Basically the real value of the Red Hat distribution is the rights to 90 days technical support (ie Linux wannabe hand-holding). This is turning software from a product to a service. Until the public is educated into thinking that the CD is more a time limited ticket valid fomr date of purchase rather than a unlimited end-user license, there will continue to be confusion. Perhaps RedHat could spend some of their IPO into running a series of educational ads to help distinguish what exactly it is that they are selling would help the general public to value the necessity of good technical support, a strong reputation to guarentee a warenty (funny how most vendors disclaim that their software actually does anything!) and bug-free software (ObjJoke, every program has at least 1 bug and can be reduced by 1 line, therefore all programs can be reduced to 1 line with a bug in it). The alternative mechanism is to clearly state on the CD a list of reputable resellers where a support guarentee is extended due to prior agreements.
Let's face it, if a company spends X million dollars installing OS Y for its corporate needs, then it will be very relunctant to change (you mean we thrown away $M X???). Thus, the general market can be segmented into Linux (hackers/engineers/generic), Windows (corporate), Java (embedded, components), Mac/BeOS (Education/Consumer) plus variants depending on specialised nichese (portables, real-time, etc). Thus a package can be developed on one platform, then the rights to port it can be sold off (e.g. Lokki with games). In this scenario everyone wins because it concentrates on your expertise, ports only occur for perceived successful applications reducing risk, and the market is charged what it can bear due to asymmetric information (ie you pay for not knowing how to grok the open source).
However, the point that probably hasn't escaped most people is that it is very hard becoming a large successful application specialist on the Windows platform as a certain comapany's desire for maximum profits has a tactic of embracing markets initially created by others. In an OpenSource environment, different business rules apply and it would be foolish of people to expect the ideals of OpenSource will translate unchanged into a CloseSource Environment.
If the point of a business is to make profits, then creating artificial scarcity is an obvious tactic. This can be seen in arenas such as the consumers diamond market which is actually over-supplied for the average wedding ring wannabe (just try reselling it to get an idea of its real value). Gaining sole ownership of a trademark is part of that strategy of limiting consumer choices as the brand (and associated perceived quality) will create a distinction with competitors.
One of the fundamental driving forces behind OpenSource which ESR didn't mention in his Magic Cauldron paper is the transition of software from a scarce item (requiring teams of highly trained mainframe programmers) to a commodity form (scripts/visual languages, etc). Prices are then determined by gross supply and demand which, with limitless digital replication, leads to infinite supply and thus theoretically zero price. In this market awash with junk mixed with gems, the problem is to stand out, hence it is easier to use software as a loss-leader for other non-portable services (e.g. AOL) than it is selling it as a distinct manufactured product. There are some software categories which do lead themselves to natural scarcity (e.g. highly specialised physics problems) but the economic payoff is so diffuse that there is no financial incentive to accelerate the time to market (but then science always thinks long-term as well as shoe-string).
In the long-term, trademarks are going to be a problem as it is so easy to create new ones. Hence the interest of big firms to buy up fast growing competitors and absorb their "name" before the wider public gains permanent "mindshare" and dilutes their purchasing draw (notice how one not-so-well-loved company brands every single product under their name?). While the "technical merits" of a piece of software may be obvious to the digital cognosti (ie. engineering geeks with too much time on their hands), the average public is not exposed to a wide range of implementations and therefore has no comparative point of reference to make an informed choice (plus the good-enough motive applies here). Trademarks is just one weapon in creating non-obvious barriers to pure competition which in a capitalistic society equates to near zero profits (anathema to a growing share price). Hence the desire of corporations not to sell software (perish the thought of the unwashed masses getting their hands on their pristine intellectual property:-) ), but to painstakingly dole out limited (and thus scarce and controllable) rights of use to maximise revenue streams. It's not the software but the control that companies are seeking.
OpenSource software developers have to understand the corporate mindset no matter how non-intuitive it appears because it has to coexist with the wider business ecosystem. No matter how desirable, free beer just does not magically appear.
Let's look forward into the future in 10 years time. With fast (and hopefully cheap) bandwidth, people will be offering software services over the net. Why bother patenting when any trade secrets can be embodied in the backend machine and effectively hidden from public view. If competitors can't duplicate it then by definition it must be "non-obvious" and therefore you can keep on charging a premium for that service.
The fundamental problem is that software is nearly pure information and thus spreads like a wave. Attempting to control it like a particle source is inherently doomed to failure as the rest of the world will just innovate around it. It doesn't help when convergence forces technology closer together and idiots attempt to patent obvious variations but using a slightly different mechanism. We really need higher quality barriers to recognise the really innovative ideas. Peer review and not patent offices are the only way to go and OpenSource will hopefully be better in the long term at really identifying talent.
Why would you even bother with cops and the internet when you can have an electronic tag and automatic bank deductions? This is already in place for a Toll-Road project in Melbourne Australia and Singapore CBD carzone control. It's only a matter of time before someone else uses them for speed gates and then for parking meters.
Despite a certain segment of the population's outcry of fascist control and outright paranoia of an overbearing brother, no matter how government is perceived, there is the very real issue of public transportation management. The no-limits on personal mobility and the pork-barrelling of highway construction has basically created cogestion effects which has led to road-rage, discrimination against alternative transportation (bikes, etc), and the usual hourly traffic stress. Given the increase in population and prosperity, can you imagine the problems if every person in India and China had the american average of 2.5 cars? After watching the driving habits of 3rd world countries, I'd be either scared out of my mind or investing in insurance companies.
This is going to be one example of how the information age will change society. If you know the position (GPS) and connectivity (IP address) of every single car, imagine how traffic flows can be managed. Someone has to pay for the roads and ongoing maintenance and an electronic user-pays scheme seems to make sense on economic grounds. With some luck, this will eliminate one layer of road/petrol taxes or at least make it more transparent although some people may debate the merits of substituting inefficient government for efficient private fine collection. Just make sure that the gas companies don't own the petrol stations AND the tollroads otherwise you might as well redefine the word monopoly and extortion.
The military has always been an important driver of technology (with roots from Enigma code cracking) as it is one of the few industry sectors that can ignore commercial returns on interest or practical coding considerations (can we say Ada here?). As such, the experimentation with Jini is a very important win for Sun as a successful rollout can establish a long-term commercial need for Java (once the US industrial-complex mandates the federal "need" for Java, imagine the profit nargins!). Compared with some of the negative press about Windows crippling an AEGIS (someone correct my memory if I'm wrong), if the perception of Jini/Java as a SECURE network interoperability framework takes off, then Sun are in clover. Of course whether the reality matches perception, only the coders and marketers know for certain. It is odd that in an industry with replacement cycles of less than 3 years, long-term credibility is still key to convincing people whether to adopt a given technology or not (or maybe it is not a contradiction after all when you consider the investment in training).
As a side note, the issue of perception came up in another thread, the Brazillian who worried about "unnecessary" violence in American cities then ignoring the "real" local killings, it does touch on a very important issue of perceived vs actual crime levels. Studies have shown that often a public's fear of events is way out of proportion with the actual frequency (thus the popularity of travel insurance when the airline industry has the best statistical passenger=mile safety record). Given the rather consistent message coming from Hollywood, it's not surprising that the rest of the world thinks the US consist of right-wing gun-totting ex-military thugs (die hard, etc) or mindless love-sick clueless dorks (sleepless in seattle etc). The reality is that the average person around the world would be pretty much indistinguihable as most societies converge to a norm given comparative social-economic levels. However the perception (or taken to the extreme - cultural mythology) is a very important driver for a lot of individual actions. Witness the US glorious history in westerns with the calvery opening up the frontier that has extended through 2 world wars and many regional conflicts. This has resulted in giving the US a global absolute advantage in organised violence, especially with the continual recruitment of citizen soldiers (a la Roman Empire style) based on an personal commitment to freedom. We should all salute the US model where the common solider is prepared to give up their lives so that their commander-in-chief can play around with interns. Now IMHO that is a strong military system based on its own internal memes which will persist and remain vigilant to its mission regardless of perceptions or reality of civilian leadership.
If people take the time to actually look up the technical details (see Microprocessor Report April 19, v15 i5) they'll probably get a more realistic idea of the capabilities instead of third-hand info from marketing flacks (not that I've got anything against salesdroids but slashdot is suppose to be targetted at a technical audience).
You can think (very broadly speaking) of the EmotionEngine as a R5K (like in SGI O2) coupled with 2 vector and 1 image unit. I wish people luck in developing partly asynchronous parallel/threaded algorithms that can get anywhere near the "peak" 6.2 Gflops. Also the I/O processor (ie the old Sony chip) only has space for 1 PCMCIA card on a 32 bit I/O bus. The other major constraint is total memory of 32 MBytes with no hard disk for swap space unless some bright spark can do some magic with an IEEE-1394 peripheral. As one wag noted, its easy to create a fast chip when you don't need to worry about memory hierarchies. At least it will have a lot of graphic functions built in (fog, sprites, particles, etc) so you can dazzle people with gee-whiz effects.
A go-kart is still a go-kart even if powered by a formula 1 engine. It will be a very useful and amusing console toy but don't expect it fufill your fantasties of having a supercomputer in your bedroom.
Why are companies bending over backwards to satisfy a rather opinionated crowd of nerds? Here lies the marketing line of logic.
In any mature market, there are several categories of buyers - innovators eagar to try new ideas (2%) - early adoptors integrated into social networks and are opinion leaders (5-10%) - early majority who deliberate for some time before adopting (30-40%) - late majority who tend to be sceptical and follow peer pressure (30-40%) - laggards - suspicious of new (10%)
The innovators are too flightly, instead companies try to get a critical mass of early adopters whether they are yuppies, TV stars or rag journals. IT companies are now realising that in the internet world, the balance of power has now shifted to the front-line staff holding together systems with (metaphorical) duct tape and hacked up code. These people with the deep technical knowledge of a company's IT assets are critical for that 24x7 operation which the business world is heading. If these people walk, you can kiss goodbye to your competitive edge for the next 6 months.
Hence gaining that critical early 10% penetration of a rather cynical set of experts is key to dominating the rest of the market. The early success of the PalmPilot was catalysed by a legion of hackers contributing code to the application base and thus getting past the bullshit filter, much less the "whats-in-it-for-me" barrier.
So though nerds will never grace the society times, they will progressively have more perks like trade shows and freebies (compare with doctors who prescribe drugs) showered on them. These are dangerous times because with this new-found respect, there is the danger of falling for the ego-stroking and forgetting professional responsibilities. A hacker's code of conduct should realise and emphasise the responsibilities that come with holding together societies electronic infrastructure, as well as educating the next generation of hackers.
So enjoy the little perks that make work worth living but be vigilant against decay into the seductive side.
> Surely the idea with clusters (and SMP, for that matter) is to use truckloads of cheap stuff to get a better result than one big expensive thing
Just because a car is cheap doesn't mean it is not going to cost you in terms of repairs and shoddy engineering down the road. What people forget is that price is a function of scale of economy and functionality. Would you base a car purchase purely on the number of cylinders and torque? Similarly the computer purchasing experts look at the overall system, the balance between components, cost of parts, availability of drivers and software, cost of ownership and human learning curve. Personally I think too many people without a clue are hoodwinked by fast talking salesdroids and bells and whistles. If you look at the hard evidence, you might even find that MIPS chips (in the Origins) give better sustained real-world application performance for a certain class of problems than even the highly touted Alphas. Also if you look at say SGI's O2, you find that it is designed for easy rack-mount maintenance. All this comes at a premium.
Sheesh... the PC makes a barely adequate car for roaming around the information backlanes but some people want big gruntly trucks for industrial computing. Companies pay real money for smarts who can evaluate the difference between the two.
One may ask the rather pertinant question for SGI, where are the profits in the computing industry?
If you compare it with other industries, two models usually work, high volume & thin margins or a specialist niche. Big players like IBM have discovered that the PC market is basically a money drain within their corporate structure. I think SGI are discovering the same phenomena with their smaller volume of sales. The big problem is that their traditional primary market (graphics) has suddenly shifted into the mainstream PC workstation range and until they can define/create a new market for their high-end wares, are effectively losing customers at the margin which is hurting them as they have longer product cycles than the current disposable PC (when you get a 486 chip in mobile phones/PDAs with shelf life of 12-18 months, you know something is crazy).
Personally I see them releasing people wholescale as a bad move as their core strength is in the knowledge that their staff carry around in optimising software for high-end kickass machines. SGI could provide a valuable service by hiring their people out to other companies in order to tune specific applications to go like the hammers of hell and provide a cost-effective hardware platform and software APIs to support them. In other words, loosen the reins before they jump ship. Without the skills and detailed know-how of coding to the wire, we will otherwise be left with bloatware relying on Moore's law to progress.
Economists have concluded there are several ways of making a guarenteed profit, rent-seeking monopoly (no guesses here as to the most successful culprits), efficiency differential between low and high cost providers, permanent gains due to unique technological breakthroughs and inelastic demand. As SGI is in no position to be a monopoly and their cost structure is inappropriate for a high-volume box shuffler, IMHO their long-term hope is to develop/morph new technologies, and spin them and associated staff off into as many subsidiary companies as quickly as possible to gain the capital appreciation. Given the rather dynamic, coopetitive and fluid nature of the computing market, I suspect it is easier to ally with siblings than with potential competitors.
Well, if not then the fate of DEC is illustrative of one path that SGI could face.
SGI and IBM have demonstrated their belief in OpenSource by releasing quite sizeable pieces of code and/or APIs. If and when Adobe does the same, then they might have some street credibility, especially in getting the community to port stuff like pdf converters to smaller platforms.
By the way, despite people's attachment to GIMP as a toy, Adobe have quite a prescence among commercial professional typesetting/publishing hardware that will never be replicated by the hacker community in the immediate future. This automated desktop-to-printing press market is worth some serious bikkies and I doubt whether anyone commercial vendor is going to abandon this gold mine. A serious stoush between Xerox, Cannon, and the document/image specialists might cause a market ruckus before things get standardised (funny how each digital camera manufacturer is pushing their favourite image format).
The real question is whether postscript or pdf is an appropriate file format to store digital documents/images for the long-term. TeX is probably the closest equivalent (even still used in scientific publishing) but it is not really a typesetting language. I won't mention Word which seems to come out differently on different configurations and framemaker SGML is rather complex for the average joe. XML is a simplified version designed for Web publishing but does it have the same richness suitable for paper publishing? It would be nice to be able to retrieve and view documents 50 years down the track after PCs have been replaced by whatever gee-whiz vr hype that will be the marketing ploy of the decade.:-)
One possibility is the Simple Document Format (http://www.mincom.com/mtr/sdf/) which separates somewhat the content and the formatting engines but I'm sure there's better alternatives. Any suggestions?
Rather than one of those dry historical anthologies, would it be better to base it on the collected real-life stories of Jon and the IETF? What are writings but a meeting of minds across space and time? When an author sits down and composes something, it's like having a conversation with a future soul. What is missing is the context, the idealism and passion of the times. Living during the French revolution is vastly different from reading about it. What would you want your peers to know about the celebration of the individual and challenge technical mastery?
Perhaps slashdot could keep an open archive for a few months to collect people's memories of the evolution of the internet, along with crystalised documents, with the aim of producing a cdcard (youanthologies know one of those shape CDs) to (yeah, cheesy but the rest of the population thinks a date is important) commemorate the new millenium?
We can then call it BE and AI for before and after internet:-)
To understand the move, you have to realise what business SGI is in. Their customer base is a) Big Data (Federal/Banks/Servers) b) Complex Graphics (Entertainment) c) Raw Science (Universities/Medical)
Finito.
Selling piddly graphics cards (that is compared with Infinite Reality Engine++ !!... drool) for a high volume, fad-driven market is most definitely not their market.
What they are doing with their Linux strategy (and to some extent ditto for BigBlueIBM) is creating a scorched earth policy by releasing source code in the low-end server/workstation market to prevent certain software/hardware companies (you know who) from cutting into their margins and preparing a migration path from their proprietary Unices to value-added Linux components/tuning. You can see it in their MIPS processor strategy, developing for high-end then migrating to the embedded market (guess what's in your routers and printers?) once the R&D dollars are written off.
Anything that is a distraction at this point in time when their CEO is still turning the company around is a waste of energy. Despite this forum's fascination with technology, you have to understand that companies are hard-nosed businesses with zero tolerance for sentimental gestures. You get paid money for doing the things that are hard and other people are willing to pay money for, not sexy 3D games which take up a disproportionate amount of space on freshmeat.
When you understand the difference between work and play (or are rich enough not to care), then you can pontificate to your heart's content.
ajdavis writes They should just release the whole damn MacOS X. Sure, some people would port it to cheaper Intel hardware, costing Apple revenue, but most people would run MOSX on G3/G4 hardware, and MOSX would kick butt even more if it were open-source. Its OSS development would be particularly fast since a) most OSS developers are already familiar with Mach & BSD, and b) this is so political.
I'd strongly disagree that Apple should release ALL of their MacOS. Their strength is in their user-oriented interface which is specifically engineered for (dare I say it?) non-nerds. The underlying kernel and OS functions are pretty much commodities with a bit of flash technology but their memeware (ie conceptual interface) is unique and should be leveraged for what it's worth (between zero and infinity depending on how they hype it). If their OS was completely freed, then it would be harder to set consistent standards and interaction modes. Quality or quantity? Your choice.
IBM (a hardware company with a few impressive forays into software) A few? Minor point but IBM has one of the largest software development team around (someone has to keep their corporate mainframe customers happy).
It makes sense in a way, having the Linux standard base, then companies such as Sun, SGI, IBM? including value-added components on top. If this scenario comes into play, there will be a above-line, below-line divide with everything below the line being free (and coincidently similar functionality to a certain unnamed OS). Companies are not stupid, after seeing what integrating browsers can do to the competition, they're going to return the favor in spades.
As for Java, it will be interesting to see how far Sun/IBM will push that barrow in their server centric platforms. The question that Linux has to face is
a) What modularity/extensions will be supported (kernel+application) in the future b) How do applications talk to each other in a coherent language (XML?) c) Can one increase complexity while maintaining stability?
I don't know about others but IMHO component technology is still relatively painful to use. Either that or I'm so far behind the learning curve all I can see is a brick wall to bash my head against.
It is good to see SGI taking a leadership role in advancing graphics and OpenSource for Linux. To expand a little bit on their roadmap (http://www.sgi.com/fahrenheit/support.html)
IRIS Performer IRIS Performer is primarily aimed at developers of real-time, multiprocessed, interactive graphics applications... While a number of developers currently using IRIS Performer will be able to migrate to the new Fahrenheit Scene Graph API, we do not foresee the first version of Fahrenheit meeting the needs of the entire marketplace served by IRIS Performer. That being said, SGI will continue to invest in enhancements to IRIS Performer in support of real-time applications and the Onyx2 marketplace for the foreseeable future.
I hope that the new Farenheit APIs can be massaged to retrofit the Performer APIs (though anything mixing multithreading, real-time and fast graphics is sure to be mind-numbing). The challenge for the developer community now is to work out higher-level control/interaction systems and interface them with OpenGL scene graphs. One interesting trend is the use of physics based worlds (see MathEngine http://www.mathengine.com/main.htm) as the need to cleanly separate the physical reality from the representation and display reality.
Given that game boxes such as Sony have recommended the use of Linux, it will be interesting to see whether general purpose PC-based boxes are relegated to development for single-tasked appliances or whether "convergence" means everything will be mixed up.
SGI do make some nice hardware. For example, their R10K supports hardware-based performance monitoring. I just hope their products remain price-competitive to justify the premium. If the computer industry can be compared with the car, then we have the consumers (Ford/Mazda, Intel, AMD) versus the specialists/industrial (Volvo, SGI, IBM). However, learning to drive in one should make it easier to cross over to the next level (no foreign instructions, custom gear boxes or lockin licenses). That's why I think we should applaud SGI for porting their APIs across and why it should benefit the computer industry as a whole.
> Jini development team at Sun, acting as "shepherds" for new projects and working groups
A rather interesting attitude towards future developers, so long as you stay within the nice little fenced-in commercially viable paddock available, you won't get eaten by those nasty wolves (or penguins) that are out there to steal your code and ideas:-{.
Does controlled chaos really work? Can they motivate week-end hobbyists to become mercenary developers, much less cynical ISVs? What are the motivating factors to encourage individuals to excel? All the greedy capitalists are forming startups, all the zealots are gnuing away, and the erratic geniuses are beavering away on the glory of Linux. Who's left to grab developer's mindshare from?
Given the complexity of software systems nowadays, I suspect the limiting factorfor growth is the learning curve and ease of development. It will be interesting in the medium term to see how well Java/Jini ranks against Windows and Linux as the developer's playground.
In a technical domain, the number of hours worked may not correlate to results as much as people like. Let's face it, if you take the economist view that as rational actors, individuals maximise their income utility then it makes sense to put in overtime in expectation of an increase in the ownership value (don't ask what happens to options in a falling share market though!) For those who don't wish to place a portion of their salary at risk, a steady income to support family activities is preferred. However, one should ask the careful question of exactly what is being produced? Is it accomplishing the desired goal at a reasonable rate of progress? Spending hours peddling around in redundant debugging cycles is a waste of energy. Also human nature indicates that there is a natural plateau of performance beyond which is difficult to penetrate. Working more than 18 hours a day in a burst mode cna lead to burnout (hence the return to simplicity that some managers are advocating).
History has shown that great leaders inspire ordinary people to achieve the extraordinary. Natural people skills are just as rare (or even more so) as creative programming skills. One shortcut is the zealotry of religion or enforced groupthink (can we say mindless corporate culture here?). However, this is self-defeating as there is a limit to people suspectible to self-delusion (no flames on actual religion please, this is just a comment on the number of gullible idiots out there). In the long-term, only bottom-up grassroots activities have any chance of sustainability as successful thoughts/habits become engrained into mainstream.
Unforunately we don't have good measures of soft factors such as human creativity and coding productivity. Absolute hours worked is not always useful as quantity != quality hours. All companies can really do is create a framework and culture that encourages the results that is desired and leave it up to individual talent and capabilities (while somehow eliminating obstructive PHBs).
Didn't RMS indicate that code documentation would come under GPL? What about paper documentation? Where should the distinction be drawn between free software and stuff that is worth paying for?
1) copying/distribution OK but must keep GPL license 2) modification/derivative works iff acknowledged + nondiscriminatory relicensing & redistribution 3) binaries OK iff source provided directly or indirectly 4) even if you screw up, your audience can do what they want so long as they don't screw up 5) above terms accepted if you modify/distribute source 6) can't deny others the same rights 7) if patented, can't distribute 8) if distribution blocked, allow geographical exclusion
Therefore if you generate the output through a web interface, you are not redistributing the original source, just output (although it would be smart to reference the original software). Hmmmm... does this mean an unencrypted command can be automatically sent to a server outside the US, encrypted code added, then forwarded to a third party without breaking any export laws?
It really puzzles me how people wish to skimp on the hardware then set themselves up for later hassles and risk of (expensive) failure.
Lets take a look at a baseline video server from SGI, Origin 200 with MediaBase with say 100 seat configuration, web management tools + 200 Gbyte FibreChannel and network bits would cost about $50K minimum upfront (guesstimate here based on educational discounts and extrapolation of bits and pieces we've purchased over the years) + 10% maintenance/year. Extra for their FailOver system.
OK, now the hand-rolled Linux version. You need to look for a) streaming software (Darwin?) for multiprocessor b) decent high-end file system (port SGI XFS?) c) tuning the sucker for the best SCSI and network parameters d) video library management software (none as yet, perhaps someone port SGI OpenVault?) e) system management to monitor the whole thing
Lets assume you've got a collection of genius hackers at 1 man-years worth at each task, working for nothing except glory, you can probably get it done for $20K and 5 man-years worth of pizzas and coke.
Cheap at that price.
Rule #2 - If you don't know what you're doing, make sure you get damn good advice from people who've done it before.
I'm not sure how much money an ISP spends on hardware but I suspect that the box is probably not their biggest cost (except for the big eCommerce hosting sites). I'd like to see a breakdown but I suspect salaries, communications, consumables (disks, software etc) then CPU in that order. The advantage of Linux over AIX is that there is a significantly larger portion of the population familiar with Linux, hence the learning curve reduces and thus hireability of talent improves. Also, it would be nice to get people who have a clue about balanced systems, especially I/O bottlenecks!
As for the number of systems out there, off the top of my head
- SGI offer nice balanced systems with good real-application/price performance and from memory one outfit also offers mips-based linux systems. - HP with their PA-RISC have solid engineering, though probable a little pricey for just ISP work, - Compaq Alphas are supposed to be good file/web servers - Your standard Sun Sparc covers the low-mid range server market quite nicely - El Cheapo Intel box - Apple servers
For a large ISP, rock-solid reliability, good service and security, decent software, plus someone to sue in case things goes really wrong are probably more important than the actual price of a box.
The properties of a unit of current are related to its functions. 1) long-term store of value 2) convenient medium of exchange 3) measure of economic goods and services
Gold has some nice properties 1) extremely stable and durable 2) divisible into smaller subunits 3) considered a luxury item and thus independent of goods and services.
The history of money is rather intriguing as historically, people have used different mediums for exchange (eg. roman solders used salt from which we've derived salary). The reason why gold has traditionally been used is that it is very hard for governments to dilute it's value. However, relying on it has rather negative consequences when everybody is using it as there are physical costs and circulation (ie liquidity) problems. Oh, plus the rather inconvient fact that the world's largest suppluers are in South Africa and Russia so whether you are leftist or rightist, you're sure to alienate half your population by depending on it as the sole measure of your economic future.
The modern fiat money system is based on trust (e.g. social security) backed by the laws of the host country. That is why the feds get rather irritated with things like counterfeiting and trying to escape the tax system as it dilutes the trust and ultimate value of the currency. Of course, they're also hoping that nobody points out the clothesless emporor in by inflating the currency through increasing the M3 supply, they automatically devalue the currency. Considering the US owes nearly 6 trillion to the rest of the world and, as a global reserve currency, is not on the gold standard, they can theoretically inflate away the entire debt through accounting tricks. (one reason why the Euro has appeared and indebted countries are so pissed off).
Governments and banks would dearly like to move people into an electronic currency (consider it as the pure economic laws without the paper or cheque bits fiddly bits) as it eliminates the horded value locked up in circulating paper (one reason why Russia is so bad is that all the rubles have been tranformed into illiquid assets and horded, thus denying their use as a circulating unit of exchange) and allows financial firms to cut handling costs (not to mention lend on the stock exchange for the hours in the day you're not using it).
Modern capitalistic countries are already seeing this with the move towards a credit-based society. Of course, if the system screws up (as the Singapore case shows), the population gets rather annoyed as nothing brings out the fear of mob hysteria than losing your shirt. However, no matter what currency you use, it still has to satisfy the functions of a universal store of value, common medium of exchange and independent unit of pricing.
LL
PS As an interesting academic exercise, you can figure out the instrinsic value of software by creating synthetic measures of its value, much like you can trade Hollywood futures on the future value of stars and films.
Following up on someone else's comment, the specs for MPEG4 were released last year with some reference code for structured audio.
As for being innovative, I would be careful to distinguish between invention, incremental improvement and radically new. Universities are more likely to focus on the radically new, especially exotic languages which tests out specific ideas that eventually get incorporated into mainstream (orthogonal persistance is one example coming through the current pipeline). Application developers focus on the inventions that make like simpler, creating the killer apps of the day (VisiCalc for spreadsheets), followed by later imitators and refinements. It generally takes something about 10 years in moving major technologies from university to mainstream (assuming that anyone is interested:-( ).
Considering that most people can't live on air for 10 years, Linux hackers usually end up with (hopefully) decent jobs and play around in their spare time. The amazing factor is that the relatively recent arrival of the web which allows many slices of people's spare time to accumulate into solid products, especially when they have the time luxury to reengineer a clean architecture.
Commercial vendors on the other hand have to keep in mind certain things like pleasing the stock-holders whose gracious generosity has lent them some trifling few billion to accelerate development and hype their products. Time is not a luxury and corporate secrecy (due to requirements for patenting) is an absolute. This leads to a rather closed worldview in which old techniques applied to a different setting is interpreted as "innovation". In my book truely innovative companies are those that have creately completely new sectors of the computer industry (Adobe for desktop publishing, SGI for OpenGL, AutoDesk for CAD) that wouldn't have existed otherwise.
The big problem that the Linux crowd has to address is to separate proprietary from open from expensive. Some code by it's very nature is expensive to develop (safety/fault tolerant stuff because of extensive testing). Other stuff like compilers are needed in the intermediate stages before creating the sale goods to consumers and business. Despite what people think, there is no free beer (unless you're prepared to go out and plant the crops and brew it yourself).
The power and abstraction of software is both its strength and its Ancilles heel. Any thought can only be expressed using the current available languages and concepts. We have evolved from assembly-level bit twiddling, to macros, imperative languages, functional, object-oriented and now high levle scripting/glue systems. Yet each level of language abstraction invalidates any patented technique below it. If someone patents the use of the letter "E" and charges monopoly rents, think how painful communications would be.
The fundamental problem is that it is impossible to prevent other people having the same idea. Given that there are 5 billion people all trying to survive, sooner or later, everything will be independently rediscovered or reinvented. Should being the first to discover automatically extinguish the rights of others who wish to develop the idea further? One can think of an example of a certain plant MNC who has been awared the exclusive rights to bioengineer cotton fibre and has basically done nothing as yet but refuses to allow anyone else encroach on a potentially lucrative market.
Business nowadays has less to do with innovation than issues of control in a market nichce and thus extract long-term profits. Such is the power of the almighty dollar. Idealism my be good for the pampered classes but it doesn't pay for your kid's medical bills.
Slashdot Mirror
1 Red Hat CD =
$1 CD + $2 download time + Right to support
Basically the real value of the Red Hat distribution is the rights to 90 days technical support (ie Linux wannabe hand-holding). This is turning software from a product to a service. Until the public is educated into thinking that the CD is more a time limited ticket valid fomr date of purchase rather than a unlimited end-user license, there will continue to be confusion. Perhaps RedHat could spend some of their IPO into running a series of educational ads to help distinguish what exactly it is that they are selling would help the general public to value the necessity of good technical support, a strong reputation to guarentee a warenty (funny how most vendors disclaim that their software actually does anything!) and bug-free software (ObjJoke, every program has at least 1 bug and can be reduced by 1 line, therefore all programs can be reduced to 1 line with a bug in it). The alternative mechanism is to clearly state on the CD a list of reputable resellers where a support guarentee is extended due to prior agreements.
LL
Let's face it, if a company spends X million dollars installing OS Y for its corporate needs, then it will be very relunctant to change (you mean we thrown away $M X???). Thus, the general market can be segmented into Linux (hackers/engineers/generic), Windows (corporate), Java (embedded, components), Mac/BeOS (Education/Consumer) plus variants depending on specialised nichese (portables, real-time, etc). Thus a package can be developed on one platform, then the rights to port it can be sold off (e.g. Lokki with games). In this scenario everyone wins because it concentrates on your expertise, ports only occur for perceived successful applications reducing risk, and the market is charged what it can bear due to asymmetric information (ie you pay for not knowing how to grok the open source).
However, the point that probably hasn't escaped most people is that it is very hard becoming a large successful application specialist on the Windows platform as a certain comapany's desire for maximum profits has a tactic of embracing markets initially created by others. In an OpenSource environment, different business rules apply and it would be foolish of people to expect the ideals of OpenSource will translate unchanged into a CloseSource Environment.
LL
If the point of a business is to make profits, then creating artificial scarcity is an obvious tactic. This can be seen in arenas such as the consumers diamond market which is actually over-supplied for the average wedding ring wannabe (just try reselling it to get an idea of its real value). Gaining sole ownership of a trademark is part of that strategy of limiting consumer choices as the brand (and associated perceived quality) will create a distinction with competitors.
:-) ), but to painstakingly dole out limited (and thus scarce and controllable) rights of use to maximise revenue streams. It's not the software but the control that companies are seeking.
One of the fundamental driving forces behind OpenSource which ESR didn't mention in his Magic Cauldron paper is the transition of software from a scarce item (requiring teams of highly trained mainframe programmers) to a commodity form (scripts/visual languages, etc). Prices are then determined by gross supply and demand which, with limitless digital replication, leads to infinite supply and thus theoretically zero price. In this market awash with junk mixed with gems, the problem is to stand out, hence it is easier to use software as a loss-leader for other non-portable services (e.g. AOL) than it is selling it as a distinct manufactured product. There are some software categories which do lead themselves to natural scarcity (e.g. highly specialised physics problems) but the economic payoff is so diffuse that there is no financial incentive to accelerate the time to market (but then science always thinks long-term as well as shoe-string).
In the long-term, trademarks are going to be a problem as it is so easy to create new ones. Hence the interest of big firms to buy up fast growing competitors and absorb their "name" before the wider public gains permanent "mindshare" and dilutes their purchasing draw (notice how one not-so-well-loved company brands every single product under their name?). While the "technical merits" of a piece of software may be obvious to the digital cognosti (ie. engineering geeks with too much time on their hands), the average public is not exposed to a wide range of implementations and therefore has no comparative point of reference to make an informed choice (plus the good-enough motive applies here). Trademarks is just one weapon in creating non-obvious barriers to pure competition which in a capitalistic society equates to near zero profits (anathema to a growing share price). Hence the desire of corporations not to sell software (perish the thought of the unwashed masses getting their hands on their pristine intellectual property
OpenSource software developers have to understand the corporate mindset no matter how non-intuitive it appears because it has to coexist with the wider business ecosystem. No matter how desirable, free beer just does not magically appear.
LL
Let's look forward into the future in 10 years time. With fast (and hopefully cheap) bandwidth, people will be offering software services over the net. Why bother patenting when any trade secrets can be embodied in the backend machine and effectively hidden from public view. If competitors can't duplicate it then by definition it must be "non-obvious" and therefore you can keep on charging a premium for that service.
The fundamental problem is that software is nearly pure information and thus spreads like a wave. Attempting to control it like a particle source is inherently doomed to failure as the rest of the world will just innovate around it. It doesn't help when convergence forces technology closer together and idiots attempt to patent obvious variations but using a slightly different mechanism. We really need higher quality barriers to recognise the really innovative ideas. Peer review and not patent offices are the only way to go and OpenSource will hopefully be better in the long term at really identifying talent.
LL
Why would you even bother with cops and the internet when you can have an electronic tag and automatic bank deductions? This is already in place for a Toll-Road project in Melbourne Australia and Singapore CBD carzone control. It's only a matter of time before someone else uses them for speed gates and then for parking meters.
Despite a certain segment of the population's outcry of fascist control and outright paranoia of an overbearing brother, no matter how government is perceived, there is the very real issue of public transportation management. The no-limits on personal mobility and the pork-barrelling of highway construction has basically created cogestion effects which has led to road-rage, discrimination against alternative transportation (bikes, etc), and the usual hourly traffic stress. Given the increase in population and prosperity, can you imagine the problems if every person in India and China had the american average of 2.5 cars? After watching the driving habits of 3rd world countries, I'd be either scared out of my mind or investing in insurance companies.
This is going to be one example of how the information age will change society. If you know the position (GPS) and connectivity (IP address) of every single car, imagine how traffic flows can be managed. Someone has to pay for the roads and ongoing maintenance and an electronic user-pays scheme seems to make sense on economic grounds. With some luck, this will eliminate one layer of road/petrol taxes or at least make it more transparent although some people may debate the merits of substituting inefficient government for efficient private fine collection. Just make sure that the gas companies don't own the petrol stations AND the tollroads otherwise you might as well redefine the word monopoly and extortion.
LL
As a side note, the issue of perception came up in another thread, the Brazillian who worried about "unnecessary" violence in American cities then ignoring the "real" local killings, it does touch on a very important issue of perceived vs actual crime levels. Studies have shown that often a public's fear of events is way out of proportion with the actual frequency (thus the popularity of travel insurance when the airline industry has the best statistical passenger=mile safety record). Given the rather consistent message coming from Hollywood, it's not surprising that the rest of the world thinks the US consist of right-wing gun-totting ex-military thugs (die hard, etc) or mindless love-sick clueless dorks (sleepless in seattle etc). The reality is that the average person around the world would be pretty much indistinguihable as most societies converge to a norm given comparative social-economic levels. However the perception (or taken to the extreme - cultural mythology) is a very important driver for a lot of individual actions. Witness the US glorious history in westerns with the calvery opening up the frontier that has extended through 2 world wars and many regional conflicts. This has resulted in giving the US a global absolute advantage in organised violence, especially with the continual recruitment of citizen soldiers (a la Roman Empire style) based on an personal commitment to freedom. We should all salute the US model where the common solider is prepared to give up their lives so that their commander-in-chief can play around with interns. Now IMHO that is a strong military system based on its own internal memes which will persist and remain vigilant to its mission regardless of perceptions or reality of civilian leadership.
LL
If people take the time to actually look up the technical details (see Microprocessor Report April 19, v15 i5) they'll probably get a more realistic idea of the capabilities instead of third-hand info from marketing flacks (not that I've got anything against salesdroids but slashdot is suppose to be targetted at a technical audience).
You can think (very broadly speaking) of the EmotionEngine as a R5K (like in SGI O2) coupled with 2 vector and 1 image unit. I wish people luck in developing partly asynchronous parallel/threaded algorithms that can get anywhere near the "peak" 6.2 Gflops. Also the I/O processor (ie the old Sony chip) only has space for 1 PCMCIA card on a 32 bit I/O bus. The other major constraint is total memory of 32 MBytes with no hard disk for swap space unless some bright spark can do some magic with an IEEE-1394 peripheral. As one wag noted, its easy to create a fast chip when you don't need to worry about memory hierarchies. At least it will have a lot of graphic functions built in (fog, sprites, particles, etc) so you can dazzle people with gee-whiz effects.
A go-kart is still a go-kart even if powered by a formula 1 engine. It will be a very useful and amusing console toy but don't expect it fufill your fantasties of having a supercomputer in your bedroom.
LL
Why are companies bending over backwards to satisfy a rather opinionated crowd of nerds? Here lies the marketing line of logic.
In any mature market, there are several categories of buyers
- innovators eagar to try new ideas (2%)
- early adoptors integrated into social networks and are opinion leaders (5-10%)
- early majority who deliberate for some time before adopting (30-40%)
- late majority who tend to be sceptical and follow peer pressure (30-40%)
- laggards - suspicious of new (10%)
The innovators are too flightly, instead companies try to get a critical mass of early adopters whether they are yuppies, TV stars or rag journals. IT companies are now realising that in the internet world, the balance of power has now shifted to the front-line staff holding together systems with (metaphorical) duct tape and hacked up code. These people with the deep technical knowledge of a company's IT assets are critical for that 24x7 operation which the business world is heading. If these people walk, you can kiss goodbye to your competitive edge for the next 6 months.
Hence gaining that critical early 10% penetration of a rather cynical set of experts is key to dominating the rest of the market. The early success of the PalmPilot was catalysed by a legion of hackers contributing code to the application base and thus getting past the bullshit filter, much less the "whats-in-it-for-me" barrier.
So though nerds will never grace the society times, they will progressively have more perks like trade shows and freebies (compare with doctors who prescribe drugs) showered on them. These are dangerous times because with this new-found respect, there is the danger of falling for the ego-stroking and forgetting professional responsibilities. A hacker's code of conduct should realise and emphasise the responsibilities that come with holding together societies electronic infrastructure, as well as educating the next generation of hackers.
So enjoy the little perks that make work worth living but be vigilant against decay into the seductive side.
LL
Just because a car is cheap doesn't mean it is not going to cost you in terms of repairs and shoddy engineering down the road. What people forget is that price is a function of scale of economy and functionality. Would you base a car purchase purely on the number of cylinders and torque? Similarly the computer purchasing experts look at the overall system, the balance between components, cost of parts, availability of drivers and software, cost of ownership and human learning curve. Personally I think too many people without a clue are hoodwinked by fast talking salesdroids and bells and whistles. If you look at the hard evidence, you might even find that MIPS chips (in the Origins) give better sustained real-world application performance for a certain class of problems than even the highly touted Alphas. Also if you look at say SGI's O2, you find that it is designed for easy rack-mount maintenance. All this comes at a premium.
Sheesh
LL
One may ask the rather pertinant question for SGI, where are the profits in the computing industry?
If you compare it with other industries, two models usually work, high volume & thin margins or a specialist niche. Big players like IBM have discovered that the PC market is basically a money drain within their corporate structure. I think SGI are discovering the same phenomena with their smaller volume of sales. The big problem is that their traditional primary market (graphics) has suddenly shifted into the mainstream PC workstation range and until they can define/create a new market for their high-end wares, are effectively losing customers at the margin which is hurting them as they have longer product cycles than the current disposable PC (when you get a 486 chip in mobile phones/PDAs with shelf life of 12-18 months, you know something is crazy).
Personally I see them releasing people wholescale as a bad move as their core strength is in the knowledge that their staff carry around in optimising software for high-end kickass machines. SGI could provide a valuable service by hiring their people out to other companies in order to tune specific applications to go like the hammers of hell and provide a cost-effective hardware platform and software APIs to support them. In other words, loosen the reins before they jump ship. Without the skills and detailed know-how of coding to the wire, we will otherwise be left with bloatware relying on Moore's law to progress.
Economists have concluded there are several ways of making a guarenteed profit, rent-seeking monopoly (no guesses here as to the most successful culprits), efficiency differential between low and high cost providers, permanent gains due to unique technological breakthroughs and inelastic demand. As SGI is in no position to be a monopoly and their cost structure is inappropriate for a high-volume box shuffler, IMHO their long-term hope is to develop/morph new technologies, and spin them and associated staff off into as many subsidiary companies as quickly as possible to gain the capital appreciation. Given the rather dynamic, coopetitive and fluid nature of the computing market, I suspect it is easier to ally with siblings than with potential competitors.
Well, if not then the fate of DEC is illustrative of one path that SGI could face.
LL
Actions speak louder than words ....
:-)
SGI and IBM have demonstrated their belief in OpenSource by releasing quite sizeable pieces of code and/or APIs. If and when Adobe does the same, then they might have some street credibility, especially in getting the community to port stuff like pdf converters to smaller platforms.
By the way, despite people's attachment to GIMP as a toy, Adobe have quite a prescence among commercial professional typesetting/publishing hardware that will never be replicated by the hacker community in the immediate future. This automated desktop-to-printing press market is worth some serious bikkies and I doubt whether anyone commercial vendor is going to abandon this gold mine. A serious stoush between Xerox, Cannon, and the document/image specialists might cause a market ruckus before things get standardised (funny how each digital camera manufacturer is pushing their favourite image format).
The real question is whether postscript or pdf is an appropriate file format to store digital documents/images for the long-term. TeX is probably the closest equivalent (even still used in scientific publishing) but it is not really a typesetting language. I won't mention Word which seems to come out differently on different configurations and framemaker SGML is rather complex for the average joe. XML is a simplified version designed for Web publishing but does it have the same richness suitable for paper publishing? It would be nice to be able to retrieve and view documents 50 years down the track after PCs have been replaced by whatever gee-whiz vr hype that will be the marketing ploy of the decade.
One possibility is the Simple Document Format
(http://www.mincom.com/mtr/sdf/) which separates somewhat the content and the formatting engines but I'm sure there's better alternatives. Any suggestions?
LL
Rather than one of those dry historical anthologies, would it be better to base it on the collected real-life stories of Jon and the IETF? What are writings but a meeting of minds across space and time? When an author sits down and composes something, it's like having a conversation with a future soul. What is missing is the context, the idealism and passion of the times. Living during the French revolution is vastly different from reading about it. What would you want your peers to know about the celebration of the individual and challenge technical mastery?
:-)
Perhaps slashdot could keep an open archive for a few months to collect people's memories of the evolution of the internet, along with crystalised documents, with the aim of producing a cdcard (youanthologies know one of those shape CDs) to (yeah, cheesy but the rest of the population thinks a date is important) commemorate the new millenium?
We can then call it BE and AI for before and after internet
LL
To understand the move, you have to realise what business SGI is in. Their customer base is
... drool) for a high volume, fad-driven market is most definitely not their market.
a) Big Data (Federal/Banks/Servers)
b) Complex Graphics (Entertainment)
c) Raw Science (Universities/Medical)
Finito.
Selling piddly graphics cards (that is compared with Infinite Reality Engine++ !!
What they are doing with their Linux strategy (and to some extent ditto for BigBlueIBM) is creating a scorched earth policy by releasing source code in the low-end server/workstation market to prevent certain software/hardware companies (you know who) from cutting into their margins and preparing a migration path from their proprietary Unices to value-added Linux components/tuning. You can see it in their MIPS processor strategy, developing for high-end then migrating to the embedded market (guess what's in your routers and printers?) once the R&D dollars are written off.
Anything that is a distraction at this point in time when their CEO is still turning the company around is a waste of energy. Despite this forum's fascination with technology, you have to understand that companies are hard-nosed businesses with zero tolerance for sentimental gestures. You get paid money for doing the things that are hard and other people are willing to pay money for, not sexy 3D games which take up a disproportionate amount of space on freshmeat.
When you understand the difference between work and play (or are rich enough not to care), then you can pontificate to your heart's content.
LL
ajdavis writes
They should just release the whole damn MacOS X. Sure, some people would port it to cheaper Intel hardware, costing Apple revenue, but most people would run MOSX on G3/G4 hardware, and MOSX would kick butt even more if it were open-source. Its OSS development would be particularly fast since a) most OSS developers are already familiar with Mach & BSD, and b) this is so political.
I'd strongly disagree that Apple should release ALL of their MacOS. Their strength is in their user-oriented interface which is specifically engineered for (dare I say it?) non-nerds. The underlying kernel and OS functions are pretty much commodities with a bit of flash technology but their memeware (ie conceptual interface) is unique and should be leveraged for what it's worth (between zero and infinity depending on how they hype it). If their OS was completely freed, then it would be harder to set consistent standards and interaction modes. Quality or quantity? Your choice.
IBM (a hardware company with a few impressive forays into software)
A few? Minor point but IBM has one of the largest software development team around (someone has to keep their corporate mainframe customers happy).
LL
It makes sense in a way, having the Linux standard base, then companies such as Sun, SGI, IBM? including value-added components on top. If this scenario comes into play, there will be a above-line, below-line divide with everything below the line being free (and coincidently similar functionality to a certain unnamed OS). Companies are not stupid, after seeing what integrating browsers can do to the competition, they're going to return the favor in spades.
As for Java, it will be interesting to see how far Sun/IBM will push that barrow in their server centric platforms. The question that Linux has to face is
a) What modularity/extensions will be supported (kernel+application) in the future
b) How do applications talk to each other in a coherent language (XML?)
c) Can one increase complexity while maintaining stability?
I don't know about others but IMHO component technology is still relatively painful to use. Either that or I'm so far behind the learning curve all I can see is a brick wall to bash my head against.
LL
It is good to see SGI taking a leadership role in advancing graphics and OpenSource for Linux. To expand a little bit on their roadmap (http://www.sgi.com/fahrenheit/support.html)
... While a number of developers currently using IRIS Performer will be able to migrate to the new Fahrenheit Scene Graph API, we do not foresee the first version of Fahrenheit meeting the needs of the entire marketplace served by IRIS Performer. That being said, SGI will continue to invest in enhancements to IRIS Performer in support of real-time applications and the Onyx2 marketplace for the foreseeable future.
IRIS Performer
IRIS Performer is primarily aimed at developers of real-time, multiprocessed, interactive graphics applications
I hope that the new Farenheit APIs can be massaged to retrofit the Performer APIs (though anything mixing multithreading, real-time and fast graphics is sure to be mind-numbing). The challenge for the developer community now is to work out higher-level control/interaction systems and interface them with OpenGL scene graphs. One interesting trend is the use of physics based worlds (see MathEngine http://www.mathengine.com/main.htm) as the need to cleanly separate the physical reality from the representation and display reality.
Given that game boxes such as Sony have recommended the use of Linux, it will be interesting to see whether general purpose PC-based boxes are relegated to development for single-tasked appliances or whether "convergence" means everything will be mixed up.
SGI do make some nice hardware. For example, their R10K supports hardware-based performance monitoring. I just hope their products remain price-competitive to justify the premium. If the computer industry can be compared with the car, then we have the consumers (Ford/Mazda, Intel, AMD) versus the specialists/industrial (Volvo, SGI, IBM). However, learning to drive in one should make it easier to cross over to the next level (no foreign instructions, custom gear boxes or lockin licenses). That's why I think we should applaud SGI for porting their APIs across and why it should benefit the computer industry as a whole.
LL
> Jini development team at Sun, acting as "shepherds" for new projects and working groups
:-{.
A rather interesting attitude towards future developers, so long as you stay within the nice little fenced-in commercially viable paddock available, you won't get eaten by those nasty wolves (or penguins) that are out there to steal
your code and ideas
Does controlled chaos really work? Can they motivate week-end hobbyists to become mercenary developers, much less cynical ISVs? What are the motivating factors to encourage individuals to excel? All the greedy capitalists are forming startups, all the zealots are gnuing away, and the erratic geniuses are beavering away on the glory of Linux. Who's left to grab developer's mindshare from?
Given the complexity of software systems nowadays, I suspect the limiting factorfor growth is the learning curve and ease of development. It will be interesting in the medium term to see how well Java/Jini ranks against Windows and Linux as the developer's playground.
LL
In a technical domain, the number of hours worked may not correlate to results as much as people like. Let's face it, if you take the economist view that as rational actors, individuals maximise their income utility then it makes sense to put in overtime in expectation of an increase in the ownership value (don't ask what happens to options in a falling share market though!) For those who don't wish to place a portion of their salary at risk, a steady income to support family activities is preferred. However, one should ask the careful question of exactly what is being produced? Is it accomplishing the desired goal at a reasonable rate of progress? Spending hours peddling around in redundant debugging cycles is a waste of energy. Also human nature indicates that there is a natural plateau of performance beyond which is difficult to penetrate. Working more than 18 hours a day in a burst mode cna lead to burnout (hence the return to simplicity that some managers are advocating).
History has shown that great leaders inspire ordinary people to achieve the extraordinary. Natural people skills are just as rare (or even more so) as creative programming skills. One shortcut is the zealotry of religion or enforced groupthink (can we say mindless corporate culture here?). However, this is self-defeating as there is a limit to people suspectible to self-delusion (no flames on actual religion please, this is just a comment on the number of gullible idiots out there). In the long-term, only bottom-up grassroots activities have any chance of sustainability as successful thoughts/habits become engrained into mainstream.
Unforunately we don't have good measures of soft factors such as human creativity and coding productivity. Absolute hours worked is not always useful as quantity != quality hours. All companies can really do is create a framework and culture that encourages the results that is desired and leave it up to individual talent and capabilities (while somehow eliminating obstructive PHBs).
LL
What's the fine line between code and data?
Didn't RMS indicate that code documentation would come under GPL? What about paper documentation? Where should the distinction be drawn between free software and stuff that is worth paying for?
LL
GNU copyleft covers
... does this mean an unencrypted command can be automatically sent to a server outside the US, encrypted code added, then forwarded to a third party without breaking any export laws?
1) copying/distribution OK but must keep GPL license
2) modification/derivative works iff acknowledged + nondiscriminatory relicensing & redistribution
3) binaries OK iff source provided directly or indirectly
4) even if you screw up, your audience can do what they want so long as they don't screw up
5) above terms accepted if you modify/distribute source
6) can't deny others the same rights
7) if patented, can't distribute
8) if distribution blocked, allow geographical exclusion
Therefore if you generate the output through a web interface, you are not redistributing the original source, just output (although it would be smart to reference the original software). Hmmmm
What a wierd world we live in.
LL
Rule #1 - what is the limiting factor?
Bandwidth? Machine costs? or your time?
It really puzzles me how people wish to skimp on the hardware then set themselves up for later hassles and risk of (expensive) failure.
Lets take a look at a baseline video server from SGI, Origin 200 with MediaBase with say 100 seat configuration, web management tools + 200 Gbyte FibreChannel and network bits would cost about $50K minimum upfront (guesstimate here based on educational discounts and extrapolation of bits and pieces we've purchased over the years) + 10% maintenance/year. Extra for their FailOver system.
OK, now the hand-rolled Linux version. You need to look for
a) streaming software (Darwin?) for multiprocessor
b) decent high-end file system (port SGI XFS?)
c) tuning the sucker for the best SCSI and network parameters
d) video library management software (none as yet, perhaps someone port SGI OpenVault?)
e) system management to monitor the whole thing
Lets assume you've got a collection of genius hackers at 1 man-years worth at each task, working for nothing except glory, you can probably get it done for $20K and 5 man-years worth of pizzas and coke.
Cheap at that price.
Rule #2 - If you don't know what you're doing, make sure you get damn good advice from people who've done it before.
Rule #3 - You pay peanuts, you get monkeys.
LL
I'm not sure how much money an ISP spends on hardware but I suspect that the box is probably not their biggest cost (except for the big eCommerce hosting sites). I'd like to see a breakdown but I suspect salaries, communications, consumables (disks, software etc) then CPU in that order. The advantage of Linux over AIX is that there is a significantly larger portion of the population familiar with Linux, hence the learning curve reduces and thus hireability of talent improves. Also, it would be nice to get people who have a clue about balanced systems, especially I/O bottlenecks!
As for the number of systems out there, off the top of my head
- SGI offer nice balanced systems with good real-application/price performance and from memory one outfit also offers mips-based linux systems.
- HP with their PA-RISC have solid engineering, though probable a little pricey for just ISP work,
- Compaq Alphas are supposed to be good file/web servers
- Your standard Sun Sparc covers the low-mid range server market quite nicely
- El Cheapo Intel box
- Apple servers
For a large ISP, rock-solid reliability, good service and security, decent software, plus someone to sue in case things goes really wrong are probably more important than the actual price of a box.
Think of the system, not the box
LL
OK guys, some basics
The properties of a unit of current are related to its functions.
1) long-term store of value
2) convenient medium of exchange
3) measure of economic goods and services
Gold has some nice properties
1) extremely stable and durable
2) divisible into smaller subunits
3) considered a luxury item and thus independent of goods and services.
The history of money is rather intriguing as historically, people have used different mediums for exchange (eg. roman solders used salt from which we've derived salary). The reason why gold has traditionally been used is that it is very hard for governments to dilute it's value. However, relying on it has rather negative consequences when everybody is using it as there are physical costs and circulation (ie liquidity) problems. Oh, plus the rather inconvient fact that the world's largest suppluers are in South Africa and Russia so whether you are leftist or rightist, you're sure to alienate half your population by depending on it as the sole measure of your economic future.
The modern fiat money system is based on trust (e.g. social security) backed by the laws of the host country. That is why the feds get rather irritated with things like counterfeiting and trying to escape the tax system as it dilutes the trust and ultimate value of the currency. Of course, they're also hoping that nobody points out the clothesless emporor in by inflating the currency through increasing the M3 supply, they automatically devalue the currency. Considering the US owes nearly 6 trillion to the rest of the world and, as a global reserve currency, is not on the gold standard, they can theoretically inflate away the entire debt through accounting tricks. (one reason why the Euro has appeared and indebted countries are so pissed off).
Governments and banks would dearly like to move people into an electronic currency (consider it as the pure economic laws without the paper or cheque bits fiddly bits) as it eliminates the horded value locked up in circulating paper (one reason why Russia is so bad is that all the rubles have been tranformed into illiquid assets and horded, thus denying their use as a circulating unit of exchange) and allows financial firms to cut handling costs (not to mention lend on the stock exchange for the hours in the day you're not using it).
Modern capitalistic countries are already seeing this with the move towards a credit-based society. Of course, if the system screws up (as the Singapore case shows), the population gets rather annoyed as nothing brings out the fear of mob hysteria than losing your shirt. However, no matter what currency you use, it still has to satisfy the functions of a universal store of value, common medium of exchange and independent unit of pricing.
LL
PS As an interesting academic exercise, you can figure out the instrinsic value of software by creating synthetic measures of its value, much like you can trade Hollywood futures on the future value of stars and films.
As for being innovative, I would be careful to distinguish between invention, incremental improvement and radically new. Universities are more likely to focus on the radically new, especially exotic languages which tests out specific ideas that eventually get incorporated into mainstream (orthogonal persistance is one example coming through the current pipeline). Application developers focus on the inventions that make like simpler, creating the killer apps of the day (VisiCalc for spreadsheets), followed by later imitators and refinements. It generally takes something about 10 years in moving major technologies from university to mainstream (assuming that anyone is interested
Considering that most people can't live on air for 10 years, Linux hackers usually end up with (hopefully) decent jobs and play around in their spare time. The amazing factor is that the relatively recent arrival of the web which allows many slices of people's spare time to accumulate into solid products, especially when they have the time luxury to reengineer a clean architecture.
Commercial vendors on the other hand have to keep in mind certain things like pleasing the stock-holders whose gracious generosity has lent them some trifling few billion to accelerate development and hype their products. Time is not a luxury and corporate secrecy (due to requirements for patenting) is an absolute. This leads to a rather closed worldview in which old techniques applied to a different setting is interpreted as "innovation". In my book truely innovative companies are those that have creately completely new sectors of the computer industry (Adobe for desktop publishing, SGI for OpenGL, AutoDesk for CAD) that wouldn't have existed otherwise.
The big problem that the Linux crowd has to address is to separate proprietary from open from expensive. Some code by it's very nature is expensive to develop (safety/fault tolerant stuff because of extensive testing). Other stuff like compilers are needed in the intermediate stages before creating the sale goods to consumers and business. Despite what people think, there is no free beer (unless you're prepared to go out and plant the crops and brew it yourself).
LL
The power and abstraction of software is both its strength and its Ancilles heel. Any thought can only be expressed using the current available languages and concepts. We have evolved from assembly-level bit twiddling, to macros, imperative languages, functional, object-oriented and now high levle scripting/glue systems. Yet each level of language abstraction invalidates any patented technique below it. If someone patents the use of the letter "E" and charges monopoly rents, think how painful communications would be.
The fundamental problem is that it is impossible to prevent other people having the same idea. Given that there are 5 billion people all trying to survive, sooner or later, everything will be independently rediscovered or reinvented. Should being the first to discover automatically extinguish the rights of others who wish to develop the idea further? One can think of an example of a certain plant MNC who has been awared the exclusive rights to bioengineer cotton fibre and has basically done nothing as yet but refuses to allow anyone else encroach on a potentially lucrative market.
Business nowadays has less to do with innovation than issues of control in a market nichce and thus extract long-term profits. Such is the power of the almighty dollar. Idealism my be good for the pampered classes but it doesn't pay for your kid's medical bills.
LL