simply a straw man which distracts from the issue at hand.

Pre-WIPO/DMCA the general public had every legal right in most countries to make copies of VHS/DVD/CD/TAPE/LP for personal use. The very fact that we now need some special product because manufacturers are reluctant and or have been prevented from providing a modern consumer recordable technology is in my opinion is quite odd.

Why DVD suddenly requires this special attention is beyond me. And before might suggest that consumers can simply copy their DVD to a current technology such as VHS, how does this protect the consumer, academic, or librarian of the future when these technologies are no longer available?

Mike

Arron replies

When DVD was first introduced there was no outcry about it being copy protected. Nobody complained that they couldn't "take backups." And people bought millions of DVDs. Nobody was complaining that their "fair use" rights were being infringed. For five of DVD's six years there were only players were available. Even the set-top DVD recorders now on the market can't make copies of commercial DVDs but nobody is campaigning against that.

Why is it that only now that piracy tools might become available there are suddenly people complaining about the copy protection? Why is it that only now that piracy tools might become available that "fair use" is suddenly so important?

Arron

Subject: P. R. Engel - closet AMD fan

I have been noticing the style of P. R. Engel, and I think he is a closet AMD fan.

The way he writes his pieces with oh so openly flawed reasonings. I think that he does it just to show how naive people have to be to think that Intel offers a superior product.

Look at the facts. He frequents a seemingly pro-AMD news site (no problem with news being pro-AMD; they are the enthusiast's chip!) and he sends flames that make us all laugh at Intel and their "loyal" fans.

Yeah, look at the pen name he uses it's pretty close to an anagram of Intel PR (come on, he only changed 2 letters!) and we all know how useless they are at hiding the truth from us.

P R Engel - You have been unmasked!

Hope this gave you a laugh. =)

Regards,
Richard.

The only excuse of publishing the letters of Mr. Paul R. Engel at Your site is the usage of him as the local entertainer (foolish clown) - I don't see any other explanation :-).

as for 'AMD vapour launch' - could You pls publish this link?

Sincerely Yours,
Michael Stoliarov

Early Nanotechnology was "ASBESTOS" which is a few nanometer diameter Alumina Silicate inorganic mineral fibers. Also carbon black in Xerox machines and laser printers.

Modestly telling.......

Dr. Colvin at Rice University's Nanotechnology Center, studies the safety risks of nanotech, says there is no evidence of safety concerns. She did not reply to this missive below, but did read it.

--

I noticed you [Dr. Colvin] testified we don't know about hazards of nanoscale particles. It is not worth exaggerating, but neither would someone experienced in industry avoid due caution associated with any finely divided nanoscale material.

Regrettably before the advent of nanotubes and buckyballs there were many examples of nanoscale structured materials present - in nature and manmade.

Example #1 is asbestos - its toxicity arises from the nanoscale fibers penetrating cell walls quite readily.

It did not result in recent nanotechnology science and it turned out to be a huge health and liability problem, probably slightly exaggerated and overlitigated but still a significant health concern.

Pretty obvious example. The aluminum silicates? comprising asbestos are relatively benign if not in nanoscale form and not breathed in or contacted directly in nanoscale form.

But in a form or environment where small particles are liberated and waft in the air and are breathed in, it is pretty obvious that there is a hazard. Just ask any of the firms which are defendants in the many cases related to asbestos, and the industrial toxicologists and cancer specialists who can describe all the gory details

Likewise well known with carbon materials, is that Xerox toners, which are carbon black, are associated with cancer incidences in maintenance technicians of copy equipment and others who work with carbon black, the main ingredient of toner.

It is plain that nanotubes will be no more benign than carbon black and might well be at least a little more hazardous because again the smaller size will under the "wrong" circumstances increase the possibility of breathing into the lungs, and initiating the same cancers that are observed caused by carbon black, since the smaller particles permit longer distances for the particles to be airborne, and greater risk of compromising the cell wall integrity correlated with cancer incidences.

Some food for thought.

I did evaluate the use of asbestos nanoscale fibers for some applications touted as requiring carbon nanotubes. Much cheaper and in several instances even very useful. In Quebec, where I am from, the mining of Asbestos is significantly less expensive than any price contemplated for nanotube production.

But there remains the toxicology challenges for which there exists no obvious rigorous data yet in nanotubes. This is not a reason to be Pollyanna about potential industrial and consumer safety hazards, just more reason to practice greater caution once outside the environment of small scale lab experiments.

You might mull this over, since there is more than strong anecdotal evidence that history will repeat itself with carbon nanotubes, if due care is not taken.

--

start calling it "nano-asbestos", and some will get the idea. I found out the dimensions of asbestos fiber diameters after I visited the Smalley Lab at Rice Univ to mount tiny multiwalled carbon nanotubes on Atomic Force Microscope probes ( nanostyluses ) and gave some thought to checking into asbestos. When I returned to Montreal ( home ) for a visit I spoke to some geologists at McGill and found in a book on silicate minerals that Asbestos was a NANO-fiber, far ahead in time of the discovery of carbon Nanotubes.

Name supplied

p.s. I have about 20 years working in microfabrication in the Integrated Circuits industry and 8 of these was working on nanotech sensors.

A word to the wise might be sufficient, but not always.

Paypal has now started emailing users who haven't yet signed up to the new "progressive" terms and conditions on a daily basis. Their original time limit to agree to the new terms has now expired, and they seem to have extended it by 55 days for the remaining users and are emailing them on every day (so far) to "remind" (spam?) them into submission.

I've had day 55, 54 and 53's emails so far. Could they possibly expect this tactic to work?!?

Amused,
Richard

Subject: AMD says no to thermal grease

OK thats stupid on AMD's part but the article focuses on what overclockers will think.

NEWS FLASH
Overclocking any hardware voids warranty. Dont think they will care.
Now for the regular everyday consumer who becomes unlucky and recieves a bad processor or processor goes bad and in the process of installation used thermal grease.... boy they going to be enraged. Probably not think twice bout jumping ship to chipzilla and drag all there friends with them.

Subject: Rambus article

As an investor for over 5 years, I am glad to see you and others finally opening your eyes to the injustices that have been done to Rambus by Jedec, Micron, Hynix, and Infineon. It has been a painful and very costly conspiracy for investors; I think it is a criminal act for big companies like this to conspire to ruin a company just because they have invented innovative and superior technology (at that time). Please keep digging; there is still a lot more of information to come about the injustice done to Rambus. The Motley Fool Rambus board continues to uncover, collect, and discuss some very compelling evidence. Thank you for being open to all the evidence. I think it is increasingly clear that the FTC has been wrongly influenced by the Micron connections to the FTC. The FTC is supposed to protect the public, not Micron. The FTC is out of control and is going out of its way to selectively destroy Rambus while ignoring evidence that supports Rambus' case. They are also holding Rambus to a higher standard than they are holding other Jedec members to. Micron and Infineon are doing a power play with this because they are threatened by the technology that Rambus invented.

Mike Danford

Subject: Teel on Rambus

Bill Teel's summary of the Rambus adventure totally ignores some of the basic facts and events of the Rambus escapade, and seems to be an attempt to insert a "grassy knoll" into PC history.

Bill didn't feel obligated to mention some of the background facts that led to the Rambus fiasco in the first place. Intel was under real pressure from the unexpected success of the AMD Athlon. When the Athlon breezed past the 1GHz barrier, Intel tried to crank the clock on the PIII one too many times, suffered great (and well-deserved) public humiliation accordingly, and set to work to bring their P4 design to market earlier than planned. [Even now, I'm not sure that Intel originally intended that the .18 P4 would ever see the light of day as a production chip.]

At the time Intel accelerated the P4 schedule, the Dramurai were in disarray (situation normal) and there was no assurance that DDR would be appearing anytime soon. This complicated Intel's life, because the P4 design needs significant memory bandwidth; P4 performance with PC133 was a joke - the poor cpu was starved for data. Intel's search for a faster memory led them to Rambus and their RDRAM, which, with sufficient influx of Intel cash, could be brought to market with the P4. The initial RDRAM offerings highlighted a distinct production problem in meeting the PC800 spec - PC600 was commonly supplied with any system which was not at the top of the speed range. This did not inspire a lot of buying - most customers do not buy the top-of-the-line, and the slower-clocked P4s with PC600 were abysmal performers. PC800, where supplied, was satisfactory as far as bandwidth was concerned, but came with an unexpected latency penalty which hurt the P4 in office apps - so the 1.5GHz P4 was barely quicker overall than the 1.2GHz Athlon/PC133.

[Bill's column does not mention the stock options granted by Rambus to Intel at the time of the choice of RDRAM by Intel - options which made it very much in Intel's interest - to the tune of multiple billions of dollars - for RDRAM to succeed as the new "industry standard" for PC memory.]

The Willamette/Rambus combination never took command of the PC marketplace. Intel's market share eroded throughout the Willy/RDRAM period, thanks to weak sales of the P4 lineup. (At the same time, the Celery was being clobbered by the Duron on both performance and price, which didn't help matters.) Despite Intel's disdain, DDR did finally appear in the market, and after several stumbles, worked well enough with Athlon platforms to maintain parity with the Willy/RDRAM combo, as both AMD and Intel 'wound the clock' on their processors.

I'm sure that if the Willy/RDRAM combo had ever shown a decisive peformance lead over the Athlon/DDR, sales of Willy/RDRAM would have soared, and more of the Dramurai would have started producing RDRAM. Instead, RDRAM showed little market growth, and rather than invest their own money to develop RDRAM production, the other Dramurai watched - and saw the continued absence of any boom in the RDRAM market. When Willy/DDR platforms were finally introduced, they were not that much slower than Willy/RDRAM - several of the non-Intel DDR boards performed on a par with the Intel RDRAM boards - and the DDR boards fairly quickly began to outsell the RDRAM boards in the marketplace, thanks in no small part to the cheaper memory and the fact that DDR became standard in mid-range P4 PCs from folk such as Dell.

Some words on memory prices, the first two not mentioned by Bill:

(1) RDRAM chips are larger than SDRAM/DDR chips for any given process. All of us know what that means - bigger chips are more expensive chips, so RDRAM chips will always be more expensive than DDR chips when built on the same process - a fact of silicon life. [At one point, Samsung held a press conference to proclaim that they had beat the price barrier between RDRAM and DDR, and were producing RDRAM chips as cheaply as they produced DDR chips - but a quick look at the details of their presentation revealed that they were comparing the prices of RDRAM chips produced on a .15 process with the prices of DDR chips produced on a .18 process - the "price parity" thus proclaimed by Samsung would suggest that on the same process, RDRAM chips would be about 44% more expensive to make than DDR chips.]

(2) For the first few years of RDRAM production, some individual chip tests could not be performed on individual RDRAM chips, but could only be performed on the chips after they were assembled onto DIMMs - a condition which would either increase rework costs or significantly decrease effective yields - another fact of silicon life which added to the cost differential between DDR and RDRAM. [I suspect, but do not know, that this problem was solved by the development of some very specialized test equipment - another investment cost.]

(3) In addition to the added production costs associated with the RDRAM design, there was/is the cost of the royalty payment to Rambus.

Given these 3 cost factors, it was and is a fact that RDRAM could never be produced or sold as cheaply as DDR, a fact which Rambus spent several years pointlessly denying with their claims of "price parity " (the projected 'time to parity' bounced around, from presentation to presentation), complete with accompanying bogus graphs. This continued obvious lie was one factor in the decline of their credibility.

Bill's contention that the spirit of brotherhood among the Dramurai is strong enough for all of them to act in concert - on virtually anything - is amusing. The idea that they could (or would have the brass to even attempt to) act in concert to "strong-arm" Intel is hilarious - each of the Dramurai would be too worried that one of the others was talking to Intel in secret. I would suggest that the only "meaning" associated with Intel's dropping of RDRAM chipsets is that sales were flagging - most customers just wouldn't pay the extra money for what they perceive as a negligible performance gain. [In recent times, Intel has never really been a 'niche' supplier - their business model is to make money by producing and selling huge numbers of identical parts, and their facilities are well designed and built for that very purpose.]

I suppose Bill's stock in Rambus can excuse some of the "Rah, rah, SiS boom bah" in his letter/ column when he tells us how wonderful some of Rambus's future designs will be. Rather than take his word for it, though, I think I'll wait to see how those future products perform - and how much they cost - compared to DDR-II, or whatever the "industry standard" is at that point.

John

Subject: Benefits of Opteron aside from >4GB

I think everyone who's said that we're not gonna need or gonna need 64-bits anytime soon is missing an important point. Anyone who's done Assembly Programming would know what a pain the x86 architecture's registers are to work with. By having just a few registers you have to do a lot of memory accesses, which slow things down a lot. BUT... the Opteron significantly increases the number of registers to 16 x 64-bit registers. I'm not sure, but I think I've read somewhere that these registers can also be used as 32-bit registers, much like the way you could use AX as AH and AL. That would result in 32 x 32-bit registers, a 4-fold increase from the x86-32's 8 x 32-bit registers. With all these registers memory accesses can be reduced greatly, thereby speeding up performance. This is the most important benefit as of this time.

And about the 4GB thing everyone's been talking about, it's gonna be quite a while before home users need anything beyond the 32-bit address space (of course, pro users are an entirely different matter). Given that most home users today use 512MB (that's using 29- out of 32-bits), and my estimate that folks upgrade RAM by doubling it (256MB to 512MB, 512MB to 1GB, 1GB to 2GB, etc.) every 1.5 to 2 years, it's gonna take +/- 4.5 to 6 years before we reach the 32-bit barrier and begin to use 64-bit addressing. 4.5 to 6 years gives AMD enough time to credibly establish AMD64. Their timing couldn't be better.

yours,
Ronald

Subject: the Opteron's bus

I'm just writing to ask about the Opteron and it's FSB. AMD is, and has been using the Alpha EV7 bus, right? That bus is been in use for a long while because it had the ability to scale up to ~400mhz. Now we have Bartons with 400 mhz (200 mhz but DDR, same for Opteron) busses, and the Opteron, I think, comes with a 400mhz bus.

What I don't know is, if the 400 mhz was DDR, or if AMD still has headroom in the FSB to compete with Intel, so that they could scale it up another 200 mhz to be 400 mhz non DDR or 800 mhz effectively through DDR. If AMD has finally maxed out their EV7 bus, how are they going to compete by the end of the year in their server chips? Intel could, I would imagine, make a update to the Xeon to have its 400/800 mhz bus if they wanted to, if they have not already.

What will AMD do once the current bus is maxed out? Does AMD even realize this? What will they have to do next? Design their own, license another Alpha bus (if there is one?) This might be something to start asking AMD.

I am a fanboy for AMD, so as much as I am excited over the whole AMD64 tech, I wonder what they will do in the future, because the FSB is one of the bottlenecks being worked on now. I realize they have the ondie memory controller, but what about the a dual channel bus? that has to use a mobo's controller at the moment, which would be at max 200/400 if not overclocked, and what IT admin will want to overclock his servers?

Email address supplied

Subject: Itanic's big caches

It seems only fair to wonder if the Itanium's 64bit competitors will follow suit with the huge caches to get bigger performance.

Not immediately, because the Itanic needed them more. Its code density is notoriously worse than its competitors'. And it already had unused load/store bandwidth, so the cache change didn't need matching rework of the core.

Cache coherent multiprocessors have the interesting property that the _collective_ cache is available to each CPU. (Not as quickly as local cache, but quicker than main memory.) In well made CC MPs, like the Power 4, there's a significant speed benefit. It would be an uncommon workload that didn't fit in the collective caches of a 32-way Power4, but _did_ fit after they doubled things. Unless, of course, it was something that had to be simultaneously in _every_ local cache - like code!

Don

"At that point, the PR side of the business stepped in, a gag was placed around this fellow's mouth, and someone put a striped square thing filled will yellow goo into my hand, as I was led off."

I'm a little slow. What was this thing?

Thanks,
Dave Feustel

All members of our department have been racking our brains to understand what phrase or saying this refers to, or what it means. Can you elaborate? The best we can come up with is a griddled toasted cheese sandwich or an latticed custard danish, but square. What phase or saying this refers to is still a mystery.

Many Thanks
Richard Phillips

[none of us knows either. Ed.]

Re: "Windows Server 2003 won't run major Microsoft apps"

Microsoft obsoleting older versions of software such as SQL Server and Exchange makes more sense when you look at it from the point of view of their recent acquisition of Connectix's virtual machine technology. First they release Windows Server 2003, and make sure none of the old software runs on it, with "security" being the excuse. They'll get some new deployments, but upgrades from Server 2000 aren't likely, which is why they keep comparing it to NT 4.0. They want to get those people off (and since they are cutting off support for it, they'll succeed with a lot of them) That'll guarantee that everyone coming off NT 4.0 upgrades to the very latest version of their software. Well, except for those who switch to Linux, but as hard as upgrading to Windows Server 2003 and the latest versions of SQL Server and Exchange will be, upgrading to Linux with Oracle and Samsung Contact is probably even more challenging. Even if they may happen to have licenses for stuff that's got a couple years of support left in it won't be usable for them. Their fault for trusting Microsoft.

Once the NT 4.0 drop dead dates passes, look for them to stick the Connectix virtual machine technology into an service pack of Windows Server 2003 to give people the ability to emulate Windows 2000 (but not NT 4.0, they'll give some excuse why that's not doable) so those older versions of software people are running on Windows Server 2000 can continue to run until they expire. That'll be a good move for them because a lot of people who might have stuck with Windows Server 2000 until support expired will upgrade, perhaps to the very expensive Datacenter version to consolidate smaller servers (where it won't be an upgrade, but a new purchase at full price)

This whole strategy is probably a test run of what they want to do when they introduce Longhorn with Palladium technology, since they'll want get everyone on that and lock up their data in proprietary format with DRM, thus guaranteeing they can't ever consider Linux or any other alternative. They'll follow the DOJ settlement and give access to the file format, but the out they got for security in that settlement means it'll be useless without getting through the DRM. So nothing will be able to get to a Microsoft customer's data at this point except Microsoft software. And by providing their own virtual machine technology, they'll kill the market for virtual machine technology on Windows. VMware will survive for virtual machine technology on Linux, but in the long run I fear it'll become useless of running Microsoft software because they rely on cooperation with Microsoft to make it all work. Cooperation that'll surely disappear when they introduce their own software (if it hasn't already)

Email address supplied

Subject: I really like the way you keep the flames intact...

I don't know if I could say that I'm impressed enough. The way you guys keep the flames in one piece without editing (even the hilarious albeit rude insults) amazes me. Keep it up! :)

Tamer Metwalli

Subject: Re: Alpha technology could be in 32-bit Itanium kludge

Or for an alternative theory Intel have done some kind of deal with Transitive.

I've seen their technology in action and it's very impressive. In some specialised cases the translated code can actually run faster than the original.

The problem with both solutions is that they are operating system specific. As opposed to something like Transmeta's code morphing technology which is chip level and OS neutral.

p

ANY READERS RECALL FX!32, the fab 32-bit extension developed that allows 32-bit X86 Windows programmes to run at high high speeds on the Alpha and Windows NT?

Yes, I do remember FX!32 and hope you were being facetious when you described it as running at "high high speeds". Probably most of your readers have never used the product and may not catch the joke. My experience at the time was that windows programs ran on a 166Mhz Alpha slower than they did on a 90 Mhz Pentium. Assuming that said programs didn't require any special drivers, in which case they wouldn't work at all.

FX!32 was an impressive piece of software that made no sense to use - why spend a fortune on an Alpha and run slowly under FX!32 when you could instead buy a cheaper Pentium and run faster natively? I suspect that the FX!32 clone on Itanium will suffer from the same ecomonics.

Regards,
David Mathog

Subject: WinXP Right Click Gobbles Up CPU Time

I think that this is terrible. M$ should be ashamed of themselves. And then refusing to fix it? What kind of BS is that? Right clicking is an efficient way of managing your files, no matter what it is that you are doing. I think that that could cause problems in some areas of computing, as well. Imagine you are burning a CD and you right click on a file and the lack of CPU cycles to the program produces you a coaster? (THank God for BURN-PROOF!) Anyway, I think that this is just another sad example of why M$ is detested by so many. (Betcha a dollar that it isn't fixed in Longhorn ;) )

Jason Trivilino

Subject: Windows XP right click gobbles up CPU time

I've noted this one ages ago. Very annoying as it even chews so much CPU time that media player stops playing.

Anyway, the problem only occurs if you have "Show common tasks in folders" selected on the General tab of Folder options. If switch that option to "Use Windows Classic Folders" the problem goes away. It's probably just CPU time take by extracting info from the file to fill out all that bullshit on the left hand common tasks pane. It's a waste of space IMO, so I just turn it off and never get the problem as a result.

HTH,
David

Subject: RE: Windows XP right click gobbles up CPU time

In a very unscientific test, I opened up Asus Probe temp moniter, opened up task manager, and right-clicked on a drive in My Computer, sure enough the CPU usage went to 100%. But at the start my CPU temp was at 40 degrees, after 5 minutes or so it was at 45 degrees. Hmm.

Regards,
Huw Davies

Subject: Via's Nehemiah SMP? Hogwash!

I like Dave Jones, he is a nice Linux kernel hacker. But he probably wouldn´t suspect that quotes from his diary would end up as news at the Inq.

There is a lot more to making a CPU SMP-compatible than Dave´s diary mentions. VIA would have to redesign the caches: the present cache design is not coherent i.e. not usable in SMP configurations.

Then additionally VIA would have to find a market for a processor that even in dual or quad SMP configuration, would still be slower than a single P4.

I also doubt very much VIA would dedicate part of its development team to any low-volume, specialty product, let alone something as complex as a CPU + cache + bus interface redesign. They are smarter than that.

SiS bought the Rise CPU core IP but we have yet to see a usable product from them. They are probably busy churning out fast, cost-efficient 65x P4 chipsets, and making a lot of money in the process. Again, money-wise smart, out of necessity.

Email address supplied

Subject: RIAA media reports are misleading !

The story on "RIAA file share ruling" demonstrates a number of incorrect conclusions by the judge and some in the media. First off if software can only be used for illegally duplication of copyright protected software then the software producers and sellers are co-conspirators rather they want to admit it or not. If the software has other legitimate uses then there should be a means to block it's use for all illegal copy activity. While use of file duplicating software may be in some ways similar to buying a conventional video camera or VCR it is NOT the same, as these devices are usually purchased to play or record a video. To duplicate a video you need two machines or a dedicated machine for the purpose of duplication. Most VCRs and video cameras do not have duplication features, they have only record or playback features. Yes I am aware VCR's are often attached to TV's and illegally used to record copyright protected material, but consumers who do this are fully aware that they are in violation of law and they should be prosecuted whenever possible to deter this practice.

I have seen a rash of headlines on this judicial ruling that state that the judge has ruled that "file sharing is legal". This is NOT what the judge ruled at all. He ruled that the software company who was the defendant in this case, was not legally responsible for the illegal file sharing of it's customers. This in no way makes file sharing legal. In addition as stated above, a software mfg. has a legal obligation (which the judge failed to acknowledge) to design their software so that it can't be used to duplicate copyright protected materials. Perhaps all future video and audio recordings will need to be distributed with an "effective" copyright code that the machines and software read to prevent duplication. While not a total solution, it's a start. There's no denying that the motivation for producing bogus duplication software is financial gain. And when money is involved, large quantities of money in particular, many people can "rationalize" theft of intellectual property even when they know it is illegal. Theft is theft, plain and simple. Until society comes to grips with the reality they should not have been making illegal audio and video tapes twenty years ago, they won't accept that they are criminals for duplicating current music or video files, which they are. Denial is not an acceptable or practical legal defense and eventually the courts will emphasize this point. It will be a long and painful journey but the law will be enforced and people had better learn to accept reality.

Even though this court ruling may be poorly conceived it is NOT an endorsement of file sharing. The ruling sets a bad precedent as the judge is confused on technology aspects. The RIAA will have no choice but to sue individuals if this court ruling is upheld, which is unlikely as the judge missed too many important aspects of the case and technology. In the meantime the RIAA should file appropriate suits against any individual in violation of the law to send a clear message that illegal file sharing of copyright protected material is wrong and illegal and there is a high price to pay for theft of intellectual property.

Randy Hubbard

Subject: Opteron/Athlon64

I was interested to see just what results Anand might get - and what they might mean - with his attempt to compare an Opteron with desktop 32-bit CPUs. Frankly, I was *amazed* that the 1.8GHz Opteron in a server board swept all of the game benches, winning every game benchmark over a faster-clocked Barton *and*, much more remarkably, over the 3.0GHz/800MHz Woodie/Canterwood.

There's still a mystery afoot which won't be resolved until Athlon64 actually arrives. The Opteron uses a 144-bit wide memory interface, while the Athlon64 mem interface will only be 64 bits wide, so RAM access may be slower. Offsetting that is the Opteron's use of ECC-only memory and the server design (stability above any performance consideration) of the current Opteron board, so surely the Athlon64 will show considerable improvement in tests where membus latency is important - which it typically is, in "office" apps, and in those scientific apps that are heavily weighted with BS&I calculations. [BS&I = Brute Strength and Ignorance]

Anyway, it seems that the Hammers (oops! - naughty word) will be hella gaming machines as well as office/scientific machines. (Winning over the gaming community will be a *huge* victory for AMD.) Bring on the 'desktop' nF3, and wrestle it against the Via/AMD duallie boards - it'll give us something to look at while we're waiting for Athlon64 - and some of us might not wait, after all.

John

Subject: MacOS X continues Crapintosh tradition....

Thought you might like to arc that plasma torch flame up a bit and get some comment from those rabid Mac users.

Over the years we've attended onsite for many of our Mac user clients for the "flashing question mark" of death. Usually they're printing via regular MacOS with Print Monitor and the machine crashes. Somehow part of the HD gets corrupt and the machine won't boot.

For years Apple have not thought it fit to provide a decent disk recovery tool as "Disk First Aid" can not repair these problems. "Disk First Aid found problems with this volume, but can not repair them!". Panic usually sets in for these users. It is then necessary to use the excellent Norton Utilities for Mac.

Happening across the same problem today and knowing the client had dual hard disk drives, one with MacOS X 10.2.5 and the other with MacOS 9.2.2 we had the client boot to MacOS X to run Disk Utility. One tab in this tool is the "new" Disk First Aid. Sure enough, this "First Aid" tool could not repair the drive!!

Thinking that perhaps the Unix "fsck" would do a good job, we talked the client through running that on the drive in question. The whole machine locked up. Try again - same again. Apple support's response? "fsck is not supported on non-UnixFS disks. Your client should have a professional disk recovery tool".

Giving up, one of our staff went up there, ran Norton Utilities and the problem was fixed in 5 minutes.

We sell/support Windows machines too, and I can not recall a hard disk drive that was at least spinning which we could not "recover" with the OS built-in tools. Can you say "product quality" Steve Jobs? Crashintosh, or less endearingly "F#$kintosh".

Email address supplied

µ