Despite the rather black box connotations of reverse engineering, it is a legitamate R&D exercise. Car manufacturers reverse engineer their competitors, chefs try to dissect recipes, etc. In the computer context according to the Centre for Software Maintenance,
Reverse Engineering is the process of analysis of an existing software system to create representations of a different form or higher level of abstraction.
and
Reengineering is the process of analysis and modification of an existing software system to reconstitute it in a new improved form.
Given the preenial occupation of engineers is to make things better, faster, or cheaper, tinkering with the electronic toys or source code is a natural pasttime. It is only the marketeers and financial managers that want things to be "hidden" so that the cost (and thus in their mind = value) is higher (basic economics, remember scarcity == higher price). Obfuscation of code is an obvious mechanism to exclude competitors, however, it significantly adds to the long-term cost of maintenance and also reduces the potential market. How many times have you've been given a piece of code with the design specs/architecture residing in someone's head who've just left? Wouldn't it be nice if some intelligent bit of software did the analysis and gave you the answer (yeah, wishful thinking but still....).
Perhaps people don't realise it but there are 2 information monopolies, one when one party controls everything and there is no alternative, the other when everything is freely available so that there is no competition (and thus no alternative).
LL
Uncoordinated driver having hard time learning
on
CNN Installs Linux
·
· Score: 5
These god-fangled Model-Ts may be Ford's hotest piece of metal but are they the car of the people?
Would you recognise a z-crank if you tripped over one?
My month-long effort to be cool and try and move from point A to point B without crashing at point C,D,E,F gives me chilling flashbacks to learning how to ride a horse without having people laugh at me. For now, I'm glad to have ol' Silver back with me to rely on.
The good ol' boys warned me beforehand. They had trying finding the pedals and their friends throw bits of metal around on that silly assembly line thingy. I'm car-literate since I've actually washed an early model before becoming a news flack. We're in the 90s but I'm a bit clueless about all the moving odd bobs inside.
My advice is if you've barely mastered whip-cracking and never changed a tire, stick with the ol' feller or take the train. If you know an oil change from elbow grease, perhaps you'd even find this fun (gasp of shock and horror).
In truth, stalling made the task of backing out the driveway wasted the whole day, wasting at least 3 hours figuring out those crazy levers. In the end, with no decent horse sense and as chummy with a wheel as I was with phys-ed torture sessions, you know, the one where the teach' nursemaided those skiiny geeks and prevented us jocks from having fun.
I opened the engine and figured out the hole for that metal thingy you call the crank handle, right? I try to yank the gear as written on that flimsy bit of paper they call instructions manual but those gunshot backfires nearly gave me a heart attack. Not a good omen.
I turn to the "Cleaning Engine" page in teh manual but couldn't figure out that picture with all the bits and pieces meant. The next page says something about changing oil but all I want to do is drive this silly hunk of junk, not build it. Opening up the engine didn't show any spots to hook the reins.
I turn the handle but think something's wrong as it was making all these funny noises like my horse has colic or something. I press the pedal and nearly wiped out my favourite mail box. At least a horse is smart enough to avoid impaling itself.
Well, it least it seems to move but don't know why I have to keep looking over my shoulder to see where I'm going all the time. I survive smashing into the barn door but the manual warns against driving without flags and horn blowing. I get outright dirty trying to count the cylinders and rpms as suggested, taking half an hour to motor 50 feet back to the house.It then dies for some strange unknown reason so I know it's time for lunch.
After a nice big juicy steak, I give the od' editor a hollar but he mentions something about gas fill-up and to check the tank (as if I'd containminate my washing water!). After a while, a neighbour drives by so I swallow my pride (yes siree, gave me indigestion for resta the day) and asks what a gas tank looks like. He mocks me "And you're learning how to drive?". Next time he gives any mouth, I've gotta shotgun handy. He helps me fill it up a bit (at least I guess which is the right hole) but then some shit musta hit the fanbelt so I drop this project for a couple of hours.
At this point, I'm burnt out. I start wildly guessing buttons to push. God must have been with me as it started again (must remember to buy a new can for that mailbox). I can make it to the barn (and even stop!) without hitting anything too important. That's good enough for me.
Then I head home, bleary-eyed. My superficial knowledge of gasoline engines made this project frustrating. I would have helped if I were a mechanic. On the other hand, I'm surprised I got as far as I did, just like trying to connect with a baseball bat. It was mildly fulfilling in a mysterious way. I may have no idea how to use this car but I got my hands dirty.
{Parody mode off}
I sure would go a lot better when someone invents the automatic GUI transmission! No disrespect to the poor guy but unfortunately it is relatively early days and the only way to learn is to be willing and get those hands dirty. Congrats on making a start and a warning to the rest of the Linux mechanics that exotic details of kernel file spaces is as relevant as quantum physics to the average driver.
Perhaps E-Bay is trying to eliminate future corruption of their customer profile?
If you read the marketing-speil coming from all the current e-commerce sites, be very afraid. They want to know how who, what why, where, when and what you spend. Now if a robot that emulates a human search (or aggregate metasearches) then GIGO (Garbage In, Garbage Out) rules apply. For people who work in the database field, the technology is the least of the worries as cleaning up the data is probably 80% of the work.
similar analogies would be - creating artifical users on Hotmail (their name space is NOT infinite) - inserting mega dynamic indexes to Yahoo (overwhelm the human classifiers) - creating recursive reviews to Amazon (drowning the human comments with noise)
In short, anything automated can be screwed up to some degree. If the script kiddies ever get a clue, I suspect there will be some sort of arms race between robot automation and exclusion. Fun and games.
Hmmmm..... looks like some hardware assist to help binary retargetting. For people not familiar with the concept, take a look at an overview. The concept is sound in that as ESR points out 95% of the programming jobs out there are spent in maintaining old code on old machines. However, if there was a way of abstracting and specifying the hardware characteristics and mapping from one to another, then old binaries could be shifted onto newer and cheaper hardware with less hassle. I can think of cases like old Cray binaries where porting them to a new MPP would be too painful manually, some of those timing cases can be really subtle. Given that computer companies are very relunctant to support hardware which isn't current (ie not profitable) and others could potentially go belly-up (correct me if I'm wrong, I think only IBM is one of the few giants left from the 60's), there is a need to protect the million of man-years spent on specific packages. Of course, research has shown that retargetting works better with availability to the original compiler source:-).
Given the rate of corporate take-overs, you could quite easily end up running a zillion different systems and lose valuable time in trying to consolidating everything.
Oh well, add this to the speculation pile along with everyone else.
In your paper The Magic Cauldron you talk about the sale value (final cost to consumer) and use value (what economics would consider a capital good). Modern capitalist societies have evolved very complex and sophisticated instruments (a la the profit motive) to price these goods and signal to the market what is valued. With OpenSource, this pricing information is missing and thus time/effort is spent on "sexy" projects like 3D interfaces (what economists call malinvestments) instead of really important stuff like good optimising compilers.
Question 1) Pricing of OpenSource Software How can OpenSource software be fairly priced given that it is always possible to undercut a distributor?
Question 2) Distribution of Resources Instead of vertically integrating all the profits at the sale end (distributions like Red Hat), how can the creators of the intermediate goods get enough funding to continue refining their products?
Question 3) Scaling to Megaprojects Given the limitations of no capital pool of funding so that intermediate software can maintained, will OpenSource projects be limited to "small" projects that can be supported by 1-6 key designers and wouldn't this be an inherent constraint?
Question 4) Bazaar Rules of Conduct At the moment, the Software Bazaar is controlled by gentleman rules of conduct (no encroaching on projects, equal sharing, etc). Do you see this continuing with the increasing commercialisation (and potential moral corruption) of the hacker's "gift culture"?
Question 5) Software Patent Roadblocks In a situation where time-to-market is becoming a key factor in dominating the bulk of the profits (see some notes on game theory of software patents), how can OpenSource work around limitations of key locks on irreplaceable algorithms?
Question 6) Freedom to Copy. While big companies like SGI have the resources and name-brand equity to release and protect their OpenSource efforts, how will smaller entry level players survive long enough without their ideas being poached by bigger companies?
I'd hate to rain on anyone's parade but wouldn't this military wet dream be superfluous? Let's suppose I'm one of these countries with tac-nukes. Why would I bother announcing the launch site and invite retaliation by using balistic missiles? Better still, just to ship it into a anonymous cargo hauler and detonate it within some strategic harbor or even Panama canal. That way the source (assuming you can disguise the origin of the manufactured weapon) can be anonymous. Given the gung-ho way the US been acting around the world in the last few decades, I'm sure there's no shortage of splinter or fanatic groups to spread the net of suspicion. If people are interested in the military mindset, take a look at their parameters magazine, in particular the article by Peters on "Our New Old Enemies". Very interesting.
People don't go to war for no good reason. If you create a threat, then people will respond in kind. Defining enemies through an arms race might be good for the military-industrial complex (correct me if I'm wrong... I believe US and Britain are still the largest exporters of arms) but does little to create long-term goodwill. Exporting organised violence seems to be a self-fufilling prophency as it propagates a climate of fear. Afterall, if you think someone is an enemy. then what are the chances that every action you perceive is hostile? Psychopaths are not the only people with a warped mind-view, a entire culture can be infected in rather subtle but destructive ways (Andy Grove "Only the Paranoid Surive", Bill Gates "Technogy is great, but 90% market share is better"). Very successful but at what cost?
This century has seen 2 world wars, numerous regional conflicts and ongoing bushfires. I would hope the next century has a better record.
belswick wrote Note that SGI is showing all the signs of entering the death throes stage. Another 30% of the workforce laid off, abandoning major initiatives, CEO bailing (to MS!!), loss of faith by major customers.
Unless you've got inside information (which the SEC would be very interested in hearing about), I think the slashdot audience would appreciate more evidence than mindless parrotting of popular press. For your information, they are spinning off several portions of their divisions into separate business entities. Now while some people may consider this akin to kicking fledgings out the nest, the rate of turnover in Silcon Valley is such that the difference between working for one company vs another is just which branded T-shirt you wear. Think of it as a beehive with clumps forming and dispersing to form interesting new combinations. Abandoning major initiatives?, how many announcements have you've heard from major companies that have died the silent death of being irrelevant to real needs.
As for the CEO, well, I'm sure there will be some interesting books a few years down the track but for many hard-core SGI purchasers, the shift into Intel consumerism where they did not have any competitive advantaged showed some very wooly thinking (for the cognosti, there is nothing technically inferior about the MIPS architecture). The loss of customers is not surprising considering that many applications that used to be top-end in the 70s can now run on a single modern processor and big cache (the refuge of the lazy microarchitect). Getting a free ride from Moore's Law is not the same as coming up with innovative new software applciations that can really take advantage of increased CPU capacity (apart from molecular simulations which will chew up any CPU cycle you throw at them).
Customers will buy SGI equipment if SGI can show they offer a value proposition that is worth the premium over mainstream machines, whether it is memory latency, quality engineering, coolness factor or whateever, people will buy (oh and getting their manufacturing/distribution process to be more efficient would help a lot). Computers are becoming so prevalent that the only distinguishing feature nowadays for PCs is image and lifestyle (does the color clash with the decor:-) ).
Reasonable people must expect that SGI goes Chapter 11 RSN (barring a government bailout) and then what happens to people who need supercomputers? Would you say Apple devotees are unreasonable? Don't you understand that given a planet of 5 billion odd people, not everyone is interested in the toys you are? Cries of doom and gloom have always been around in any industry in one form or another as it gives paper pushers a reason to justify their existance instead of getting their hands dirty coding or designing. You have to realise that SGI serves a fairly specialised market (data intensive, high-end graphics, scientific back-end grunt machines) in the 50K-50M range. Much like Porsche and BMW cater towards a cliental that wants absolute performance and not cheap consumer junk (admitedly the Japanese have given the US auto industry a shot in the arm since 80s), there will always be people who appreciate the qualities that SGI offers. Provided SGI can continue to support those companies and not go around trying to push Porshes for people wanting bicyles (amazing how hype can convince people they need a Pentium III to browse the web) at an affordable price, they will survive.
If you work for a company, you'd realise that the first law is survival which is depedent on their market relevance. SGI will continue so long as their is a demand for their expertise as priced compared with other market alternatives.
As Matt Welsh noted, it is not exactly a trivial problem. If you look very closely at the article, the LCC wants to occupy a happy ground between the share-nothing crowd (Microsoft, Tandem) and the share-everything (Oracle). The share nothing pardigm is rather simplistic in its approach and reflects the fact that throwing together a bunch of machines with a cheap interconnect is a comparatively straight-forward re-engineering approach. The share-everything come froms the extension of shared-bus architectures (e.g. Sun Starfire) which enforces a multiple lock strategy. Companies like SGI have thrown million of R&D dollars into the middle-ground which is why their cc-NUMA architecture and cellular IRIX is quite popular. I wish the LCC luck but there is a reason why a successful working solution is expensive as it requires a savvy combination of hardware+software+smart routing (the SGI solution uses a cache directory). You are effectively paying for some very sophisticated know-how as part of every SGI machine.
Given the direction that SGI is heading (Linux for entry-level&apps + IRIX kernel extensions for high-end) I would wonder whether the LCC would produce anything practical in a realistic time-frame. This is not to decry their laudable efforts and I would hope businesses are patient enough to wait for robust and cheap solutions. If nothing else, it will hopefully offer a shardardised set of software extensions (a la OpenMP) and coding practices so that a single source tree can support 1 to n processors.
Who knows, they might be able to come up with a few tricks that the pros have missed.
People forget that one of the original purpose of companies was to raise money to fund risky ventures to plunder... errr.... trade with the New World. Given the state of navigation technology at the time, nobody in their right minds ventured everything on a single bet. Hence the development of syndicates, then Loyds to insure against loss, and stock exchanges to allow people to enter/exit various syndicates. Over time, the stock exchanges have evolved to the best gambling joint on earth (and of course you know that the only people that are guarenteed to make money is the house). As the old wag goes, "I invest, you speculate, he gambles". The essential role of modern stock exchanges is to shift risk onto people who are willing (and bright/dumb depending on viewpoint) to bear it with the mantra that returns match risk. Thus as a youngster with no life, you can afford to bet everything on obscure startups whereas retirees prefer steady income streams from things like utilities and blue chips.
If you think about human activities, a major determinant is the climate and weather. Sales of ice-cream soar on sunny days. Harsh winters push up demand of heating fuel. Derivatives are a mechanism for rapidly signalling events and allowing people to hedge and avoid being caught out by unexpected circumstances. Energy markets are going to be a key beneficiary as well as agriculture, tourism and discretionary events (if it gets rained out then you don't lose everything). Betting on the weather hedge funds that are likely to appear sooner is probably a surer thing that IMHO some of the current IPOs which seem to be scraping the barrel of good ideas.
All your points are valid and I'll briefly explain the nitty gritty:
1) global circulation models are actually done by people in the US, downscaling via nested regional models are limited to this part of the world and if and when the system becomes operationalised, is expected to be distributed. Think cooperating groups around the world sharing the CPU burden
2) the 100m models are interfaced to streamflow and catchment models which are only a comparatively small region set within with the wider desert (rather uninteresting). Think sparse multi-resolutional hierarchy.
3) futher submodels are inherently linear in space/time, while the climate fields are calculated once, the bulk of the operational landscape runs the scientists are interested in are multiple ensembles which require lots of memory, hence some rather painful use of staging and compression. Think conversion to streaming media rather than static files.
If you're interested in more details, send me your email and I'll point you to some of my papers.
I'm a little amused by people asking for a comparison of the costs of a Cray SV2 compared with a Beowulf. For a certain class of applications, they remain dominant and some groups are willing to pay the extra premium for that niche, regardless of the absolute costs (and don't forget the cooling/storage/manpower multipliers). Much like a train (vector computers) is suited for different terrain than buses (shared-mem) or trucks (SMP), vector computers provide very cost effective REAL computing power, often obtaining 50-60% of peak performance whereas you'd be lucky to see beyond single digits for MMPs (and before I get roasted, I'd qualify that by noting decent compilers and reworking algorithms often overcome initial technical limitations).
As for the US support of Cray, well, jaded veterns of comp.sys.supercomputer and HPCC practitioners are well aware of the historical situation with federal funding, technical advantages and bang-for-buck comparisons with Fujitsu and NEC vector computers. For people interested in what the Japanese are doing, I believe NEC are planning on introducing a 1 Teraflop machine with the goal of hitting 5 Teraflops for their Whole Earth Simulator project. Some scientists' idea of heaven is a dedicated vector box and for their purpose and types of code, it is a valid desire.
The SV2 is a curious beast, effectively the first stage in the merging of the Origin cc-NUMA memory subsystem and vector chips. You can think of it as a hybrid box allowing various combination of graphics pipes, MIPS/Merced nodes and vector nodes. The gripe of some people is that they are looking for a successor to the top-end T90 and they are impatient. However, developing at the high end is always trickier than people expect (witness Merced) as you need to increase capabilities along a multitude of dimensions (memory latency, I/O subsystem, heat dissipation, networking) rather than relying from the automatic boost from Moore's Law. Unfortunately there are very few applications which demand absolute performance regardless of actual cost.
To paraphrase crass consumerism, if you have to ask about the price, you can't afford it:-).
Signal 11 wrote You know, in all seriousness, remote learning won't achieve widespread acceptance within the next 30 years for one simple reason: you can't flirt with somebody who's in another state. Seriously.. human interaction is a key component to learning.
Once you've got your hormones under control, I'd like to respectfully point out that the college crowd is making an increasingly smaller portion of the higher education market. For people serious about getting access to further education and not interested in chasing skirts (due to existing family life), web-based material (if correctly designed and supported) may offer time savings not possible otherwise. I am amazed at the dedication (not to mention stamina) shown by some colleagues who juggle a full-time job, raise a kid, build up a technical web site, and still have time for part-time postgraduate study. Also your comments about the requirement for human interaction may be correct for kids with short attention spans but IMHO the biggest bottleneck is trying to work out the conceptual gaps in a student, then suggesting a course of study that fits their inclinations. Given the incredible discrepency in talents and backgrounds, even within a small class, it becomes a Herculean task to create a learning structure that satisfies everyone.
Now, the online "classroom" idea I dig - multimedia presentations are a great way to convey lots of information quickly. Just turn on the disovery channel, or "Bill Nye the Science Guy"
General science/tech/business edutainment aimed at mainstream would have little intrinsic value except in building a broad general knowledge. Solving problems, mapping theoretical techniques to applications, and simply knowing where to go for help can't be taught easily through a passive medium. I suspect that whole-scale uptake of remote instructional technology will be primarily by the corporate universities (perhaps outsourced to specialist firms). Let's face it, mental geniuses they may be, universities do not always have a lot of relevant real-world experience. The real competition for Stanford, Harvard, or ColumbiaU is not going to be other universities, but professional courses offered by the big 5 accounting firms or IBM/Microsoft/McDonald/etc.
I think people are simply not aware of how much it costs to put a course online (and you wondered why your ol' prof recycled the same material decade after decade:-) ). I recall the Open University (UK) estimate that it costs ~$1 million to create a single 3 year subject, plus the cost of revising it in another 3 years. Putting notes on the web and adding email support are a natural complement to normal lectures but nothing can replace face-face interaction with an expert, nor collaborative work with peers. Despite the desires of corporate educational providers, teaching will continue to be labor intensive until someone invents 3D AI avatars that can read minds.
Perhaps there are some roles where computer intensive training would be more effective (e.g pilot training) or can compenstate for disadvantages (braille translation) but for technical stuff, there's no alternative to getting your hands dirty (so why am I procrastinating reading this instead of trying to find that obscure timing bug?:-( ).
Give me a decent book by an engaging author anyday.
His simplistic response: support the Web Consortium in its fight as guardian of the Web.
Given the rather decentralised nature of the web, I'm not sure whether the concept of a "guardian" is valid. Let's face it, if a big enough organisation thought it was in its interest, W3C would be completely ignored (golden rule, he who has the gold makes the rules). Self-awareness of potential pitfalls, the ability to code up a superior alternative, and rapid distribution of the anti-body is the only defense against wholescale assimilation. From this perspective, one can view OpenSource as a necessary counterweight to the rapidly mutating virus of shifting specifications disguised as standards.
But Berners-Lee is almost naïve in interpreting the political, cultural and economic context in which the battle for an open Web will will take place...
Geez, give the guy a break. You transplant a scientist from the collegial atmosphere of a research group into the shark-infested waters of ego-driven CEOs of the business world and you expect the same skillset and world-view to undertake a Jekell and Hyde transformation? I suspect he will grow into the role after a few decades of trying to pacify all those market gorillas to have table manners and use common standards.
He underestimates the rapacious power of American capitalism, especially the new media variety. The Information Age has spawned a whole new breed of corporate monsters - CBS Viacom, Time Warner, Microsoft, the Sun-Netscape Alliance, AOL, the computer companies, the telcoms - whose very existence depends on controlling chunks of the new digitally-sparked economy, sealing territory off and charging for access to information and services.
What wooly piece of thinking led to this conclusion? If you study the lessons of history, you'd find that opening up new worlds have a nasty habit of undermining previous socio-economic structures. The discovery of the New World and subsequent exploitation of the gold/silver mines led to such initial prosperity and later inflation that the royal houses of Europe could not withstand the tides of republicism (think French Revolution). Given that average Joe could have access to their own media making tools (digital cameras, video editing etc), what makes you think they will want to watch reruns of recycled storylines? It is quite likely that capitalism will adapt and evolve to morph into a form which would be unrecognisable to us as we know it. When a corporation becomes bigger than a mid-sized country, then suddenly social factors such as career structures and social/environmental responsibility becomes major concerns.
In the United States, corporations have only one ideology: make the most money at all times in the most expedient way. There's no room in their management philosophies for ceding money to equalizing fantasies about the Web.
This is a rather simplistic view of the role of corporations. Corporations are legal entities that reflect the values of the owners (shareholders). Money in this context is a somewhat imprecise measure of the efficiency of achiving the stated corporate mission. One can point to highly effective nonprofit entities such as Salvation Army which deliver highly valued social services at a lower cost than government departments. Corporations exist only within the framework of the wider communal desires of organised groups, whether city-states or national constitutions. For example, if a corporation suddenly decided that their mission in life was to be the best mercenary army on earth, I suspect governments would suddenly move to put a crimp on their activities. If the perceived popular view of corporations is that of greedy exploitative imperalists, then it is a reflection of the society and its values. If bread and circuses are desired by the population, then corporations will spring up to provide bread and circuses. Economics is driven by different beliefs in what is valued, whether a faster car or more interesting book. Decrying cultural mindsets and mores just because they are different to the "good ol' days" have been a continual theme down through history but social movements are mechanisms for permanently shifting group values (think reformation, think feminism, eco-movement, etc). If you believe corporate greed (which at least is designed to fight each other rather than the customer) is a major problem, what alternative system can you offer to achieve superior results to fufill individual and group expectations?. Of course, if you convince enough people to believe in anti-consumerism, then kudos to your karma.
jkorty wrote The article mentioned one database company that gathers together Massachusetts court records and then charged fees for viewing. In the perfect, future world, each court would instead make the raw data of all its decisions available directly on the web. Researchers then mine this raw data to their heart's content. In this scenerio there is no place for database compilation company to insert added value.
Conceptually this argument makes sense. However, the implementation may easily be hijacked. The cost of retrofitting computer databases and the associated infrastructure (sysadmins, hardware, software) is going to impose a major cost on the public coffers. I can easily see private providers stepping into to offer "free" web sites in return for "exclusive access". One example is an enterprising soul offering discounted high school application hosting in return for ads and access to a captive audience of school kids and teachers. Another example was the recent attempted licensing of (Californian?) license numbers to mass marketers. Would you be willing to hostage your legal system to similar schemes? All we can hope is that some clued-up clerks are knowledgeable about OpenSource so that the implementation as well as content is open to public scrutiny. As they say, the devil is in the details of implementation.
I suspect I would not be the only person concerned about the conflict between public and private data collection. With federal databases, at least there are some legistlative guidelines and open media scrutiny of public office. Only vigilent watch of corporate practices can avoid similar abrogations of privacy rights in the private sphere. As other companies have shown, controlling the source of information whether news or publicity is akin to sitting on a gold mine, especially with the increasing popularity of spin-doctoring. If and when the legal system becomes subverted by vested interests, you might as well emigrate for the law is (supposedly) the only protection the weak have from oppression by the majority (assuming a public gun-fight is overkill for making a protest).
scientific data should be priced at cost of distribution
if the distributor subsequently adds value to the data, then the price should be no higher than the marginal cost of adding value
advocate the incorporation of equivalents of "fair use" as part of any regulatory structure applying to databases as such
While there has been some debate with scientific data (rather useless outside specialist fields), the case of commercial data is less obvious. There have been a few historical cases which have given people pause about the monopolisation of data. One specific example was the privatisation of some early LandSat satellite imagery which according to one viewpoint, was immediately priced to the legislated maximum which effectively stunted academic research into algorithms for processing satellite imagery and any follow-on applications. Other raw data by definition can only have value if shared, e.g. is meterological data spread across multiple countries. With the increasing automatic data collection and computerisation, the potential of conflict for owners and users of databases will only increase.
One of the biggest issues is how to "price" the assembly and aggregation of disparate data. Even pure scientific data could have some commercial value (e.g. genetic codes) under the right circumstances. One solution may be to provide the raw data and the processed value-added stuff and let the market judge whether it is cheaper to massage the raw themselves or save time in purchasing the processed.
Another approach is to create data rights limited by geographical, time, or functional scope. However, this in turn raises more problems in debating to what extent data can be altered before it is considered a unique "new" work (compare with music mixes or composition of existing recognisable art scenes). How far down the value chain is one allowed to claim a slice of the action (compare music score composers claiming a slice of movie soundtracks of their songs)? These are still unanswered questions.
A collection of innoculous facts (e.g. mouse-clicks) can be transformed into a perpetual watch on your web-browsing habits. Given enough time and persistance, any digital event can be tied to a personal profile. Who "owns" this data? A satellite can take pictures of people sun-bathing, some countries would be paranoid to define this as invasion of their sovereign air-space.
In short, the information age will create a whole new raft of problems which will require some legistlation just to clarify any ambiguities. IMHO some time limits would be the most likely solution, even sensitive federal data can be declassified after a suitable cool-down period. But unfortunately I suspect that until some people have seen how far the system can be abused, I doubt whether there will be any popular outcry for safeguards.
Theoretically, if Unix systems can be adapted to emulate Windows (WINE, Wabi, etc), then Windows can be extended to run Unix subprocesses (whether in a shell or subkernel), especially if the source code is widely available. Given Moore's Law and assuming a factor of 100% overhead, then you should be able to execute today's programs with comparable performance on cutting-edge processors in 2 years time (which should please certain chip-makers). An ascetic might disprove of the bloat but then storage is reducing in price even faster than CPUs. Given the shortage of human talent and that software development is the time consuming component, then it makes financial sense, even if you have to fork out thousands for a faster CPU.
Given two equivalent OS markets, if it is easier to port from one to the other than vice-versa, then over time the asymetric flow will benefit the lower cost solution (think game theory, think thermodynamics and the flow of heat from high to low regions). As this is a structural shift (ignoring any short-term pricing tactics), it will only become obvious in the long-term.
As others have noted, Microsoft is composed of some very smart people, savvy managers and... ummm (yeah).... zealous and dedicated marketeers. The question to the wider IT community should be does Linux or Microsoft offer the best value solution in the long-term assuming it is possible to cross-port between the two?
Universities have always evolved over the centures, from the birthing grounds of monastaries (still reflected in graduation regalia) teaching reading/writing, to librarians of classical times, to liberal colleges after renaisance, to modern technology powerhouses. What form it will take in the next century, whether virtual or otherwise, remains to be seen. One thing I have noticed is that as economies switch from agarian->industrial->manufacturing->service->knowl edge, the average age at which a person becomes "employable" rises. Thus while a kid can watch over herds, nowadays you need a minimum of honors or double degrees to have half a chance of getting into a professional career.
I think people underestimate the difficulties facing tertiary education. The problem is that there is no single "university". Instead that term can cover the gamult of technical education (adult learning), teaching colleges, research universities, plus corporate labs/training campuses, each trying gain the prestigue of being called a university and thus diluting the value of the term. Also given the rising costs and reduced public resources, it will be more expensive for individuals to select the education they desire.
The unspoken question in the interview was how to sort out the "right" choice for each potential student. Given humankind's inability to predict the future, many young people have no real clue as to what they want to do in life, much less the best method to cultivate their talents. In the grand scheme of things (ignoring any taught material), universities act as a filter and sorting mechanism, allowing companies to select the top 5-20% on the basis that if you've got half-a-clue and are willing enough to slog it out for 3-4 years in a competitive environment with your peers, then at least you are marginally employable and can thus be trained:-).
On the role of teaching, there are certain basic foundations which are essential for certain disciplines such as mathematics for many of the physical sciences. How many would willingly take up maths if it wasn't forced down their throats at an early age? While Richard Feynman was correct in saying "If you can't explain it, you don't understand it", there is a serious shortage of people who are both brilliant at research and excellent at teaching especially if the institutional incentive structures are not aligned this way. In short, you are asking for superstars, and in turn these superstars want superstudents (otherwise it would be a waste of their time) and the universities set harder entry barriers which raises the average cost (fewers students per staff). A rather tough cycle to break out of, not to mention the general upmanship and educational arms race (my supercomputer is bigger than your supercomputer etc).
So how can universities change to become more effective? The major problem is that education is not really market driven, more like a semi-regulated oliopoly (state charters, national certifications, etc) with all the expected distortions. The prestige factor alone can influence students in selecting a potential course though they may well not be totally suited or happy. Demonstrating competence is a difficult feat as there are many different skills and requirements for different disciplines. It is not like open source where one can point to as the CV and others can determine the quality (or lack thereof). Given the arrival of the web, I would toss out some possible directions education can head:
students publish their work on the web and employees can eyeball their ability
fractionalised departments from different universities can join up and offer more complete courses
more independent evaluation of courses and student grapevine
you don't like your notes, copy another uni
more clueless use of wasteful technology
students end up learning more about the real world by participating in group talkfests like/. !
So to sum things up
universities have changed over the centuries
there are diverse elements in the tertiary sector
as a filtering mechanism, universities do provide a rough ranking
research and teaching superstars are rare
more transparency will (hopefully) lead to better choices
Buying the hardware is only 15-30% of the total cost. Also, in a production environment, you should not be fixated by the CPU. The question should be, within the capital budget, what is the best combination of resources that maximises the effectiveness of achieving your mission.
To give you some real-world experience, a group I'm working with is looking at continential-scale simulation at a 5km resolution with the aim of going down to 100m. Now despite what most people think, the bottleneck (in this example) is in fact the I/O, with estimated total requirements of 30 TBytes. Doing the sums show that to keep up with the CPU (say hypothetically 1 run/24 hours), you would need average throughput of 350 MByte/sec. Hardware that supports both this volume and capacity is NOT cheap. We would joke that we paid x million for the I/O and SGI would throw in the Cray for free:-).
Now as for how an Alpha cluster could be used, it would fit very nicely into the dedicated batch box category. It has a very high CPU rate and some decent compiler optimisation. As such it would augment whatever existing environment exists, reducing the workload of the more expensive machines for development which generally have better tools (just you try debugging a multi-gigabyte core dump). The biggest problem nowadays is not the algorithms, but managing the data traffic to the CPUs and this is where Linux clusters are weak with relatively slow interconnects, unbalanced memory hierarchies, and cheaper but higher latency memory. You have to accept the disadvantages and shift jobs which are not suited for this architecture off. A bit of smarts goes a long way in stretching the budget.
After the fracas with Yahoo's takeover of GeoCities, I think I would not be alone in expressing some concern as to the "ownership" of all the submitted content of slashdot. I have previously submitted personal observations on the assumption that slashdot was a private forum and not a public newspaper where libel and/or copyright issues can arise. While I think people are generally reasonable and don't object to profit-making ventures (e.g. I can see scope for a yearly compilation of the best stories/comments from slashdot... let's call it "Voices from the Source":-)) it does raise the tricky issue of the readers also being the contributors and thus technically copyright owners of their own words.
I think least there be a major misunderstanding along the lines of Yahoo/GeoCities, I would appreciate some assurance or indication of the separation between roles of the various key stakeholders.
In most countries, there are certain laws to protect consumers such as trade practices acts. While commonly used for goods and services, I have not heard of any widely-publicised cases for software. Some specific business conducts which tend to be frowned upon are:
a) refusal to supply - in vertically integrated businesses, refusing to supply an intermediate product in order to damage a competitor. This is a particular worry in the fast pace IT industry as rapid obsolescence can lead to a permanent loss of competitiveness.
b) third line forcing - making acceptance of goods from another party a condition of supply Apart from the obvious one of hardware/OS, this is a rather subtle act for pure software as it is not always clear who owns shares in what company and thus manipulate the consumer into buying closely held products.
c) limitations on resale - intrinsic value of a product (e.g. car) is often determined by the resale value but licenses tying software to particular CPUs or sites make it difficult to transfer software easily.
d) failure of implied warrenties - fair trading acts usually specify minimum rights such as clear title, merchantable quality, and fitness for specified purpose. If anyone has read the end-user licenses recently, they might get a shock as to what they are really buying.
What is surprising is the relative placidity of consumers when it comes to software quality. If cars break down as often as hardware/software, there would be mass class action lawsuits. The question should be why are we so gullible and how long will people put up with it?
What a lot of people forget is that a mass audience eventually evolves into a mediocracy which caters to the majority opinion (you know that big lump underneath the center of the bell curve) for the given cost. On the other hand, due to the lower transaction costs, the internet allows a niche audience to assemble which, with the right structure and feedback (thanks Rob for the debates on moderation), IMHO actually creates a meritocracy where the "best" (as judged by the/. audience karma criteria) ideas tend to surface and get more widely circulated. Thus more marginal and reflective opinions tend to be disseminated creating a richer and diverse base (still centered around the specialist niche though).
For some people, access to any form of semi-intelligent debate is probably a significant improvement over the opinion pages of magazines and newspapers (the real competition to/.), especially when all contributors have an equal chance of submission. It has the advantage of allowing human nature to get on a soapbox and express itself without the fear of public speaking as well as a degree of anonymity. Also/. fits a void between the very large broadcast mediums catering to millions of people (TV, newspapers) and narrowcast mechanisms (chat rooms, bullitin boards, social clubs) which can only scale to hundreds. It will be interesting to see how/. will evolve over the next few years, especially as other groups start copying its format.
ObJoke - why is TV a medium?.... because it is neither rare nor well-done.:-)
What some companies tend to overlook is that the upfront costs is the sunk infrastructure but the long-term value is in the services on top. Unfortunately it does lead to a chicken and egg situation as nobody wants to seriously risk all that capital without having some guarenteed of payback down the track (can we say cable companies here?) but you can't create new services without experimental test-beds and cheap access. Thus the Canadian development is rather bold in opening up high-speed networks without obvious immediate commercial paybacks (and anyone mentioning video-on-demand should look at the distinct lack of consumer interest in trials so far) but then I suppose that is the definition of R&D.
Apart from 500 interactive home shopping channels:-(, what applications would benefit from greater bandwidth? Perhaps security monitoring (check on baby-sitters at home), or full-screen help-line type activities but one bottleneck I see is that the traditionalists who own the infrastructure may not always be the innovative providers, ie all the gear but no idea. Trying to "add value" by competing with customer applications which sit on top of the communications pipes (now that bandwidth and connectivity is becoming a commodity) is a rather interesting way of pissing off friends and creating enemies. The appearance of ISPs outside the traditional telcos indicate that it is not always easy to look outside traditional sources of cashflow.
So can anyone name any telco which has been particularly good at bringing new services to consumers? Or do/.ers think they mainly copy (and compete/crush) the plans of internet startups?
Troy Baer wrote Something to keep in mind about the the Origin 2000 (SGI's 128-256 CPU boxes) is that they're not SMP systems. They're ccNUMA machines, and a lot of the "ccNUMAness" (including cache coherence, I think) is handled largely by the hardware.
The point of ccNUMA is to minimise the cost of porting software from uni-processors. The crux of the matter is that it is non-trivial to adapt programs to run on multiple processors efficiently. The ideal is to have a single source tree, add extensions such as OpenMP, then recompile. Kernels are a different matter as they have to be closer to the hardware. It is still a royal pain to code to the wire and manually manipulate the cache and bus protocols but that is what is needed for maximum performance. Apart from special cases such as national defence codes, the commercial imperative is time-to-market which means a ccNUMA machine can address 95% of the issues at reasonable cost would be preferred.
I wouldn't be surprised if you could boot the MIPS version of Linux on (for instance) an Origin with little or no modification. I don't know how well it would scale, though.
As far as I'm aware (correct me if I'm wrong), the SGI port of Linux has so far concentrated on older systems such as Indys and patches for their VisualWorkstation. I suspect it will take a while (2-5 years?) for them to get to the stage of having Linux+IRIX SMP extensions running on their highly scalable systems. Cellular IRIX is a single system image which is different from the way Linux is designed. Perhaps one conceptual integration approach is to follow how RTLinux works in having a separate real-time kernel embedded within the full Linux system. Also there are other multiprocessor optimisations like processor affinity which might take a while to enter into the kernel. SGI staff may be very enthusiastic and dedicated but there is a lot of work involved which will take time.
In other words, nice PR for SGI but don't hold your breath.
LL
Comparative strengths of designs ... was SunScreen
on
IBM takes aim at Sun
·
· Score: 2
OK, time for Buzz Word City
For the type of applications Sun and IBM are targetting, raw processor speed becomes a marginal parameter in the design space. Instead, the smart cookies focus on the I/O-memory subsystems which becomes the performance bottleneck and cooling system which determines the overall reliability.
The StartFire is based on cross-bar switch technology spun off from Cray when it was purchased by SGI. As such it is analogous to a shared bus where processors (ie passengers) find it relatively easy to communicate between each other inside but the maximum capacity is somewhat fixed in that it costs the same whether a few passengers or fully populated. However, it has the advantage of being easy to migrate.
The original IBM SP series from which their current distributed memory design is derived can be compared with a truck convoy using CB radio (switched memory backplane). More flexibility in adding capacity means better efficiencies in matching load to task and thus better pricing (though at these lofty levels, the profit margins are MUCH heftier than PCs so there is a lot of gap for undercutting the competition).
The SGI ccNUMA (cache-coherent non-uniform memory access) can be compared with an articulated truck with multi-channel CDMA wireless giving a hypercube topology for fast node-node communications. Specifically designed for scalability and balanced I/O throughput, it commands a premium for its complexity and sophistication.
On a sliding spectrum of shared-distributed, the order would be Sun-SGI-IBM, but as processors speeds increase, both Sun and IBM are adopting ccNUMA techniques. Now a diversified transport system would require a judicious blend of each computer, matching the capabilities to each machine's strengths. Any claims of superiority are marketing delusions as you would not use a bus where a truck is needed. That is what supposedly CIOs get paid 6 figure incomes for deciding and service arms like IBM get fat consulting fees (any reader comments on their effectiveness?).
Note that raw technical considerations can be distorted to some extent by legacy concerns and availability of drivers (sys-admins). Personally I see the high-end server space get more competitive and cut-throat as souped up cars attempt to claim a slice of the action. However, some companies will need to hit with a clue-bat as the planned technological obsolescence of consumer items do not sit well with business (there are good reasons why you'd stick with old-style mainframes based on the principle if it ain't broke, don't meddle with it as more fixes are likely to add more bugs).
Slashdot Mirror
Despite the rather black box connotations of reverse engineering, it is a legitamate R&D exercise. Car manufacturers reverse engineer their competitors, chefs try to dissect recipes, etc. In the computer context according to the Centre for Software Maintenance,
....).
Reverse Engineering is the process of analysis of an existing software system to create representations of a different form or higher level of abstraction.
and
Reengineering is the process of analysis and modification of an existing software system to reconstitute it in a new improved form.
Given the preenial occupation of engineers is to make things better, faster, or cheaper, tinkering with the electronic toys or source code is a natural pasttime. It is only the marketeers and financial managers that want things to be "hidden" so that the cost (and thus in their mind = value) is higher (basic economics, remember scarcity == higher price). Obfuscation of code is an obvious mechanism to exclude competitors, however, it significantly adds to the long-term cost of maintenance and also reduces the potential market. How many times have you've been given a piece of code with the design specs/architecture residing in someone's head who've just left? Wouldn't it be nice if some intelligent bit of software did the analysis and gave you the answer (yeah, wishful thinking but still
Perhaps people don't realise it but there are 2 information monopolies, one when one party controls everything and there is no alternative, the other when everything is freely available so that there is no competition (and thus no alternative).
LL
{Parody mode on of original}
These god-fangled Model-Ts may be Ford's hotest piece of metal but are they the car of the people?
Would you recognise a z-crank if you tripped over one?
My month-long effort to be cool and try and move from point A to point B without crashing at point C,D,E,F gives me chilling flashbacks to learning how to ride a horse without having people laugh at me. For now, I'm glad to have ol' Silver back with me to rely on.
The good ol' boys warned me beforehand. They had trying finding the pedals and their friends throw bits of metal around on that silly assembly line thingy. I'm car-literate since I've actually washed an early model before becoming a news flack. We're in the 90s but I'm a bit clueless about all the moving odd bobs inside.
My advice is if you've barely mastered whip-cracking and never changed a tire, stick with the ol' feller or take the train. If you know an oil change from elbow grease, perhaps you'd even find this fun (gasp of shock and horror).
In truth, stalling made the task of backing out the driveway wasted the whole day, wasting at least 3 hours figuring out those crazy levers. In the end, with no decent horse sense and as chummy with a wheel as I was with phys-ed torture sessions, you know, the one where the teach' nursemaided those skiiny geeks and prevented us jocks from having fun.
I opened the engine and figured out the hole for that metal thingy you call the crank handle, right? I try to yank the gear as written on that flimsy bit of paper they call instructions manual but those gunshot backfires nearly gave me a heart attack. Not a good omen.
I turn to the "Cleaning Engine" page in teh manual but couldn't figure out that picture with all the bits and pieces meant. The next page says something about changing oil but all I want to do is drive this silly hunk of junk, not build it. Opening up the engine didn't show any spots to hook the reins.
I turn the handle but think something's wrong as it was making all these funny noises like my horse has colic or something. I press the pedal and nearly wiped out my favourite mail box. At least a horse is smart enough to avoid impaling itself.
Well, it least it seems to move but don't know why I have to keep looking over my shoulder to see where I'm going all the time. I survive smashing into the barn door but the manual warns against driving without flags and horn blowing. I get outright dirty trying to count the cylinders and rpms as suggested, taking half an hour to motor 50 feet back to the house.It then dies for some strange unknown reason so I know it's time for lunch.
After a nice big juicy steak, I give the od' editor a hollar but he mentions something about gas fill-up and to check the tank (as if I'd containminate my washing water!). After a while, a neighbour drives by so I swallow my pride (yes siree, gave me indigestion for resta the day) and asks what a gas tank looks like. He mocks me "And you're learning how to drive?". Next time he gives any mouth, I've gotta shotgun handy. He helps me fill it up a bit (at least I guess which is the right hole) but then some shit musta hit the fanbelt so I drop this project for a couple of hours.
At this point, I'm burnt out. I start wildly guessing buttons to push. God must have been with me as it started again (must remember to buy a new can for that mailbox). I can make it to the barn (and even stop!) without hitting anything too important. That's good enough for me.
Then I head home, bleary-eyed. My superficial knowledge of gasoline engines made this project frustrating. I would have helped if I were a mechanic. On the other hand, I'm surprised I got as far as I did, just like trying to connect with a baseball bat. It was mildly fulfilling in a mysterious way. I may have no idea how to use this car but I got my hands dirty.
{Parody mode off}
I sure would go a lot better when someone invents the automatic GUI transmission! No disrespect to the poor guy but unfortunately it is relatively early days and the only way to learn is to be willing and get those hands dirty. Congrats on making a start and a warning to the rest of the Linux mechanics that exotic details of kernel file spaces is as relevant as quantum physics to the average driver.
LL
Perhaps E-Bay is trying to eliminate future corruption of their customer profile?
If you read the marketing-speil coming from all the current e-commerce sites, be very afraid. They want to know how who, what why, where, when and what you spend. Now if a robot that emulates a human search (or aggregate metasearches) then GIGO (Garbage In, Garbage Out) rules apply. For people who work in the database field, the technology is the least of the worries as cleaning up the data is probably 80% of the work.
similar analogies would be
- creating artifical users on Hotmail
(their name space is NOT infinite)
- inserting mega dynamic indexes to Yahoo
(overwhelm the human classifiers)
- creating recursive reviews to Amazon
(drowning the human comments with noise)
In short, anything automated can be screwed up to some degree. If the script kiddies ever get a clue, I suspect there will be some sort of arms race between robot automation and exclusion. Fun and games.
LL
Hmmmm ..... looks like some hardware assist to help binary retargetting. For people not familiar with the concept, take a look at an overview. The concept is sound in that as ESR points out 95% of the programming jobs out there are spent in maintaining old code on old machines. However, if there was a way of abstracting and specifying the hardware characteristics and mapping from one to another, then old binaries could be shifted onto newer and cheaper hardware with less hassle. I can think of cases like old Cray binaries where porting them to a new MPP would be too painful manually, some of those timing cases can be really subtle. Given that computer companies are very relunctant to support hardware which isn't current (ie not profitable) and others could potentially go belly-up (correct me if I'm wrong, I think only IBM is one of the few giants left from the 60's), there is a need to protect the million of man-years spent on specific packages. Of course, research has shown that retargetting works better with availability to the original compiler source :-).
Given the rate of corporate take-overs, you could quite easily end up running a zillion different systems and lose valuable time in trying to consolidating everything.
Oh well, add this to the speculation pile along with everyone else.
LL
In your paper The Magic Cauldron you talk about the sale value (final cost to consumer) and use value (what economics would consider a capital good). Modern capitalist societies have evolved very complex and sophisticated instruments (a la the profit motive) to price these goods and signal to the market what is valued. With OpenSource, this pricing information is missing and thus time/effort is spent on "sexy" projects like 3D interfaces (what economists call malinvestments) instead of really important stuff like good optimising compilers.
Question 1) Pricing of OpenSource Software How can OpenSource software be fairly priced given that it is always possible to undercut a distributor?
Question 2) Distribution of Resources Instead of vertically integrating all the profits at the sale end (distributions like Red Hat), how can the creators of the intermediate goods get enough funding to continue refining their products?
Question 3) Scaling to Megaprojects Given the limitations of no capital pool of funding so that intermediate software can maintained, will OpenSource projects be limited to "small" projects that can be supported by 1-6 key designers and wouldn't this be an inherent constraint?
Question 4) Bazaar Rules of Conduct At the moment, the Software Bazaar is controlled by gentleman rules of conduct (no encroaching on projects, equal sharing, etc). Do you see this continuing with the increasing commercialisation (and potential moral corruption) of the hacker's "gift culture"?
Question 5) Software Patent Roadblocks In a situation where time-to-market is becoming a key factor in dominating the bulk of the profits (see some notes on game theory of software patents), how can OpenSource work around limitations of key locks on irreplaceable algorithms?
Question 6) Freedom to Copy. While big companies like SGI have the resources and name-brand equity to release and protect their OpenSource efforts, how will smaller entry level players survive long enough without their ideas being poached by bigger companies?
LL
I'd hate to rain on anyone's parade but wouldn't this military wet dream be superfluous? Let's suppose I'm one of these countries with tac-nukes. Why would I bother announcing the launch site and invite retaliation by using balistic missiles? Better still, just to ship it into a anonymous cargo hauler and detonate it within some strategic harbor or even Panama canal. That way the source (assuming you can disguise the origin of the manufactured weapon) can be anonymous. Given the gung-ho way the US been acting around the world in the last few decades, I'm sure there's no shortage of splinter or fanatic groups to spread the net of suspicion. If people are interested in the military mindset, take a look at their parameters magazine, in particular the article by Peters on "Our New Old Enemies". Very interesting.
... I believe US and Britain are still the largest exporters of arms) but does little to create long-term goodwill. Exporting organised violence seems to be a self-fufilling prophency as it propagates a climate of fear. Afterall, if you think someone is an enemy. then what are the chances that every action you perceive is hostile? Psychopaths are not the only people with a warped mind-view, a entire culture can be infected in rather subtle but destructive ways (Andy Grove "Only the Paranoid Surive", Bill Gates "Technogy is great, but 90% market share is better"). Very successful but at what cost?
People don't go to war for no good reason. If you create a threat, then people will respond in kind. Defining enemies through an arms race might be good for the military-industrial complex (correct me if I'm wrong
This century has seen 2 world wars, numerous regional conflicts and ongoing bushfires. I would hope the next century has a better record.
LL
belswick wrote
:-) ).
Note that SGI is showing all the signs of entering the death throes stage. Another 30% of the workforce laid off, abandoning major initiatives, CEO bailing (to MS!!), loss of faith by major customers.
Unless you've got inside information (which the SEC would be very interested in hearing about), I think the slashdot audience would appreciate more evidence than mindless parrotting of popular press. For your information, they are spinning off several portions of their divisions into separate business entities. Now while some people may consider this akin to kicking fledgings out the nest, the rate of turnover in Silcon Valley is such that the difference between working for one company vs another is just which branded T-shirt you wear. Think of it as a beehive with clumps forming and dispersing to form interesting new combinations. Abandoning major initiatives?, how many announcements have you've heard from major companies that have died the silent death of being irrelevant to real needs.
As for the CEO, well, I'm sure there will be some interesting books a few years down the track but for many hard-core SGI purchasers, the shift into Intel consumerism where they did not have any competitive advantaged showed some very wooly thinking (for the cognosti, there is nothing technically inferior about the MIPS architecture). The loss of customers is not surprising considering that many applications that used to be top-end in the 70s can now run on a single modern processor and big cache (the refuge of the lazy microarchitect). Getting a free ride from Moore's Law is not the same as coming up with innovative new software applciations that can really take advantage of increased CPU capacity (apart from molecular simulations which will chew up any CPU cycle you throw at them).
Customers will buy SGI equipment if SGI can show they offer a value proposition that is worth the premium over mainstream machines, whether it is memory latency, quality engineering, coolness factor or whateever, people will buy (oh and getting their manufacturing/distribution process to be more efficient would help a lot). Computers are becoming so prevalent that the only distinguishing feature nowadays for PCs is image and lifestyle (does the color clash with the decor
Reasonable people must expect that SGI goes Chapter 11 RSN (barring a government bailout) and then what happens to people who need supercomputers?
Would you say Apple devotees are unreasonable? Don't you understand that given a planet of 5 billion odd people, not everyone is interested in the toys you are? Cries of doom and gloom have always been around in any industry in one form or another as it gives paper pushers a reason to justify their existance instead of getting their hands dirty coding or designing. You have to realise that SGI serves a fairly specialised market (data intensive, high-end graphics, scientific back-end grunt machines) in the 50K-50M range. Much like Porsche and BMW cater towards a cliental that wants absolute performance and not cheap consumer junk (admitedly the Japanese have given the US auto industry a shot in the arm since 80s), there will always be people who appreciate the qualities that SGI offers. Provided SGI can continue to support those companies and not go around trying to push Porshes for people wanting bicyles (amazing how hype can convince people they need a Pentium III to browse the web) at an affordable price, they will survive.
If you work for a company, you'd realise that the first law is survival which is depedent on their market relevance. SGI will continue so long as their is a demand for their expertise as priced compared with other market alternatives.
LL
As Matt Welsh noted, it is not exactly a trivial problem. If you look very closely at the article, the LCC wants to occupy a happy ground between the share-nothing crowd (Microsoft, Tandem) and the share-everything (Oracle). The share nothing pardigm is rather simplistic in its approach and reflects the fact that throwing together a bunch of machines with a cheap interconnect is a comparatively straight-forward re-engineering approach. The share-everything come froms the extension of shared-bus architectures (e.g. Sun Starfire) which enforces a multiple lock strategy. Companies like SGI have thrown million of R&D dollars into the middle-ground which is why their cc-NUMA architecture and cellular IRIX is quite popular. I wish the LCC luck but there is a reason why a successful working solution is expensive as it requires a savvy combination of hardware+software+smart routing (the SGI solution uses a cache directory). You are effectively paying for some very sophisticated know-how as part of every SGI machine.
Given the direction that SGI is heading (Linux for entry-level&apps + IRIX kernel extensions for high-end) I would wonder whether the LCC would produce anything practical in a realistic time-frame. This is not to decry their laudable efforts and I would hope businesses are patient enough to wait for robust and cheap solutions. If nothing else, it will hopefully offer a shardardised set of software extensions (a la OpenMP) and coding practices so that a single source tree can support 1 to n processors.
Who knows, they might be able to come up with a few tricks that the pros have missed.
LL
People forget that one of the original purpose of companies was to raise money to fund risky ventures to plunder ... errr .... trade with the New World. Given the state of navigation technology at the time, nobody in their right minds ventured everything on a single bet. Hence the development of syndicates, then Loyds to insure against loss, and stock exchanges to allow people to enter/exit various syndicates. Over time, the stock exchanges have evolved to the best gambling joint on earth (and of course you know that the only people that are guarenteed to make money is the house). As the old wag goes, "I invest, you speculate, he gambles". The essential role of modern stock exchanges is to shift risk onto people who are willing (and bright/dumb depending on viewpoint) to bear it with the mantra that returns match risk. Thus as a youngster with no life, you can afford to bet everything on obscure startups whereas retirees prefer steady income streams from things like utilities and blue chips.
If you think about human activities, a major determinant is the climate and weather. Sales of ice-cream soar on sunny days. Harsh winters push up demand of heating fuel. Derivatives are a mechanism for rapidly signalling events and allowing people to hedge and avoid being caught out by unexpected circumstances. Energy markets are going to be a key beneficiary as well as agriculture, tourism and discretionary events (if it gets rained out then you don't lose everything). Betting on the weather hedge funds that are likely to appear sooner is probably a surer thing that IMHO some of the current IPOs which seem to be scraping the barrel of good ideas.
Oh well, no laws against stupidity.
LL
All your points are valid and I'll briefly explain the nitty gritty:
1) global circulation models are actually done by people in the US, downscaling via nested regional models are limited to this part of the world and if and when the system becomes operationalised, is expected to be distributed. Think cooperating groups around the world sharing the CPU burden
2) the 100m models are interfaced to streamflow and catchment models which are only a comparatively small region set within with the wider desert (rather uninteresting). Think sparse multi-resolutional hierarchy.
3) futher submodels are inherently linear in space/time, while the climate fields are calculated once, the bulk of the operational landscape runs the scientists are interested in are multiple ensembles which require lots of memory, hence some rather painful use of staging and compression. Think conversion to streaming media rather than static files.
If you're interested in more details, send me your email and I'll point you to some of my papers.
Regards,
LL
I'm a little amused by people asking for a comparison of the costs of a Cray SV2 compared with a Beowulf. For a certain class of applications, they remain dominant and some groups are willing to pay the extra premium for that niche, regardless of the absolute costs (and don't forget the cooling/storage/manpower multipliers). Much like a train (vector computers) is suited for different terrain than buses (shared-mem) or trucks (SMP), vector computers provide very cost effective REAL computing power, often obtaining 50-60% of peak performance whereas you'd be lucky to see beyond single digits for MMPs (and before I get roasted, I'd qualify that by noting decent compilers and reworking algorithms often overcome initial technical limitations).
:-).
As for the US support of Cray, well, jaded veterns of comp.sys.supercomputer and HPCC practitioners are well aware of the historical situation with federal funding, technical advantages and bang-for-buck comparisons with Fujitsu and NEC vector computers. For people interested in what the Japanese are doing, I believe NEC are planning on introducing a 1 Teraflop machine with the goal of hitting 5 Teraflops for their Whole Earth Simulator project . Some scientists' idea of heaven is a dedicated vector box and for their purpose and types of code, it is a valid desire.
The SV2 is a curious beast, effectively the first stage in the merging of the Origin cc-NUMA memory subsystem and vector chips. You can think of it as a hybrid box allowing various combination of graphics pipes, MIPS/Merced nodes and vector nodes. The gripe of some people is that they are looking for a successor to the top-end T90 and they are impatient. However, developing at the high end is always trickier than people expect (witness Merced) as you need to increase capabilities along a multitude of dimensions (memory latency, I/O subsystem, heat dissipation, networking) rather than relying from the automatic boost from Moore's Law. Unfortunately there are very few applications which demand absolute performance regardless of actual cost.
To paraphrase crass consumerism, if you have to ask about the price, you can't afford it
LL
Signal 11 wrote
You know, in all seriousness, remote learning won't achieve widespread acceptance within the next 30 years for one simple reason: you can't flirt with somebody who's in another state. Seriously.. human interaction is a key component to learning.
Once you've got your hormones under control, I'd like to respectfully point out that the college crowd is making an increasingly smaller portion of the higher education market. For people serious about getting access to further education and not interested in chasing skirts (due to existing family life), web-based material (if correctly designed and supported) may offer time savings not possible otherwise. I am amazed at the dedication (not to mention stamina) shown by some colleagues who juggle a full-time job, raise a kid, build up a technical web site, and still have time for part-time postgraduate study. Also your comments about the requirement for human interaction may be correct for kids with short attention spans but IMHO the biggest bottleneck is trying to work out the conceptual gaps in a student, then suggesting a course of study that fits their inclinations. Given the incredible discrepency in talents and backgrounds, even within a small class, it becomes a Herculean task to create a learning structure that satisfies everyone.
Now, the online "classroom" idea I dig - multimedia presentations are a great way to convey lots of information quickly. Just turn on the disovery channel, or "Bill Nye the Science Guy"
General science/tech/business edutainment aimed at mainstream would have little intrinsic value except in building a broad general knowledge. Solving problems, mapping theoretical techniques to applications, and simply knowing where to go for help can't be taught easily through a passive medium. I suspect that whole-scale uptake of remote instructional technology will be primarily by the corporate universities (perhaps outsourced to specialist firms). Let's face it, mental geniuses they may be, universities do not always have a lot of relevant real-world experience. The real competition for Stanford, Harvard, or ColumbiaU is not going to be other universities, but professional courses offered by the big 5 accounting firms or IBM/Microsoft/McDonald/etc.
LL
I think people are simply not aware of how much it costs to put a course online (and you wondered why your ol' prof recycled the same material decade after decade :-) ). I recall the Open University (UK) estimate that it costs ~$1 million to create a single 3 year subject, plus the cost of revising it in another 3 years. Putting notes on the web and adding email support are a natural complement to normal lectures but nothing can replace face-face interaction with an expert, nor collaborative work with peers. Despite the desires of corporate educational providers, teaching will continue to be labor intensive until someone invents 3D AI avatars that can read minds.
:-( ).
Perhaps there are some roles where computer intensive training would be more effective (e.g pilot training) or can compenstate for disadvantages (braille translation) but for technical stuff, there's no alternative to getting your hands dirty (so why am I procrastinating reading this instead of trying to find that obscure timing bug?
Give me a decent book by an engaging author anyday.
LL
His simplistic response: support the Web Consortium in its fight as guardian of the Web.
...
Given the rather decentralised nature of the web, I'm not sure whether the concept of a "guardian" is valid. Let's face it, if a big enough organisation thought it was in its interest, W3C would be completely ignored (golden rule, he who has the gold makes the rules). Self-awareness of potential pitfalls, the ability to code up a superior alternative, and rapid distribution of the anti-body is the only defense against wholescale assimilation. From this perspective, one can view OpenSource as a necessary counterweight to the rapidly mutating virus of shifting specifications disguised as standards.
But Berners-Lee is almost naïve in interpreting the political, cultural and economic context in which the battle for an open Web will will take place
Geez, give the guy a break. You transplant a scientist from the collegial atmosphere of a research group into the shark-infested waters of ego-driven CEOs of the business world and you expect the same skillset and world-view to undertake a Jekell and Hyde transformation? I suspect he will grow into the role after a few decades of trying to pacify all those market gorillas to have table manners and use common standards.
He underestimates the rapacious power of American capitalism, especially the new media variety. The Information Age has spawned a whole new breed of corporate monsters - CBS Viacom, Time Warner, Microsoft, the Sun-Netscape Alliance, AOL, the computer companies, the telcoms - whose very existence depends on controlling chunks of the new digitally-sparked economy, sealing territory off and charging for access to information and services.
What wooly piece of thinking led to this conclusion? If you study the lessons of history, you'd find that opening up new worlds have a nasty habit of undermining previous socio-economic structures. The discovery of the New World and subsequent exploitation of the gold/silver mines led to such initial prosperity and later inflation that the royal houses of Europe could not withstand the tides of republicism (think French Revolution). Given that average Joe could have access to their own media making tools (digital cameras, video editing etc), what makes you think they will want to watch reruns of recycled storylines? It is quite likely that capitalism will adapt and evolve to morph into a form which would be unrecognisable to us as we know it. When a corporation becomes bigger than a mid-sized country, then suddenly social factors such as career structures and social/environmental responsibility becomes major concerns.
In the United States, corporations have only one ideology: make the most money at all times in the most expedient way. There's no room in their management philosophies for ceding money to equalizing fantasies about the Web.
This is a rather simplistic view of the role of corporations. Corporations are legal entities that reflect the values of the owners (shareholders). Money in this context is a somewhat imprecise measure of the efficiency of achiving the stated corporate mission. One can point to highly effective nonprofit entities such as Salvation Army which deliver highly valued social services at a lower cost than government departments. Corporations exist only within the framework of the wider communal desires of organised groups, whether city-states or national constitutions. For example, if a corporation suddenly decided that their mission in life was to be the best mercenary army on earth, I suspect governments would suddenly move to put a crimp on their activities. If the perceived popular view of corporations is that of greedy exploitative imperalists, then it is a reflection of the society and its values. If bread and circuses are desired by the population, then corporations will spring up to provide bread and circuses. Economics is driven by different beliefs in what is valued, whether a faster car or more interesting book. Decrying cultural mindsets and mores just because they are different to the "good ol' days" have been a continual theme down through history but social movements are mechanisms for permanently shifting group values (think reformation, think feminism, eco-movement, etc). If you believe corporate greed (which at least is designed to fight each other rather than the customer) is a major problem, what alternative system can you offer to achieve superior results to fufill individual and group expectations?. Of course, if you convince enough people to believe in anti-consumerism, then kudos to your karma.
LL
jkorty wrote
The article mentioned one database company that gathers together Massachusetts court records and then charged fees for viewing. In the perfect, future world, each court would instead make the raw data of all its decisions available directly on the web. Researchers then mine this raw data to their heart's content. In this scenerio there is no place for database compilation company to insert added value.
Conceptually this argument makes sense. However, the implementation may easily be hijacked. The cost of retrofitting computer databases and the associated infrastructure (sysadmins, hardware, software) is going to impose a major cost on the public coffers. I can easily see private providers stepping into to offer "free" web sites in return for "exclusive access". One example is an enterprising soul offering discounted high school application hosting in return for ads and access to a captive audience of school kids and teachers. Another example was the recent attempted licensing of (Californian?) license numbers to mass marketers. Would you be willing to hostage your legal system to similar schemes? All we can hope is that some clued-up clerks are knowledgeable about OpenSource so that the implementation as well as content is open to public scrutiny. As they say, the devil is in the details of implementation.
I suspect I would not be the only person concerned about the conflict between public and private data collection. With federal databases, at least there are some legistlative guidelines and open media scrutiny of public office. Only vigilent watch of corporate practices can avoid similar abrogations of privacy rights in the private sphere. As other companies have shown, controlling the source of information whether news or publicity is akin to sitting on a gold mine, especially with the increasing popularity of spin-doctoring. If and when the legal system becomes subverted by vested interests, you might as well emigrate for the law is (supposedly) the only protection the weak have from oppression by the majority (assuming a public gun-fight is overkill for making a protest).
LL
While there has been some debate with scientific data (rather useless outside specialist fields), the case of commercial data is less obvious. There have been a few historical cases which have given people pause about the monopolisation of data. One specific example was the privatisation of some early LandSat satellite imagery which according to one viewpoint, was immediately priced to the legislated maximum which effectively stunted academic research into algorithms for processing satellite imagery and any follow-on applications. Other raw data by definition can only have value if shared, e.g. is meterological data spread across multiple countries. With the increasing automatic data collection and computerisation, the potential of conflict for owners and users of databases will only increase.
One of the biggest issues is how to "price" the assembly and aggregation of disparate data. Even pure scientific data could have some commercial value (e.g. genetic codes) under the right circumstances. One solution may be to provide the raw data and the processed value-added stuff and let the market judge whether it is cheaper to massage the raw themselves or save time in purchasing the processed.
Another approach is to create data rights limited by geographical, time, or functional scope. However, this in turn raises more problems in debating to what extent data can be altered before it is considered a unique "new" work (compare with music mixes or composition of existing recognisable art scenes). How far down the value chain is one allowed to claim a slice of the action (compare music score composers claiming a slice of movie soundtracks of their songs)? These are still unanswered questions.
A collection of innoculous facts (e.g. mouse-clicks) can be transformed into a perpetual watch on your web-browsing habits. Given enough time and persistance, any digital event can be tied to a personal profile. Who "owns" this data? A satellite can take pictures of people sun-bathing, some countries would be paranoid to define this as invasion of their sovereign air-space.
In short, the information age will create a whole new raft of problems which will require some legistlation just to clarify any ambiguities. IMHO some time limits would be the most likely solution, even sensitive federal data can be declassified after a suitable cool-down period. But unfortunately I suspect that until some people have seen how far the system can be abused, I doubt whether there will be any popular outcry for safeguards.
LL
Theoretically, if Unix systems can be adapted to emulate Windows (WINE, Wabi, etc), then Windows can be extended to run Unix subprocesses (whether in a shell or subkernel), especially if the source code is widely available. Given Moore's Law and assuming a factor of 100% overhead, then you should be able to execute today's programs with comparable performance on cutting-edge processors in 2 years time (which should please certain chip-makers). An ascetic might disprove of the bloat but then storage is reducing in price even faster than CPUs. Given the shortage of human talent and that software development is the time consuming component, then it makes financial sense, even if you have to fork out thousands for a faster CPU.
... ummm (yeah) .... zealous and dedicated marketeers. The question to the wider IT community should be does Linux or Microsoft offer the best value solution in the long-term assuming it is possible to cross-port between the two?
Given two equivalent OS markets, if it is easier to port from one to the other than vice-versa, then over time the asymetric flow will benefit the lower cost solution (think game theory, think thermodynamics and the flow of heat from high to low regions). As this is a structural shift (ignoring any short-term pricing tactics), it will only become obvious in the long-term.
As others have noted, Microsoft is composed of some very smart people, savvy managers and
LL
I think people underestimate the difficulties facing tertiary education. The problem is that there is no single "university". Instead that term can cover the gamult of technical education (adult learning), teaching colleges, research universities, plus corporate labs/training campuses, each trying gain the prestigue of being called a university and thus diluting the value of the term. Also given the rising costs and reduced public resources, it will be more expensive for individuals to select the education they desire.
The unspoken question in the interview was how to sort out the "right" choice for each potential student. Given humankind's inability to predict the future, many young people have no real clue as to what they want to do in life, much less the best method to cultivate their talents. In the grand scheme of things (ignoring any taught material), universities act as a filter and sorting mechanism, allowing companies to select the top 5-20% on the basis that if you've got half-a-clue and are willing enough to slog it out for 3-4 years in a competitive environment with your peers, then at least you are marginally employable and can thus be trained
On the role of teaching, there are certain basic foundations which are essential for certain disciplines such as mathematics for many of the physical sciences. How many would willingly take up maths if it wasn't forced down their throats at an early age? While Richard Feynman was correct in saying "If you can't explain it, you don't understand it", there is a serious shortage of people who are both brilliant at research and excellent at teaching especially if the institutional incentive structures are not aligned this way. In short, you are asking for superstars, and in turn these superstars want superstudents (otherwise it would be a waste of their time) and the universities set harder entry barriers which raises the average cost (fewers students per staff). A rather tough cycle to break out of, not to mention the general upmanship and educational arms race (my supercomputer is bigger than your supercomputer etc).
So how can universities change to become more effective? The major problem is that education is not really market driven, more like a semi-regulated oliopoly (state charters, national certifications, etc) with all the expected distortions. The prestige factor alone can influence students in selecting a potential course though they may well not be totally suited or happy. Demonstrating competence is a difficult feat as there are many different skills and requirements for different disciplines. It is not like open source where one can point to as the CV and others can determine the quality (or lack thereof). Given the arrival of the web, I would toss out some possible directions education can head:
So to sum things up
LL
Buying the hardware is only 15-30% of the total cost. Also, in a production environment, you should not be fixated by the CPU. The question should be, within the capital budget, what is the best combination of resources that maximises the effectiveness of achieving your mission.
:-).
To give you some real-world experience, a group I'm working with is looking at continential-scale simulation at a 5km resolution with the aim of going down to 100m. Now despite what most people think, the bottleneck (in this example) is in fact the I/O, with estimated total requirements of 30 TBytes. Doing the sums show that to keep up with the CPU (say hypothetically 1 run/24 hours), you would need average throughput of 350 MByte/sec. Hardware that supports both this volume and capacity is NOT cheap. We would joke that we paid x million for the I/O and SGI would throw in the Cray for free
Now as for how an Alpha cluster could be used, it would fit very nicely into the dedicated batch box category. It has a very high CPU rate and some decent compiler optimisation. As such it would augment whatever existing environment exists, reducing the workload of the more expensive machines for development which generally have better tools (just you try debugging a multi-gigabyte core dump). The biggest problem nowadays is not the algorithms, but managing the data traffic to the CPUs and this is where Linux clusters are weak with relatively slow interconnects, unbalanced memory hierarchies, and cheaper but higher latency memory. You have to accept the disadvantages and shift jobs which are not suited for this architecture off. A bit of smarts goes a long way in stretching the budget.
LL
After the fracas with Yahoo's takeover of GeoCities, I think I would not be alone in expressing some concern as to the "ownership" of all the submitted content of slashdot. I have previously submitted personal observations on the assumption that slashdot was a private forum and not a public newspaper where libel and/or copyright issues can arise. While I think people are generally reasonable and don't object to profit-making ventures (e.g. I can see scope for a yearly compilation of the best stories/comments from slashdot ... let's call it "Voices from the Source" :-)) it does raise the tricky issue of the readers also being the contributors and thus technically copyright owners of their own words.
I think least there be a major misunderstanding along the lines of Yahoo/GeoCities, I would appreciate some assurance or indication of the separation between roles of the various key stakeholders.
LL
In most countries, there are certain laws to protect consumers such as trade practices acts. While commonly used for goods and services, I have not heard of any widely-publicised cases for software. Some specific business conducts which tend to be frowned upon are:
a) refusal to supply - in vertically integrated businesses, refusing to supply an intermediate product in order to damage a competitor. This is a particular worry in the fast pace IT industry as rapid obsolescence can lead to a permanent loss of competitiveness.
b) third line forcing - making acceptance of goods from another party a condition of supply Apart from the obvious one of hardware/OS, this is a rather subtle act for pure software as it is not always clear who owns shares in what company and thus manipulate the consumer into buying closely held products.
c) limitations on resale - intrinsic value of a product (e.g. car) is often determined by the resale value but licenses tying software to particular CPUs or sites make it difficult to transfer software easily.
d) failure of implied warrenties - fair trading acts usually specify minimum rights such as clear title, merchantable quality, and fitness for specified purpose. If anyone has read the end-user licenses recently, they might get a shock as to what they are really buying.
What is surprising is the relative placidity of consumers when it comes to software quality. If cars break down as often as hardware/software, there would be mass class action lawsuits. The question should be why are we so gullible and how long will people put up with it?
LL
What a lot of people forget is that a mass audience eventually evolves into a mediocracy which caters to the majority opinion (you know that big lump underneath the center of the bell curve) for the given cost. On the other hand, due to the lower transaction costs, the internet allows a niche audience to assemble which, with the right structure and feedback (thanks Rob for the debates on moderation), IMHO actually creates a meritocracy where the "best" (as judged by the /. audience karma criteria) ideas tend to surface and get more widely circulated. Thus more marginal and reflective opinions tend to be disseminated creating a richer and diverse base (still centered around the specialist niche though).
/.), especially when all contributors have an equal chance of submission. It has the advantage of allowing human nature to get on a soapbox and express itself without the fear of public speaking as well as a degree of anonymity. Also /. fits a void between the very large broadcast mediums catering to millions of people (TV, newspapers) and narrowcast mechanisms (chat rooms, bullitin boards, social clubs) which can only scale to hundreds. It will be interesting to see how /. will evolve over the next few years, especially as other groups start copying its format.
.... because it is neither rare nor well-done. :-)
For some people, access to any form of semi-intelligent debate is probably a significant improvement over the opinion pages of magazines and newspapers (the real competition to
ObJoke - why is TV a medium?
LL
What some companies tend to overlook is that the upfront costs is the sunk infrastructure but the long-term value is in the services on top. Unfortunately it does lead to a chicken and egg situation as nobody wants to seriously risk all that capital without having some guarenteed of payback down the track (can we say cable companies here?) but you can't create new services without experimental test-beds and cheap access. Thus the Canadian development is rather bold in opening up high-speed networks without obvious immediate commercial paybacks (and anyone mentioning video-on-demand should look at the distinct lack of consumer interest in trials so far) but then I suppose that is the definition of R&D.
:-(, what applications would benefit from greater bandwidth? Perhaps security monitoring (check on baby-sitters at home), or full-screen help-line type activities but one bottleneck I see is that the traditionalists who own the infrastructure may not always be the innovative providers, ie all the gear but no idea. Trying to "add value" by competing with customer applications which sit on top of the communications pipes (now that bandwidth and connectivity is becoming a commodity) is a rather interesting way of pissing off friends and creating enemies. The appearance of ISPs outside the traditional telcos indicate that it is not always easy to look outside traditional sources of cashflow.
/.ers think they mainly copy (and compete/crush) the plans of internet startups?
Apart from 500 interactive home shopping channels
So can anyone name any telco which has been particularly good at bringing new services to consumers? Or do
LL
Troy Baer wrote
Something to keep in mind about the the Origin 2000 (SGI's 128-256 CPU boxes) is that they're not SMP systems. They're ccNUMA machines, and a lot of the "ccNUMAness" (including cache coherence, I think) is handled largely by the hardware.
The point of ccNUMA is to minimise the cost of porting software from uni-processors. The crux of the matter is that it is non-trivial to adapt programs to run on multiple processors efficiently. The ideal is to have a single source tree, add extensions such as OpenMP, then recompile. Kernels are a different matter as they have to be closer to the hardware. It is still a royal pain to code to the wire and manually manipulate the cache and bus protocols but that is what is needed for maximum performance. Apart from special cases such as national defence codes, the commercial imperative is time-to-market which means a ccNUMA machine can address 95% of the issues at reasonable cost would be preferred.
I wouldn't be surprised if you could boot the MIPS version of Linux on (for instance) an Origin with little or no modification. I don't know how well it would scale, though.
As far as I'm aware (correct me if I'm wrong), the SGI port of Linux has so far concentrated on older systems such as Indys and patches for their VisualWorkstation. I suspect it will take a while (2-5 years?) for them to get to the stage of having Linux+IRIX SMP extensions running on their highly scalable systems. Cellular IRIX is a single system image which is different from the way Linux is designed. Perhaps one conceptual integration approach is to follow how RTLinux works in having a separate real-time kernel embedded within the full Linux system. Also there are other multiprocessor optimisations like processor affinity which might take a while to enter into the kernel. SGI staff may be very enthusiastic and dedicated but there is a lot of work involved which will take time.
In other words, nice PR for SGI but don't hold your breath.
LL
OK, time for Buzz Word City
For the type of applications Sun and IBM are targetting, raw processor speed becomes a marginal parameter in the design space. Instead, the smart cookies focus on the I/O-memory subsystems which becomes the performance bottleneck and cooling system which determines the overall reliability.
The StartFire is based on cross-bar switch technology spun off from Cray when it was purchased by SGI. As such it is analogous to a shared bus where processors (ie passengers) find it relatively easy to communicate between each other inside but the maximum capacity is somewhat fixed in that it costs the same whether a few passengers or fully populated. However, it has the advantage of being easy to migrate.
The original IBM SP series from which their current distributed memory design is derived can be compared with a truck convoy using CB radio (switched memory backplane). More flexibility in adding capacity means better efficiencies in matching load to task and thus better pricing (though at these lofty levels, the profit margins are MUCH heftier than PCs so there is a lot of gap for undercutting the competition).
The SGI ccNUMA (cache-coherent non-uniform memory access) can be compared with an articulated truck with multi-channel CDMA wireless giving a hypercube topology for fast node-node communications. Specifically designed for scalability and balanced I/O throughput, it commands a premium for its complexity and sophistication.
On a sliding spectrum of shared-distributed, the order would be Sun-SGI-IBM, but as processors speeds increase, both Sun and IBM are adopting ccNUMA techniques. Now a diversified transport system would require a judicious blend of each computer, matching the capabilities to each machine's strengths. Any claims of superiority are marketing delusions as you would not use a bus where a truck is needed. That is what supposedly CIOs get paid 6 figure incomes for deciding and service arms like IBM get fat consulting fees (any reader comments on their effectiveness?).
Note that raw technical considerations can be distorted to some extent by legacy concerns and availability of drivers (sys-admins). Personally I see the high-end server space get more competitive and cut-throat as souped up cars attempt to claim a slice of the action. However, some companies will need to hit with a clue-bat as the planned technological obsolescence of consumer items do not sit well with business (there are good reasons why you'd stick with old-style mainframes based on the principle if it ain't broke, don't meddle with it as more fixes are likely to add more bugs).
LL