110 hour workweeks? 72 hour project rosters? 24x7 callouts? Abusive working conditions? Continuous and expensive professional retraining?
These are some of the actual realities of life that medical interns had to suffer to achieve mastery of their speciality (and the rights to their 4 day golfing weekend). However, there is a big big difference in skill levels between a top-notch heart surgeon and a herbal dispensor with matching salary scales.
The big problem that I see is that there is no high level professional accreditation process for software development. Anyone can code, but it is harder to put together cost-effective solutions to solve a problem. The Microsoft/Cisco/etc certifications are mere first-aid certificates in comparison with the amount of training doctors insist on for their professional qualifications. And just like any high-level profession, not everybody has the tolerance of stress or the right qualities to be a super-achiever.
There are a few professional IT bodies evolving, the System Administrator's Guide (SAGE) is one example. In this frontier environment, it might be profitable peddling silicon snake oil to the ignorant masses but it does not instil long-term confidence in the customer. Unfortunately there is no easy way of separating out the good from the bad and I suspect the whole IT industry is just too immature at this stage (plus too many contentious political in-fighting for the spoils) to formalise any discipline specific review processes (much less decide on disciplines).
The alternative is to foster a culture that encourages the right long-term behaviour and allow talent to rise to the top. Much like the Hippocratic Oath defines the medical profession, I would suggest that we need an equivalent Hacker's "Code of Conduct" to decisively shift OpenSource from fighting proprietary silicon snake oil to being a peer reviewed professional quality service.
One of the most interesting things the article mentioned (I thought) was that the thing that "galvanized" research into quantum computing was a report that showed a quantum computer could break virtually any digital security system. Keeping that statement in mind, is it any wonder why people get excited about breaking into a secure computer system? Most mainstream people are appalled at crackers attempting to gain access to a secure system, yet here the same impulse is driving research that could fundamentally change our lives.
There is a big difference between exploration and conquest. Remember that technology is amoral, it is the applications/outcomes which are determined by our positive or negative tendencies, and even that is defined to some extend by social context. The psychology of "crackers" (essentially ego driven) is completely different from the pure form of hacking (curiosity driven), about the same difference between a vandal and an artist. Both might use similar techniques but the creative tendencies of an artist are channeled into positive outcomes.
Researchers are a different breed from paranoid security cryptoanalysis (at least I hope so as my taxes are being levied for them to be paranoid). Science and research operate in a climate of open and free exchange of ideas and anything which is regarded as a "difficult challenge" is of interest. Security on the other hand is dealing with control and exclusion (private property) which tends to bring out the acquisitive side of human nature (remember wanting a toy just so that your sibling can't have it?). The impulses that drive research and finding knowledge (and to some extent pure hacking) are not the same as cracking and destorying data.
I just always find it interesting when human nature morally cris-crosses itself....
Society has developed informal rules (social norms) and institutions (courts, non-profits) to minimise human nature to destroy itself. Morality is a rather complex abstraction in its own right. As one wit once noted on the dispartity of income distribution along the age axis
"If you're not a communist when you're young, you've no heart. If you're not a capitalist when you're old, you've no brain"
Good, bad, indifferent, we are all evolving bitstreams in the global memory of human conciousness. I only hope that people have enough self-awareness of their own nature and act accordingly.
The author raises a rather important point. At the moment, Linux development can be compared with a cottage industry with all the charm of custom design and individual craftmanship. Can it evolve into the formal processes of mass production for a consumer market? This is a tough call as it pits the love and talent invested into a hand-made Rolls-Royce compared with need for testing and production control needed for an assembly-line model.
The spirit of hacking is exploring new heights and creating elegant structures. It is a beguilding song, speaking against shoddy software engineering and praising open source feedback. But who controls this dialog? Does the story-tellers shape the audience or are they slaves to the fans? Would Linux be changed by imposing the legal contracts and compelled performance of the almighty dollar? Or can the two coexist peacefully?
The software industry is a complex beast, ranging from the complete free source, to the patented intellectual property. For OpenSource to succeed in mainstream, it has to offer a superior business model than existing solutions. To a large extent, the OS is invisible to end-users as what people are interested in are the services that live on top. Linux may offer more stable platform at a lower cost but unless independent software vemdors see a viable market (and that means a financial justification for porting), then it will remain marginalised compared with Windows or even Java.
NB - The suggestions were just that, some ideas for people to mull over. Obviously as artistic editor, CT would think very carefully over them so as to maintain the group dynamics, keep the pros coming in with their wide experience (I don't know what cocktail parties others attend but I choose ones where colleagues are either amusing or professionally relevant) and general "look and feel" of the experience:-).
Above all, he should retain the element of choice for the user. Maybe some people would like browsing ACs comments but at other times when rushed, they may only want the highlights. Choice in the post-modern era - tsall good (see http://www.perl.com/pace/pub/perldocs/1999/03/pm.h tml for the reference, if slashdot could have something like main points are condensed at the top, it would be fantastic).
The success of slashdot has more to do with the fact that it is a specialist rag with a good self-selective moderation policy. I would see it more as a cross between a cocktail party hosted at a professional conference with someone sneaking around with a mike, rather than the traditional newspaper which is tailored towards a mass audience (cae we say devolution to mediocratity here?). Like any specialise rag/zine whatever, it is highly dependent on the audience in mind, in this is case, generally highly educated computer professionals which results in a better signal to noise ratio than traditional newsgroups or ad-based paper publications (which of course aids for the widest possible coverage in their field).
I suspect that now that CT has some spare change to hire some Perl hackers, he could make a nice little sideline in licensing the software. (I personally would be interested in something similar to provide a grassroots community helpline). Some things that may be worthwhile tuning include
1) clear policy for submission (e.g. compulsary 6 month gag period for new subscribers while they observe the community norms) 2) more flexible extraction mechanisms (e.g. more sophisticated threading to reduce bandwidth) 3) perhaps more powerful navigation/selection features (e.g. only read articles from people who consistenty rank>2) 4) forward/backward links to related topics (actually not as easy as it sounds) so that people can understand the context and trajectory of the information vector
The internet is not radio, TV or newspaper and I personally think slashdot is a good step forward in this brave gnu world:-).
But it does remove your control and dilute the value of your works. Economics is the art of artificial scarcity. If there is a shortage of wheat, prices go up. One off artistic works like scultures are unique. That is why professional artists take great care in evolving and protecting a particular style.
However, to sell it to a museum, they need to state its authencity and source. Without the ability to control the marketing and subsequent distribution (what Yahoo are claiming IMHO) your bargining power is severely reduced.
The other point is scarcity defines value. Collectors pay fortunes for once-off items that cannot be duplicated again. By giving themselves the right to infinite duplication, Yahoo is claiming a large share of any future economic goods that a user may produce, including the power to give it away to destroy your livelihood if your income is dependent on your intellectual efforts.
Nothing technically illegal but it is not a balanced contract.
> By submitting Content to any Yahoo property, you automatically grant, or warrant that the owner of such Content has expressly granted, Yahoo the royalty-free, perpetual, irrevocable, non-exclusive and fully sublicensable right and license to use, reproduce, modify, adapt, publish, translate, create derivative works from, distribute, perform and display such Content (in whole or part) worldwide and/or to incorporate it in other works in any form, media, or technology now known or later developed.
This I suspect is a little too broad for what Yahoo is at the moment (a catalog service). I would argue on an individual level that Yahoo is a publishing house and therefore the only rights that it should receive are a right to reproduce (cache) and republsh (serve web pages to the public). If it is a multimedia studio, then it should at least have the professional courtesy to define the services it is providing to the customers (someone else can argue about the legitimacy of those services).
The scary part is the "technology now known or later developed". This effectively extinguishes all future rights to derivative products. Hypothetically if someone is the next da Vinci who started say a new trendy artform, then Yahoo can put a claim to that. I wonder what law courts would say about implicit contracts imposed by a knowledgeble party to create a lien on future income. Capitalism is one thing but selling out your future life-earnings for a pittance is another!
I think a closer examination of the situation will help explain the situation.
The basic principle of modern citizenship is individual sovereign rights with one of the outcomes is that every person's house is considered their castle. Within one's house, you can do whatever you wish. Effectively if you are sold a piece of music, within your private sphere (let's call it the domosphere in the style of ESR's noosphere + domos=house) you can copy, listen, mangle, whatever the music to your heart's content.
However, once you step outside your door, those rights are subsumed within the wider community (nomosphere? nomos=natural laws or laws made by government). Here you are govered by the laws made by politicians (from polis=political community or state), one of which is the public broadcasting and recording act (or whatever the US equivalent is). This creates certain economic incentives such as the (non)exclusive right to play music over the airwaves which has given rise to such wonderful business models such as radio stations, TV sitcoms and the wonderful world of advertising (ain't Western technology wonderful:-} ).
Now the recording industry and its associated bodies can be said (loosely speaking) to be originally formed to protect these rights and provide fair return for the artistic creators as in those days it was expensive to create complex performances.
The big big problem is that technology has inverted the economic incentives. For example, it is so easy (relatively speaking) to create your own music that copyright is effectively diluted. Storage and dissemination of information is so cheap that you can probably set up a virtual radio station through the internet and netcast your private playlist. This of course causes a few heads to start worrying about where their next paycheck will be coming from as people switch from one mechanism to another with zero loss in perception and effective demand (ie good music - though tastes may vary). Throughout history, vested interests have maintained profitable trade routes and enforced them through violent acts of war if necessary. Today's battles are being fought in the media box and courtroom with it's heavy cost of regulatory oversight and potential for exploitation of loopholes.
This is an era of great change and slashdot readers and hackers are on the front-line. In a technology battle for the consumer market, the media and public forums are being used as active weapons to both influence (and is constrained) by the legal framework. Buying a technology product is not just "buying music" but making a choice as to what laws one wishes to respect. I see the challenge to the wider hacker community is to develop technology which is both practical and deployable in the context of the wider social domain and existing business environment.
The solution is not obvious. Technoloy can be developed (patents, secure digital copying, etc) but technology (can we dare say a technology arms race is now underway between the companies?) can always be sidesteped which just illustrates the futility of unenforceable regulations. Consumers are very smart although they might not conciously realise the implicit rights embedded within the technology. DIVX is a good example as it tried to impose technology better suited for the public sphere (renting of music from a library) into the private household. Interesting technology but bad market research.
For those people willing to shell out the big money, fast I/O is certainly possible. However, I/O is not just the raw hardware but also the file system, data organisation and other important points like data integrity. RAID systems are available but figuring out an automatic mechanism of splitting up higher level files to optimise the access patterns is non-trivial. For the average consumer, there is no real need to stream multiple Gbytes off the disk. On the other hand, businesses in the document storage business don't blink when salespeople quote 6 or 7 figure prices for high-end enterprise storage systems (one reason why EMC and kin are earning 50%+ gross margins). Sure beats the 5% on PC sales.
> John Mashey (SGI, was MIPS, Bell Labs) has been lecturing recently on "Infrastress"
John is a smart cookie. I/O has always been problem. The issue is partly technical, and partly markets. The technical constraints are that electronics have been getting a free ride by Moore's law whereas communications is limited by the physical properties of the copper/fibre and the signalling sensitivity/intelligence. As others have pointed out, the storage is constrained by mechanical properties and the cost model of factory assembling lines (no high throughput fabs with automatic line shrinks).
The marketing is due to certain companies hyping the raw MHz of the machine rather than looking at it in the context of a balanced architecture. What's the point of a Porche that goes x zillion kms/hour if the road limit is only 50? So the cost structure tends to favor electronic advances over mechanical. Hence you see the general trend of microprocessors migrating to where the storage is (processor in memory, embedded DVDs like Nuon, and hopefully smarter SANs).
Fast I/O is possible but it is most definitely not cheap (laws of supply and demand). Bulk storage is also not cheap as plugging in a drive doesn't scale up well when you're considering multi-terabyte systems. When there is a compelling market reason, the forces of capitalism will act with its usual speed.
So, what will be the killer app to kick off the demand for fast storage?
> Not true. When I was doing consulting work, we regularly could underbid proprietary solutions by using Linux, giving us a significant competitive advantage while actually maintaining a higher margin than our competitors.
Each software has their strengths. The advantage of MS products are that they support the most common features of the business market. The advantage of Linux is its easy customisation and ability to solve once-off problems while maintaining profits to the smaller consulting companies. It all comes down to the same basic issue, if you look after the customer properly, then the profits will look after themselves.
> Linux should be afraid. Fear is the perfect motivator.
Fear is the mind killer - Bene Gesserit(sp?)
I'm a little perturbed at this attitude which is reminescent of Intel and MS corporate culture.
IMHO "hackers" do things because of the challenge, curiosity, and sheer creative exuburance. The day hacking stops being enjoyable is the day that I'll throw away the computer and retreat to NZ (actually a fantastic place to visit).
Sure performance is a nice feature to have but as all good engineers know, there are always tradeoffs in time, price and features. Instead I suggest some fundamental thought go into a clean design which is upwards compatible with future hardware trends whilst supporting a range of architectures. Coming from an HPC arena, I know that tweaking the code can give significant short-term boosts but at the cost of portability.
The money saved on the OS can go into smarter memory subsystems or more intelligent hardware detection. A flock of penguins does not have to occupy the same business niche as a market gorilla. What the distributors like Red Hat could do is provide out of box optimised binaries for common platforms or even work closely with vendors to pretune the code (Cobalt and the QCube is one example). I would suggest using the strengths of Linux such as its diversity and open flexible architecture to provide a cost-effective solution to the low-mid server range without pandering to irrational prejudices or self-defeating chest beating PR exercises. (Of course, if there are serious performance problems, then one has to rectify them).
Remember, if you're not having fun, then it's not hacking.
The government may be a monopoly but citizens of free nations have a choice of emigrating. There are plently of first world nations which would welcome people of talent (Canada, Australia, NZ). There are also various means of influencing the outcome of the goverment actions - ballot box - jury box - bullet box
On the other hand, influencing corporations are a little trickier if the MNC has a greater cashflow than the government itself. Some possibilities suggest themselves: - proxy vote of shareholders - legal suits within recognised juristiction - protest through your spending habits - idiot box.. err... TV anti-FUD through PR
The key element is an informed individual choice LL
ESR offers a nice summary of the situation. I'll attempt to generalise it by looking a little bit more closely at the software development cycle and relating it back to some basic economic principles.
If we accept the analogy that software development (and indeed all intellectual endeavour) is based on exploring and developing the (potentially infinite) Noosphere, then we can look at the common law practices that developed and codified as a result of the gold explorations as it evolved from the goldrush days to today's highly efficient staged exploration, extraction, processing, refinement and disposal.
This can be considered the 5 stages of software development
alpha) from exploring the basic idea beta) ramping through prototypes gamma) sale of industrial strength product delta) major iterations epsilon) obsolescence and shutdown/maintenance
Now economics can be basically divided into consumption and capital goods, what ESR defines as the sale and use value in measuring software. The sale value is all important as it is the final "good" that is ultimately consumed, the end application in other words. The use goods are the intermediate software needed to produce the sale goods (compilers, drivers, OS, etc). By themselves they have high development costs but are under external cost pressure so as to minimise the final sale product. This explains the commoditisation of these products through OpenSource means as unless there is extreme specialisation (high performance compilers, embedded systems) people prefer familiar toolsets and reduced learning curve.
However, the economic justification of charging money is that the developers of these tools have to eat while spending time producing and refining the languages and techniques. No gnu compilers, minimal OpenSourced products as the entry costs for beginners is too high. This has to be balanced by the fact that for profit or personal satisfaction, the compiler writers have to eat or get reimbursed for their time. So the economic requirements at the different development stages are quite different.
alpha - exploration - grubstake, a bet on a pie in the sky that your idea might turn out to have some commercial applications
beta - extraction - seed money to dig through the possibilities and identify the gold seam
gamma - processing - development and sale of a useful entity
delta - refinement - adding increasing value and generating higher order goods like jewelery from the basic gold
epsilon - disposal - the whole enterprise has to be profitable to ensure that there is enough funds (e.g. professional indemity) to ensure that the customers are taken care of even if disaster strikes. This would be considered in the context of environmental cleanup after the seam runs out.
Now the current economic model of software patents and IPOs give rise to a rather unhealthy situation in my view. Essentially companies are rewards for the first to strike gold and gain a monopoly in a area which they can exploit to the hilt. As ESR pointed out, this leads to rather questional quality and poor customer satisfaction.
OpenSource offers a viable alternative BUT the problem is that unless it can offer a decent mechanism of supporting the intermediate tools (ie compilers, OS, 3D world construction) the only outcomes will be small toy programs which are done as academic exercises or learning experiences. This can be seen in the relatively unsophisticated capital markets of SE Asia as apart from government supported conglomerates, businesses are run by small family groups.
Thus the OpenSource movement needs to address the intellectual property rights issue to really grow from being a grassroots protest against existing proprietary software practices, to being a supported industry in its own right. Though I've got a few ideas, I'd really like people to suggest what would be a workable system given the entrenched myriad of vested interests in the current software patent approach.
This includes the usual "dirty tricks" that have plagued the development of mineral extraction. Claim jumping, absentee landlords, appraisal dilution, highway robbery, economic externalities, legal responsibility to upstream and downstream suppliers and customers, etc... Fun and games if you're young and desperate but not something that suits a family man (or woman).
Despite its appearances, the IT industry still acts like a roughshod mining town chasing a vaporush than a professional outfit!
LL
PS. Even I'm not immune from opportunistic monetary rewards as I would like to be remembered for originality in coining the phrase vaporush but unfortunately there's no way of directly capitalising on it. Such is life.
> Imagine Linux running 128 CPUs in the same box with a massive amount of memory.
As Linus noted, support for more than 4 processors would add unnecessary kernel overhead for uniprocessors. I suspect SGI would include this as one of their value added features of Linux + Irix (somebody better think of a better logo for this rather than a rainbow colored penguin).
SGI exec team are not stupid (although their marketing guys seem to be rather flat-footed), if OS innovations are being generated faster by an external group (which Linux appears to be at the moment) then it is better to harness the external momentum and build on top of it and reallocate precious staff resources to where it is really needed. Personally I would see great benefits in tuning the GNU compilers to be more aware of the underlying cache systems and their souped up bus (those cache-lines, false sharing, etc) are a pain to manually exploit and also makes the standard code base rather unportable. SGI also has got some nice memory subsystem hardware which could benefit from going mainstream to really sort out that kfluffle between NGIO and FIO.
Imagine the CPU grunt of an Alpha, the I/O, memory and graphics of an SGI, peripherals of IBM, the pricing of Dell, and the software stability of Linux as your dream box.
Let each vendor concentrate on their strengths and allow the free market to decide on the worth of their products based on a unencumbered and informed choice!
The 2bps quoted is rather interesting. Are there any studies on whether this is an absolute or indexed measure? What I mean is that its been discovered that though chess grandmasters follow the 7+-2 rule in short-term memory, they recall chunks of moves rather than individual steps. Thus 2 bits of information might be considered a compact index into a sparse information space. The more specialised a person, the more complex representation can be used.
The brain is interesting in that it generally filters out daily dross based on preconceptions and training. Flemming recognised the connection between mould and lack of bateria whereas the average lab cleaner would have thrown away the contamination without realising the significance. Thus though the memory may be limited in volume, the extreme selectivity and pattern association capabilities gives humans certain advantage in certain areas of processing. If we consider that the 2 bits can be a useful index/catalog, then you can leave the details up to a computer.
Cognitive science is going to be a fun topic over the next few decades.
How to reconcile OpenSource, intellectual property and profitability?
Part of the misunderstanding is defining exactly what is software. Some people view it as a finished tangible product, much like a car whereas other view it as an on-going conversation with multiple sub-languages and cultural variants. With this in mind, this gives rise to some fundamental differences with traditional economics which relies on artificial scarcity (e.g. a monopolist has the ability to exclude competitors and thus limit the avaiability). However, the rather striking effect is that the existance of any "free" software eventually eliminates any for-fee substitute (e.g. MS Explorer vs Netscape).
Given the rather constraits of everyone having equal access to the source, it is still possible to create sustainable business models. Ghostscript is an example where the author creates commercial versions, then turns them into GPL versions after a few years. The other mechanism is to turn the strengths of OpenSource into marketing points (trust, near zero-defect, rapid development cycle, knowledgeable pool of custom programmers for-hire) and become the low-cost and efficient supplier of a market segment. Modern intellectual property is a bit of an illusion as given that all software technology is based on previous tools, it is very hard to come up with a unique solution that 10 million other software developers won't encounter within another 2-3 years. Unlike hardware products, all the minor coding variations can be easily parameterised and investigated reducing the scope for differentiation (how many ways are there really of expressing a spreadsheet?)
Capitalism is a rather complex beast but the essential point is that it is the efficient allocation of surplus time (traditionally stored/measured in dollars) to create future economic goods. The $1 that a farmer saves by not buying a lotto ticket can be invested in funding a startup which (hopefully) can return $2 (or more) to buy future consumer products. With the limiting factor in software development being talent, the OpenSource model is one solution to give any programmer a go at improving software that they are interested in. In this sense, with open peer review, it is closer to a meritocracy than any other social system. If someone comes up with a brillant new hack to double the speed of Quake, I'm sure ID software would hire them in a flash. You can already see a similar situation with vulture... errr... venture capitalists betting as much on the quality of the people in the management team as on the actual technology.
The nature of the internet does change the traditional recruitment rules in that it opens up new sources of talent worldwide (note that Mexico has installed Linux in its education system which might lead to some interesting long-term effects). As usual, some companies will adapt and prosper, others will wither and get absorbed.
While the bounty or sponsor model have commercial overtones, I would suggest that the {rich individual | corporate} patron model so prevalent in the Renaisance era is closer to the gift economy that Raymond espouses. This is where the patron demonstrates their {benevolance | lucre} by funding artistic works for public display (Sistine chapel, etc). Given that the highest levels of programming is more akin to art rather than engineering, I would speculate this may be more attractive to the highly talented segment of the hacker community.
The biggest problem I see is that in a rather large but closed world, it takes a while for "names" to be recognised as stellar qualities (e.g. Linus) and the reputation takes a long time to build up. But maybe this can be viewed as an incentive for people to contribute to OpenSource in that by associating a name with a successful project, they can effectively provide an open job resume to the world? Much like the journey-man stage of publicly displaying goods for review by the customer, it can be viewed as a necessary stage of development towards mastering a field. I can forsee the day when employees hand-pick programmers based on recognised contributions to an OpenSource program which clearly demonstrates their level of skill. To support this model does require some work in that OpenSource projects have a clear attribution history of contributors and their degree of worth (e.g. hacker X contributed n% to the design and high-level coding).
The biggest hinderance is that nobody has resolved the inherent conflicts with existing commercial models.
... according to what some people claim, the US federal government is technically a political church trust. For details, see
http://www.ptialaska.net/~swampy/amend_14/usa.ht ml
I'm not sure exactly how to interpret this (any legal eagles out there?) but my (entirely personal and thus biased) observation is that the US acts more like a megacorporation with all citizens as shareholders, Congress as directors, and president as CEO, and with the wonders of modern capitalism, one dollar, one vote:-(.
Of course as Peter Drucker once said, governments do two things well, wage war and inflate the currency. In corporate parlance, win marketshare and dilute earnings:-).
As an aside, this may offer a rather interesting way of sidestepping the mess caused by the software patent system (ie enforced monopolies for the first to come up with an idea) with something more commonsense based on common law with full disclosure (OpenSource) between two parties. Oh well, one can dream....
Both SGI and IBM are smart in their own ways. The limitation nowadays is not hardware (constrained by physics and economics) or technology (can always be bought out) but human talent. Given the increasing software complexity, it can take years for people to learn and master sophisticated software packages and their APIs. Moore's Law dictates that whatever you buy now can be bought cheaper next year. Hence anything developed at the high end will eventually filter down to the PC level. Unfortunately human learning curves moves slower than hardware evolution cycles. The power of visualisation packages such as AVS is in the module libraries that the community develops and contributes as once the critical mass is there, it is very hard to change habits or programming practices. In a way, releasing OpenSource is like force-growing a new ecosystem niche with the hope that enough talent will feed on the transplanted energy and expand the overall food chain (always a problem in small specialised fields) in the hopes that the colony becomes self-sustaining (creating future demand for similar food). The path to growth for both IBM and SGI is to move their "proprietary" IP into mainstream as quickly as possible to extract maximum benefit from their R&D dollars. For this, they need to win over developers and new power users of which academia is a unique training ground.
I can see the day when companies attempt to buyout emerging graduates as without a constant source of talent, any company will eventually wither under the winds of global competition.
Software doesn't exist in isolation, in some ways it reflects aspects of society. Just as there are non-profit organisations aiming to look after the underpriviledged, there are megacorporations interested in using IT to gain a competitive edge and monetary value is a measure of their efficiency. GNU/Linux has its major strength in providing basic civic infrastructure for the computer world (standards/protocols/education). The corporate domain should then be welcome to add value with the understanding that the quality standards are increasing all the time from volunteer work. Its just like giving people the choice between public and private education, you pay more for 'perceived' extras (not to disparange anyone, there are some very good public schools out there).
Having a diverse software ecosystem creates a robust and resilient playing field. Dinosaurs can coexist with monkeys, afterall I believe a few (rare) places still use punch cards. Given the rapid technology churn and obsolescence, there is a need for at least a stable foundation for beginning firms to practice on before hitting the brutal business world.
The biggest problem is getting enough funds to help educate the poor consumer as to the reality of buying software as currently there is no real ombusman role. While marketing adds to the cost, it is a way to help publicise good technology that would otherwise not survive. Perhaps there is some call for a social contract? If a for-profit corporation uses GPL software, perhaps it be suggested that they donate 5% of net profits/benefits to a charitable organisation of the author's choice (e.g. FSF).
IMHO, one of the main strengths of OSS is that the developers are also the users, as such the feedback is (nearly) immediate. Consequently people only develop systems which are stable and flexible because it has direct benefits. After all, would you buy food from a chef that doesn't sample their own food? The problem comes in that by having users 1 or more step removed from the development cycle, it is difficult to produce systems that fit their needs perfectly. The concept of a free market is that it sends "pricing signals" to producers to identify important traits. Hence, from the success of integrated packages, one can say that interoperability between major packages is a highly desired trait. Now free markets are predicated on a large smaple of consumers having informed choice. If they are misinformed (hype/FUD/fraud) or restricted in choice (monopoly, trade barriers, quotas) then the signals are distorted.
OSS has shown the value of global software development, combining the 'best' memes from multiple talents. I suspect the 'best' business model that works in real-life is still yet to be found. Effectively for a points system you are asking for an economic metric that measures talent/time. But here it hits problems such as multiple value systems. What does a point mean to a 16 year old teenager? The value of having hard dollars is that it is a commonly agreed upon unit of interchange and is also a substitute for future needs, rewarding entrepreneurs that best guess what future consumers want.
Hackers tend to develop stuff which automate the boring administrative tasks, leaving their skills for the more fun and creative exploratory tasks. As such, what one releases, has immediate revelance to peers. The other difference with commercial systems is that there is no fear of competition because the real talent is in the living knowledge of the developers.
Now, based on these personal observations, I would suggest some alternative systems:
1) purely commercial - trade OSS for options/shares in commercial outfit 2) casual contractual - set up a virtual company to offer customisation and support for individual corporate needs 3) purely altruistic - social contract, request 5% of net profits be donated to charitable organisation of choice
OSS is still an evolving phenomena so it will be interesting to see what will work in practice.
Is there need to call for OpenSource Jihad against every commercial company? If a group of people have invested time and talent into developing a product, should they not enjoy some return on their efforts? If there is serious need to create GPL versions, then it should be a 'simple' matter to ask for enough donations to 'buy out' the company (ie discounted sum of future revenues) and make the software completely free. This is a straight-forward money for time tradeoff, afterall you are asking people to volunteer their precious time.
One has to keep in mind what the civic outcomes of OpenSource project observed as being currently successful are 1) educational/intellectual curiosity (unis) 2) common standards (e.g. w3c) 3) not-for-profit utilities or public awareness
Software development is a complex and expensive business. Effectively by releasing an OpenSource version, you are undermining the value of the commercial alternative. This serves a long-term purpose of continually raising the quality bar and forcing companies to keep on innovating but it should not be viewed as a call to arms to duplicate everything in sight. OpenSource can play a role in keeping in check outright exploitation but it should be applied with some awareness as to the consequences.
Another way of looking at the problem is that you get paid as recompense for time taken away from doing 'fun' stuff.
Let's face it, most work can be placed in the tedious but needs to be done category. User support, maintainence, integration, documentation, etc.... Though OpenSource can be a very effective creative force, somehow it must coexist within a very consumer-oriented society where most people do not earn >$100K programming salaries. 'Suits' are then charged with the responsibility of trying to squeeze more efficiency and defined outputs from the system. If OpenSource represents 'free energy' then it will be 'exploited' by those in a position to do so.
The interesting factor is what conditions external groups place on the 'gift culture' in an attempt to control conditions for their own benefit. For example, there is a certain amount of 'reciprocracy' in the system where it is expected that on receiving a gift, there will be a balanced contribution by the receivee in the future. In a system where there are different values (suits vs hackers), asymetric expectations can cause tension.
Designing a Formula 1 car is fun and sexy, but someone has to do boring tractors otherwise nobody eats. Given that the biggest bottleneck at the moment is lack of trained people, OpenSource makes sense in optimising the spread of knowledge. On the other hand, if everyone can program, I would suspect a proprietary system will be 'superior' as it retains a 'competitive edge' for a select group. What works in reality can be quite complex and it's probably best to be flexible and choose whatever system best suits the circumstances.
There's no way that Australia can absorb a Microsoft. We've got a serious shortage of programmers already (must be due to the real sand, real sun, and real surf up here in sunny Queensland) and we're only a 12th the population size of the US. They can take over Cambridge or Peking instead where they've already got labs.
Slashdot Mirror
110 hour workweeks? 72 hour project rosters? 24x7 callouts? Abusive working conditions? Continuous and expensive professional retraining?
These are some of the actual realities of life that medical interns had to suffer to achieve mastery of their speciality (and the rights to their 4 day golfing weekend). However, there is a big big difference in skill levels between a top-notch heart surgeon and a herbal dispensor with matching salary scales.
The big problem that I see is that there is no high level professional accreditation process for software development. Anyone can code, but it is harder to put together cost-effective solutions to solve a problem. The Microsoft/Cisco/etc certifications are mere first-aid certificates in comparison with the amount of training doctors insist on for their professional qualifications. And just like any high-level profession, not everybody has the tolerance of stress or the right qualities to be a super-achiever.
There are a few professional IT bodies evolving, the System Administrator's Guide (SAGE) is one example. In this frontier environment, it might be profitable peddling silicon snake oil to the ignorant masses but it does not instil long-term confidence in the customer. Unfortunately there is no easy way of separating out the good from the bad and I suspect the whole IT industry is just too immature at this stage (plus too many contentious political in-fighting for the spoils) to formalise any discipline specific review processes (much less decide on disciplines).
The alternative is to foster a culture that encourages the right long-term behaviour and allow talent to rise to the top. Much like the Hippocratic Oath defines the medical profession, I would suggest that we need an equivalent Hacker's "Code of Conduct" to decisively shift OpenSource from fighting proprietary silicon snake oil to being a peer reviewed professional quality service.
LL
There is a big difference between exploration and conquest. Remember that technology is amoral, it is the applications/outcomes which are determined by our positive or negative tendencies, and even that is defined to some extend by social context. The psychology of "crackers" (essentially ego driven) is completely different from the pure form of hacking (curiosity driven), about the same difference between a vandal and an artist. Both might use similar techniques but the creative tendencies of an artist are channeled into positive outcomes.
Researchers are a different breed from paranoid security cryptoanalysis (at least I hope so as my taxes are being levied for them to be paranoid). Science and research operate in a climate of open and free exchange of ideas and anything which is regarded as a "difficult challenge" is of interest. Security on the other hand is dealing with control and exclusion (private property) which tends to bring out the acquisitive side of human nature (remember wanting a toy just so that your sibling can't have it?). The impulses that drive research and finding knowledge (and to some extent pure hacking) are not the same as cracking and destorying data.
I just always find it interesting when human nature morally cris-crosses itself....
Society has developed informal rules (social norms) and institutions (courts, non-profits) to minimise human nature to destroy itself. Morality is a rather complex abstraction in its own right. As one wit once noted on the dispartity of income distribution along the age axis
"If you're not a communist when you're young, you've no heart. If you're not a capitalist when you're old, you've no brain"
Good, bad, indifferent, we are all evolving bitstreams in the global memory of human conciousness. I only hope that people have enough self-awareness of their own nature and act accordingly.
LL
The author raises a rather important point. At the moment, Linux development can be compared with a cottage industry with all the charm of custom design and individual craftmanship. Can it evolve into the formal processes of mass production for a consumer market? This is a tough call as it pits the love and talent invested into a hand-made Rolls-Royce compared with need for testing and production control needed for an assembly-line model.
The spirit of hacking is exploring new heights and creating elegant structures. It is a beguilding song, speaking against shoddy software engineering and praising open source feedback. But who controls this dialog? Does the story-tellers shape the audience or are they slaves to the fans? Would Linux be changed by imposing the legal contracts and compelled performance of the almighty dollar? Or can the two coexist peacefully?
The software industry is a complex beast, ranging from the complete free source, to the patented intellectual property. For OpenSource to succeed in mainstream, it has to offer a superior business model than existing solutions. To a large extent, the OS is invisible to end-users as what people are interested in are the services that live on top. Linux may offer more stable platform at a lower cost but unless independent software vemdors see a viable market (and that means a financial justification for porting), then it will remain marginalised compared with Windows or even Java.
LL
NB - The suggestions were just that, some ideas for people to mull over. Obviously as artistic editor, CT would think very carefully over them so as to maintain the group dynamics, keep the pros coming in with their wide experience (I don't know what cocktail parties others attend but I choose ones where colleagues are either amusing or professionally relevant) and general "look and feel" of the experience :-).
h tml for the reference, if slashdot could have something like main points are condensed at the top, it would be fantastic).
Above all, he should retain the element of choice for the user. Maybe some people would like browsing ACs comments but at other times when rushed, they may only want the highlights. Choice in the post-modern era - tsall good (see http://www.perl.com/pace/pub/perldocs/1999/03/pm.
LL
The success of slashdot has more to do with the fact that it is a specialist rag with a good self-selective moderation policy. I would see it more as a cross between a cocktail party hosted at a professional conference with someone sneaking around with a mike, rather than the traditional newspaper which is tailored towards a mass audience (cae we say devolution to mediocratity here?). Like any specialise rag/zine whatever, it is highly dependent on the audience in mind, in this is case, generally highly educated computer professionals which results in a better signal to noise ratio than traditional newsgroups or ad-based paper publications (which of course aids for the widest possible coverage in their field).
:-).
I suspect that now that CT has some spare change to hire some Perl hackers, he could make a nice little sideline in licensing the software. (I personally would be interested in something similar to provide a grassroots community helpline). Some things that may be worthwhile tuning include
1) clear policy for submission (e.g. compulsary 6 month gag period for new subscribers while they observe the community norms)
2) more flexible extraction mechanisms (e.g. more sophisticated threading to reduce bandwidth)
3) perhaps more powerful navigation/selection features (e.g. only read articles from people who consistenty rank>2)
4) forward/backward links to related topics (actually not as easy as it sounds) so that people can understand the context and trajectory of the information vector
The internet is not radio, TV or newspaper and I personally think slashdot is a good step forward in this brave gnu world
LL
But it does remove your control and dilute the value of your works. Economics is the art of artificial scarcity. If there is a shortage of wheat, prices go up. One off artistic works like scultures are unique. That is why professional artists take great care in evolving and protecting a particular style.
However, to sell it to a museum, they need to state its authencity and source. Without the ability to control the marketing and subsequent distribution (what Yahoo are claiming IMHO) your bargining power is severely reduced.
The other point is scarcity defines value. Collectors pay fortunes for once-off items that cannot be duplicated again. By giving themselves the right to infinite duplication, Yahoo is claiming a large share of any future economic goods that a user may produce, including the power to give it away to destroy your livelihood if your income is dependent on your intellectual efforts.
Nothing technically illegal but it is not a balanced contract.
LL
> By submitting Content to any Yahoo property, you automatically grant, or warrant that the owner of such Content has expressly granted, Yahoo the royalty-free, perpetual, irrevocable, non-exclusive and fully sublicensable right and license to use, reproduce, modify, adapt, publish, translate, create derivative works from, distribute, perform and display such Content (in whole or part) worldwide and/or to incorporate it in other works in any form, media, or technology now known or later developed.
This I suspect is a little too broad for what Yahoo is at the moment (a catalog service). I would argue on an individual level that Yahoo is a publishing house and therefore the only rights that it should receive are a right to reproduce (cache) and republsh (serve web pages to the public). If it is a multimedia studio, then it should at least have the professional courtesy to define the services it is providing to the customers (someone else can argue about the legitimacy of those services).
The scary part is the "technology now known or later developed". This effectively extinguishes all future rights to derivative products. Hypothetically if someone is the next da Vinci who started say a new trendy artform, then Yahoo can put a claim to that. I wonder what law courts would say about implicit contracts imposed by a knowledgeble party to create a lien on future income. Capitalism is one thing but selling out your future life-earnings for a pittance is another!
LL
I think a closer examination of the situation will help explain the situation.
:-} ).
The basic principle of modern citizenship is individual sovereign rights with one of the outcomes is that every person's house is considered their castle. Within one's house, you can do whatever you wish. Effectively if you are sold a piece of music, within your private sphere (let's call it the domosphere in the style of ESR's noosphere + domos=house) you can copy, listen, mangle, whatever the music to your heart's content.
However, once you step outside your door, those rights are subsumed within the wider community (nomosphere? nomos=natural laws or laws made by government). Here you are govered by the laws made by politicians (from polis=political community or state), one of which is the public broadcasting and recording act (or whatever the US equivalent is). This creates certain economic incentives such as the (non)exclusive right to play music over the airwaves which has given rise to such wonderful business models such as radio stations, TV sitcoms and the wonderful world of advertising (ain't Western technology wonderful
Now the recording industry and its associated bodies can be said (loosely speaking) to be originally formed to protect these rights and provide fair return for the artistic creators as in those days it was expensive to create complex performances.
The big big problem is that technology has inverted the economic incentives. For example, it is so easy (relatively speaking) to create your own music that copyright is effectively diluted. Storage and dissemination of information is so cheap that you can probably set up a virtual radio station through the internet and netcast your private playlist. This of course causes a few heads to start worrying about where their next paycheck will be coming from as people switch from one mechanism to another with zero loss in perception and effective demand (ie good music - though tastes may vary). Throughout history, vested interests have maintained profitable trade routes and enforced them through violent acts of war if necessary. Today's battles are being fought in the media box and courtroom with it's heavy cost of regulatory oversight and potential for exploitation of loopholes.
This is an era of great change and slashdot readers and hackers are on the front-line. In a technology battle for the consumer market, the media and public forums are being used as active weapons to both influence (and is constrained) by the legal framework. Buying a technology product is not just "buying music" but making a choice as to what laws one wishes to respect. I see the challenge to the wider hacker community is to develop technology which is both practical and deployable in the context of the wider social domain and existing business environment.
The solution is not obvious. Technoloy can be developed (patents, secure digital copying, etc) but technology (can we dare say a technology arms race is now underway between the companies?) can always be sidesteped which just illustrates the futility of unenforceable regulations. Consumers are very smart although they might not conciously realise the implicit rights embedded within the technology. DIVX is a good example as it tried to impose technology better suited for the public sphere (renting of music from a library) into the private household. Interesting technology but bad market research.
As we say in Australia, all the gear but no idea.
LL
For those people willing to shell out the big money, fast I/O is certainly possible. However, I/O is not just the raw hardware but also the file system, data organisation and other important points like data integrity. RAID systems are available but figuring out an automatic mechanism of splitting up higher level files to optimise the access patterns is non-trivial. For the average consumer, there is no real need to stream multiple Gbytes off the disk. On the other hand, businesses in the document storage business don't blink when salespeople quote 6 or 7 figure prices for high-end enterprise storage systems (one reason why EMC and kin are earning 50%+ gross margins). Sure beats the 5% on PC sales.
You want, you pay for it.
LL
> John Mashey (SGI, was MIPS, Bell Labs) has been lecturing recently on "Infrastress"
John is a smart cookie. I/O has always been problem. The issue is partly technical, and partly markets. The technical constraints are that electronics have been getting a free ride by Moore's law whereas communications is limited by the physical properties of the copper/fibre and the signalling sensitivity/intelligence. As others have pointed out, the storage is constrained by mechanical properties and the cost model of factory assembling lines (no high throughput fabs with automatic line shrinks).
The marketing is due to certain companies hyping the raw MHz of the machine rather than looking at it in the context of a balanced architecture. What's the point of a Porche that goes x zillion kms/hour if the road limit is only 50? So the cost structure tends to favor electronic advances over mechanical. Hence you see the general trend of microprocessors migrating to where the storage is (processor in memory, embedded DVDs like Nuon, and hopefully smarter SANs).
Fast I/O is possible but it is most definitely not cheap (laws of supply and demand). Bulk storage is also not cheap as plugging in a drive doesn't scale up well when you're considering multi-terabyte systems. When there is a compelling market reason, the forces of capitalism will act with its usual speed.
So, what will be the killer app to kick off the demand for fast storage?
LL
> Not true. When I was doing consulting work, we regularly could underbid proprietary solutions by using Linux, giving us a significant competitive advantage while actually maintaining a higher margin than our competitors.
Each software has their strengths. The advantage of MS products are that they support the most common features of the business market. The advantage of Linux is its easy customisation and ability to solve once-off problems while maintaining profits to the smaller consulting companies. It all comes down to the same basic issue, if you look after the customer properly, then the profits will look after themselves.
LL
> Linux should be afraid. Fear is the perfect motivator.
Fear is the mind killer - Bene Gesserit(sp?)
I'm a little perturbed at this attitude which is reminescent of Intel and MS corporate culture.
IMHO "hackers" do things because of the challenge, curiosity, and sheer creative exuburance. The day hacking stops being enjoyable is the day that I'll throw away the computer and retreat to NZ (actually a fantastic place to visit).
Sure performance is a nice feature to have but as all good engineers know, there are always tradeoffs in time, price and features. Instead I suggest some fundamental thought go into a clean design which is upwards compatible with future hardware trends whilst supporting a range of architectures. Coming from an HPC arena, I know that tweaking the code can give significant short-term boosts but at the cost of portability.
The money saved on the OS can go into smarter memory subsystems or more intelligent hardware detection. A flock of penguins does not have to occupy the same business niche as a market gorilla. What the distributors like Red Hat could do is provide out of box optimised binaries for common platforms or even work closely with vendors to pretune the code (Cobalt and the QCube is one example). I would suggest using the strengths of Linux such as its diversity and open flexible architecture to provide a cost-effective solution to the low-mid server range without pandering to irrational prejudices or self-defeating chest beating PR exercises. (Of course, if there are serious performance problems, then one has to rectify them).
Remember, if you're not having fun, then it's not hacking.
LL
The government may be a monopoly but citizens of free nations have a choice of emigrating. There are plently of first world nations which would welcome people of talent (Canada, Australia, NZ). There are also various means of influencing the outcome of the goverment actions
.. err ... TV anti-FUD through PR
- ballot box
- jury box
- bullet box
On the other hand, influencing corporations are a little trickier if the MNC has a greater cashflow than the government itself. Some possibilities suggest themselves:
- proxy vote of shareholders
- legal suits within recognised juristiction
- protest through your spending habits
- idiot box
The key element is an informed individual choice
LL
ESR offers a nice summary of the situation. I'll attempt to generalise it by looking a little bit more closely at the software development cycle and relating it back to some basic economic principles.
If we accept the analogy that software development (and indeed all intellectual endeavour) is based on exploring and developing the (potentially infinite) Noosphere, then we can look at the common law practices that developed and codified as a result of the gold explorations as it evolved from the goldrush days to today's highly efficient staged exploration, extraction, processing, refinement and disposal.
This can be considered the 5 stages of software development
alpha) from exploring the basic idea
beta) ramping through prototypes
gamma) sale of industrial strength product
delta) major iterations
epsilon) obsolescence and shutdown/maintenance
Now economics can be basically divided into consumption and capital goods, what ESR defines as the sale and use value in measuring software. The sale value is all important as it is the final "good" that is ultimately consumed, the end application in other words. The use goods are the intermediate software needed to produce the sale goods (compilers, drivers, OS, etc). By themselves they have high development costs but are under external cost pressure so as to minimise the final sale product. This explains the commoditisation of these products through OpenSource means as unless there is extreme specialisation (high performance compilers, embedded systems) people prefer familiar toolsets and reduced learning curve.
However, the economic justification of charging money is that the developers of these tools have to eat while spending time producing and refining the languages and techniques. No gnu compilers, minimal OpenSourced products as the entry costs for beginners is too high. This has to be balanced by the fact that for profit or personal satisfaction, the compiler writers have to eat or get reimbursed for their time. So the economic requirements at the different development stages are quite different.
alpha - exploration - grubstake, a bet on a pie in the sky that your idea might turn out to have some commercial applications
beta - extraction - seed money to dig through the possibilities and identify the gold seam
gamma - processing - development and sale of a useful entity
delta - refinement - adding increasing value and generating higher order goods like jewelery from the basic gold
epsilon - disposal - the whole enterprise has to be profitable to ensure that there is enough funds (e.g. professional indemity) to ensure that the customers are taken care of even if disaster strikes. This would be considered in the context of environmental cleanup after the seam runs out.
Now the current economic model of software patents and IPOs give rise to a rather unhealthy situation in my view. Essentially companies are rewards for the first to strike gold and gain a monopoly in a area which they can exploit to the hilt. As ESR pointed out, this leads to rather questional quality and poor customer satisfaction.
OpenSource offers a viable alternative BUT the problem is that unless it can offer a decent mechanism of supporting the intermediate tools (ie compilers, OS, 3D world construction) the only outcomes will be small toy programs which are done as academic exercises or learning experiences. This can be seen in the relatively unsophisticated capital markets of SE Asia as apart from government supported conglomerates, businesses are run by small family groups.
Thus the OpenSource movement needs to address the intellectual property rights issue to really grow from being a grassroots protest against existing proprietary software practices, to being a supported industry in its own right. Though I've got a few ideas, I'd really like people to suggest what would be a workable system given the entrenched myriad of vested interests in the current software patent approach.
This includes the usual "dirty tricks" that have plagued the development of mineral extraction. Claim jumping, absentee landlords, appraisal dilution, highway robbery, economic externalities, legal responsibility to upstream and downstream suppliers and customers, etc... Fun and games if you're young and desperate but not something that suits a family man (or woman).
Despite its appearances, the IT industry still acts like a roughshod mining town chasing a vaporush than a professional outfit!
LL
PS. Even I'm not immune from opportunistic monetary rewards as I would like to be remembered for originality in coining the phrase vaporush but unfortunately there's no way of directly capitalising on it. Such is life.
> Imagine Linux running 128 CPUs in the same box with a massive amount of memory.
As Linus noted, support for more than 4 processors would add unnecessary kernel overhead for uniprocessors. I suspect SGI would include this as one of their value added features of Linux + Irix (somebody better think of a better logo for this rather than a rainbow colored penguin).
SGI exec team are not stupid (although their marketing guys seem to be rather flat-footed), if OS innovations are being generated faster by an external group (which Linux appears to be at the moment) then it is better to harness the external momentum and build on top of it and reallocate precious staff resources to where it is really needed. Personally I would see great benefits in tuning the GNU compilers to be more aware of the underlying cache systems and their souped up bus (those cache-lines, false sharing, etc) are a pain to manually exploit and also makes the standard code base rather unportable. SGI also has got some nice memory subsystem hardware which could benefit from going mainstream to really sort out that kfluffle between NGIO and FIO.
Imagine the CPU grunt of an Alpha, the I/O, memory and graphics of an SGI, peripherals of IBM, the pricing of Dell, and the software stability of Linux as your dream box.
Let each vendor concentrate on their strengths and allow the free market to decide on the worth of their products based on a unencumbered and informed choice!
LL
The 2bps quoted is rather interesting. Are there any studies on whether this is an absolute or indexed measure? What I mean is that its been discovered that though chess grandmasters follow the 7+-2 rule in short-term memory, they recall chunks of moves rather than individual steps. Thus 2 bits of information might be considered a compact index into a sparse information space. The more specialised a person, the more complex representation can be used.
The brain is interesting in that it generally filters out daily dross based on preconceptions and training. Flemming recognised the connection between mould and lack of bateria whereas the average lab cleaner would have thrown away the contamination without realising the significance. Thus though the memory may be limited in volume, the extreme selectivity and pattern association capabilities gives humans certain advantage in certain areas of processing. If we consider that the 2 bits can be a useful index/catalog, then you can leave the details up to a computer.
Cognitive science is going to be a fun topic over the next few decades.
LL
How to reconcile OpenSource, intellectual property and profitability?
... errr ... venture capitalists betting as much on the quality of the people in the management team as on the actual technology.
Part of the misunderstanding is defining exactly what is software. Some people view it as a finished tangible product, much like a car whereas other view it as an on-going conversation with multiple sub-languages and cultural variants. With this in mind, this gives rise to some fundamental differences with traditional economics which relies on artificial scarcity (e.g. a monopolist has the ability to exclude competitors and thus limit the avaiability). However, the rather striking effect is that the existance of any "free" software eventually eliminates any for-fee substitute (e.g. MS Explorer vs Netscape).
Given the rather constraits of everyone having equal access to the source, it is still possible to create sustainable business models. Ghostscript is an example where the author creates commercial versions, then turns them into GPL versions after a few years. The other mechanism is to turn the strengths of OpenSource into marketing points (trust, near zero-defect, rapid development cycle, knowledgeable pool of custom programmers for-hire) and become the low-cost and efficient supplier of a market segment. Modern intellectual property is a bit of an illusion as given that all software technology is based on previous tools, it is very hard to come up with a unique solution that 10 million other software developers won't encounter within another 2-3 years. Unlike hardware products, all the minor coding variations can be easily parameterised and investigated reducing the scope for differentiation (how many ways are there really of expressing a spreadsheet?)
Capitalism is a rather complex beast but the essential point is that it is the efficient allocation of surplus time (traditionally stored/measured in dollars) to create future economic goods. The $1 that a farmer saves by not buying a lotto ticket can be invested in funding a startup which (hopefully) can return $2 (or more) to buy future consumer products. With the limiting factor in software development being talent, the OpenSource model is one solution to give any programmer a go at improving software that they are interested in. In this sense, with open peer review, it is closer to a meritocracy than any other social system. If someone comes up with a brillant new hack to double the speed of Quake, I'm sure ID software would hire them in a flash. You can already see a similar situation with vulture
The nature of the internet does change the traditional recruitment rules in that it opens up new sources of talent worldwide (note that Mexico has installed Linux in its education system which might lead to some interesting long-term effects). As usual, some companies will adapt and prosper, others will wither and get absorbed.
I look forward to an interesting new decade.
LL
While the bounty or sponsor model have commercial overtones, I would suggest that the {rich individual | corporate} patron model so prevalent in the Renaisance era is closer to the gift economy that Raymond espouses. This is where the patron demonstrates their {benevolance | lucre} by funding artistic works for public display (Sistine chapel, etc). Given that the highest levels of programming is more akin to art rather than engineering, I would speculate this may be more attractive to the highly talented segment of the hacker community.
The biggest problem I see is that in a rather large but closed world, it takes a while for "names" to be recognised as stellar qualities (e.g. Linus) and the reputation takes a long time to build up. But maybe this can be viewed as an incentive for people to contribute to OpenSource in that by associating a name with a successful project, they can effectively provide an open job resume to the world? Much like the journey-man stage of publicly displaying goods for review by the customer, it can be viewed as a necessary stage of development towards mastering a field. I can forsee the day when employees hand-pick programmers based on recognised contributions to an OpenSource program which clearly demonstrates their level of skill. To support this model does require some work in that OpenSource projects have a clear attribution history of contributors and their degree of worth (e.g. hacker X contributed n% to the design and high-level coding).
The biggest hinderance is that nobody has resolved the inherent conflicts with existing commercial models.
LL
Ummmm, correct me if I'm wrong but ....
t ml
:-(.
:-).
....
... according to what some people claim, the US federal government is technically a political church trust. For details, see
http://www.ptialaska.net/~swampy/amend_14/usa.h
I'm not sure exactly how to interpret this (any legal eagles out there?) but my (entirely personal and thus biased) observation is that the US acts more like a megacorporation with all citizens as shareholders, Congress as directors, and president as CEO, and with the wonders of modern capitalism, one dollar, one vote
Of course as Peter Drucker once said, governments do two things well, wage war and inflate the currency. In corporate parlance, win marketshare and dilute earnings
As an aside, this may offer a rather interesting way of sidestepping the mess caused by the software patent system (ie enforced monopolies for the first to come up with an idea) with something more commonsense based on common law with full disclosure (OpenSource) between two parties. Oh well, one can dream
LL
Both SGI and IBM are smart in their own ways. The limitation nowadays is not hardware (constrained by physics and economics) or technology (can always be bought out) but human talent. Given the increasing software complexity, it can take years for people to learn and master sophisticated software packages and their APIs. Moore's Law dictates that whatever you buy now can be bought cheaper next year. Hence anything developed at the high end will eventually filter down to the PC level. Unfortunately human learning curves moves slower than hardware evolution cycles. The power of visualisation packages such as AVS is in the module libraries that the community develops and contributes as once the critical mass is there, it is very hard to change habits or programming practices. In a way, releasing OpenSource is like force-growing a new ecosystem niche with the hope that enough talent will feed on the transplanted energy and expand the overall food chain (always a problem in small specialised fields) in the hopes that the colony becomes self-sustaining (creating future demand for similar food). The path to growth for both IBM and SGI is to move their "proprietary" IP into mainstream as quickly as possible to extract maximum benefit from their R&D dollars. For this, they need to win over developers and new power users of which academia is a unique training ground.
I can see the day when companies attempt to buyout emerging graduates as without a constant source of talent, any company will eventually wither under the winds of global competition.
LL
Software doesn't exist in isolation, in some ways it reflects aspects of society. Just as there are non-profit organisations aiming to look after the underpriviledged, there are megacorporations interested in using IT to gain a competitive edge and monetary value is a measure of their efficiency. GNU/Linux has its major strength in providing basic civic infrastructure for the computer world (standards/protocols/education). The corporate domain should then be welcome to add value with the understanding that the quality standards are increasing all the time from volunteer work. Its just like giving people the choice between public and private education, you pay more for 'perceived' extras (not to disparange anyone, there are some very good public schools out there).
Having a diverse software ecosystem creates a robust and resilient playing field. Dinosaurs can coexist with monkeys, afterall I believe a few (rare) places still use punch cards. Given the rapid technology churn and obsolescence, there is a need for at least a stable foundation for beginning firms to practice on before hitting the brutal business world.
The biggest problem is getting enough funds to help educate the poor consumer as to the reality of buying software as currently there is no real ombusman role. While marketing adds to the cost, it is a way to help publicise good technology that would otherwise not survive. Perhaps there is some call for a social contract? If a for-profit corporation uses GPL software, perhaps it be suggested that they donate 5% of net profits/benefits to a charitable organisation of the author's choice (e.g. FSF).
LL
IMHO, one of the main strengths of OSS is that the developers are also the users, as such the feedback is (nearly) immediate. Consequently people only develop systems which are stable and flexible because it has direct benefits. After all, would you buy food from a chef that doesn't sample their own food? The problem comes in that by having users 1 or more step removed from the development cycle, it is difficult to produce systems that fit their needs perfectly. The concept of a free market is that it sends "pricing signals" to producers to identify important traits. Hence, from the success of integrated packages, one can say that interoperability between major packages is a highly desired trait. Now free markets are predicated on a large smaple of consumers having informed choice. If they are misinformed (hype/FUD/fraud) or restricted in choice (monopoly, trade barriers, quotas) then the signals are distorted.
OSS has shown the value of global software development, combining the 'best' memes from multiple talents. I suspect the 'best' business model that works in real-life is still yet to be found. Effectively for a points system you are asking for an economic metric that measures talent/time. But here it hits problems such as multiple value systems. What does a point mean to a 16 year old teenager? The value of having hard dollars is that it is a commonly agreed upon unit of interchange and is also a substitute for future needs, rewarding entrepreneurs that best guess what future consumers want.
Hackers tend to develop stuff which automate the boring administrative tasks, leaving their skills for the more fun and creative exploratory tasks. As such, what one releases, has immediate revelance to peers. The other difference with commercial systems is that there is no fear of competition because the real talent is in the living knowledge of the developers.
Now, based on these personal observations, I would suggest some alternative systems:
1) purely commercial - trade OSS for options/shares in commercial outfit
2) casual contractual - set up a virtual company to offer customisation and support for individual corporate needs
3) purely altruistic - social contract, request 5% of net profits be donated to charitable organisation of choice
OSS is still an evolving phenomena so it will be interesting to see what will work in practice.
LL
Is there need to call for OpenSource Jihad against every commercial company? If a group of people have invested time and talent into developing a product, should they not enjoy some return on their efforts? If there is serious need to create GPL versions, then it should be a 'simple' matter to ask for enough donations to 'buy out' the company (ie discounted sum of future revenues) and make the software completely free. This is a straight-forward money for time tradeoff, afterall you are asking people to volunteer their precious time.
One has to keep in mind what the civic outcomes of OpenSource project observed as being currently successful are
1) educational/intellectual curiosity (unis)
2) common standards (e.g. w3c)
3) not-for-profit utilities or public awareness
Software development is a complex and expensive business. Effectively by releasing an OpenSource version, you are undermining the value of the commercial alternative. This serves a long-term purpose of continually raising the quality bar and forcing companies to keep on innovating but it should not be viewed as a call to arms to duplicate everything in sight. OpenSource can play a role in keeping in check outright exploitation but it should be applied with some awareness as to the consequences.
LL
Another way of looking at the problem is that you get paid as recompense for time taken away from doing 'fun' stuff.
.... Though OpenSource can be a very effective creative force, somehow it must coexist within a very consumer-oriented society where most people do not earn >$100K programming salaries. 'Suits' are then charged with the responsibility of trying to squeeze more efficiency and defined outputs from the system. If OpenSource represents 'free energy' then it will be 'exploited' by those in a position to do so.
Let's face it, most work can be placed in the tedious but needs to be done category. User support, maintainence, integration, documentation, etc
The interesting factor is what conditions external groups place on the 'gift culture' in an attempt to control conditions for their own benefit. For example, there is a certain amount of 'reciprocracy' in the system where it is expected that on receiving a gift, there will be a balanced contribution by the receivee in the future. In a system where there are different values (suits vs hackers), asymetric expectations can cause tension.
Designing a Formula 1 car is fun and sexy, but someone has to do boring tractors otherwise nobody eats. Given that the biggest bottleneck at the moment is lack of trained people, OpenSource makes sense in optimising the spread of knowledge. On the other hand, if everyone can program, I would suspect a proprietary system will be 'superior' as it retains a 'competitive edge' for a select group. What works in reality can be quite complex and it's probably best to be flexible and choose whatever system best suits the circumstances.
LL
There's no way that Australia can absorb a Microsoft. We've got a serious shortage of programmers already (must be due to the real sand, real sun, and real surf up here in sunny Queensland) and we're only a 12th the population size of the US. They can take over Cambridge or Peking instead where they've already got labs.
Lawrence