Magic Boiler - Foreword, Targeter Eric Raymond (June 1999) [AKA] Rover Hansb IASC and other translation
-------------------------------------------------- ------------------------------
This paper analyzes the economic foundation of open source code phenomenon in the development of open source. We first overthrew some popular myths about the investment and software price structure of software development, gave a game rule analysis on open source code collaboration stability. We gave the development model of nine open source development, two of which are not profitable, seven are profitable. Then we develop a qualitative theory, indicating when the closed code is economically reasonable. Then we examined the earnings methodology of several new open source development in the current market, including the introduction of sponsorship systems and task markets. We finally made conclusions and tried to make some predictions in the future.
Directory ==========================
Near magical
2. Beyond the talent of the master
3. The illusion of manufacturing
4. "Information Need for Free"
5. Refuting the public tragedy
6. Reasons for closing the source code
7. Use value fundraising model 7.1 Apache case: cost sharing 7.2 Cisco case: risk dispersion
8. Why do you have problems with sales?
9. Indirect sales value model 9.1 failed leader / market positioner 9.2 Candy Strategy 9.3 Dedicated recipes
10. When is open, when is it closed 10.1 What profitable? 10.2 How do they do each other? 10.3 doom: A case of study 10.4 know when to let go
11. Business operation of open original code
12. Successful replication
13. Open R & D and Second Developments
14. Thus and each
15. Conclusion: After free software change
16. References and Acknowledgments
17. Appendix: Why is the seller of the seller?
18. This document revised record
Near magical
In the myth of Wales, the Ceridwen goddess has a huge pot. When the female meditation is only the spell she knows, the pot will become a wonderful food. In modern science, Buckminster Fuller proposed a "short" concept that technology will become more effective and cheap in the case of earlier physical resources investment. Arthur C. Clarke pointed out that "Any advanced technology is not born with magic", linking them. For many people, the success of open source communities seems to be unbelievable. High quality software is free, in a realistic world full of competition and scarce, this does not seem to continue, but it is not bad. Where is it? Is Ceridwen's cauldron just a small trick? If not, in this case, "short" is working - what spells do you think?
2. Beyond the expert
Open source code culture has definitely makes many people who have learned software development confused. "Big Church and Market" describes how the scattered collaborative software development is effective overturned the law of Brooks, which has an unprecedented development method of unprecedented reliability and quality. "Pioneering Intelligence Domain" discloses social kinetics in the market model development style, which should be understood by humanologists' terms rather than routine exchange of economic terms, in this culture, members Competition in contributing. In this paper, we will start to overthrow some popular myths about software production economics; then analyze the "Cathedral and Market" and "Pioneering Intelligence Domain" two articles analysis, development one New conceptual tools, understand the reasons for the opening of the open source developer's gift and culture can continue in the exchange economy. Do not open the topic first, along the above clue analysis, we need to abandon (at least temporarily ignored) in the "Give Culture" level. The existence of cultures in "Pioneering Intelligence Domain" is based on the substance information required for survival, so that social exchange is not very important; this analysis is very persuasive in pure spiritual world. However, this explanation seems to be weak in response to the integrated economic environment of most open source developers in real life. For many people, social exchange is still driving the driving force they work hard, but they have gradually lost their attractions. It is necessary to find enough reasons for their behavior in the scarce economics of resources, in order to make these behaviors to be based in cultural materials.
Therefore, we will now think of collaboration and switching mode for maintaining open source development (from the field of lack of economics throughout the resource). In the process of analysis, through in-depth analysis and listed examples, we also answer the very practical problem: "How do I make money by open source?" However, this problem is based on the universal software development economic model contrary to the nature of software development. First, we need to show many thinking misunderstandings after hiding this problem.
(There is also the last need to be explained before the analysis is: discussions and advocates the open source development model in this article cannot be understood as the complete negation of the closed source mode, and there is no opposition to the existing software intellectual property system. Not a selfless appeal to "sharing". Although some people in the development of the source development group are still passionate about these discussions, since the "Cathedral and Market", the experience has clearly shown that these debates are not necessary. Important is Open source development models and economic benefits can create better quality, higher reliability, lower cost, and can choose more good products.)
3. The illusion of manufacturing
We need to pay attention to computer programs and other types of tools and capital goods, there are two economic value: use value and sales value.
The value of the program is its economic value of the tool; sales value is its value as a commodity value. (Specialty in economics, sales value is the final value of the product, the value of use is the product intermediate value)
When most people say the software industry, they always analyze the "factory model" economy with the following characteristics:
1. Most developers' labor is paid by the sales value income
2. Software sales value is proportional to development costs (for example, the resource cost required for function replication) and use value.
In other words, people have strong thinking inertia to assume that software has the characteristics of standard industrial products. But these two assumptions are wrong.
First, the code to be sold is only the hoss of the iceberg in the programming industry. In the early days of the microcomputer, you generally believe that the world is written in the bank and insurance companies. This is no longer true - now other industries More and more increased the strength of software development, the proportion of the financial industry to decline - but we will still see about 95% of the code in the short term is the company's internal preparation.
These code includes most MIS, financial and database software customized by medium or large-scale companies. Including professional technical code such as device drivers (almost no one is driving to make money by selling equipment, this will discuss later); Various embedded codes including increasing CNC machines - from mechanical tools and jet passenger, cars, microwaves and even toaster.
Most of this internal code is integrated with its environment, replication and reuse is very difficult (regardless of the environment of the business office or the fuel system of the combined harvester). Therefore, once the environment changes, you need to do many work to synchronize the software. This work is "maintenance". Any software engineer or system analyst will tell you that this is the source of most wages of programmers (more than 75%). Therefore, most program employees spend on the internal code written and maintained (of course, most of the programmers will be born) - the readers may be happy to check the "Cuisher Information" part on the newspaper. Programming work list checks.
I hope that readers will try to browse local newspaper recruitment information, look at programming. Data processing, and software engineering projects that include software development work. By using these work, it is used to use or sell, you will be deeply Inspire.
Obviously, even if "sales" defines the maximum range, 20 people still have at least 19 by usage value (as the middle value of the product). Why do we think that the software industry takes only 5% of the sales value driver Cause. Note that the analysis of other parts of this article is not completely dependent on this number; even if this number reaches 15% or even 20%, the economic inference is still eight or nine.
(When I gave a speech at the technical discussion, I often started two questions: how much money pays for writing software, and how much salary relies on software sales value. The first question is very popular, There are very few second problems, big and quantity listeners are very surprised by this problem)
Second, through the investigation of actual customer behavior, software sales value is easily overthrown with its development and upgrade costs. Develop and upgrade costs (for discounts) account for a large proportion - food, car, Mechanical tools, even have many intangible products - such as music, maps or database data. This product can still be kept even after the producer is closed after the production of its sales value.
It is stark contrast to the above-mentioned formation. When a software product producer is closed (or if the product development is terminated), there is almost no customer is willing to spend money, regardless of the theoretical value or the development cost of the same functional product. How high. (To test this statement, go to your nearby software store discount counter to see :-))
When the producer fails, the retailer's behavior is very revelation. They know that some producers don't know, they know: The price of customers will cost to a large extent by the seller's future service decision. (Here The 'Services' is widely understood as perfect, upgrade, and subsequent products).
In other words, the software is mainly a stable service industry, and it is considered that it is a manufacturative industry that has no reason to illusion.
In addition, check why we have these inertial thinking. They may come from sales products that software producers have promoted, these is a small part of the software industry, and the only part of promotion, most of them are obvious and The product of the heavy advertisement is a short-term product, which is like a game. They have hardly need to provide follow-up services (except for contract regulations)
In addition, it is worth noting that the price system advocated by the manufacturing illusion is in fact the bottom line that keeps the development budget does not collapse. Since (like a general considered) 75% of the typical software product cycle costs in maintenance, debugging and extension Then, the typical type of pricing strategy that only uses high price, very low-related service costs, will only lead to services that are poor in all aspects.
The loss of users is that even if the software is a service industry, the factory model has caused the producer to reduce the quality of service. If the producer makes money by selling Bit, a lot of effort is manufacturing bits and selling them; helping the service part, because it is not a profit The center will become a little effort and resource that only pays, in order to avoid irritating users.
On the other hand, most producers use this factory model to lead to long-term failure. Provide funds to meet the fixed price products required to meet unlimited after-sales service and technical support, only in the market that expands is fast enough - its past Sales and future revenues can meet the cost of support and survival cycles - can survive. Once the market matures and sales declines, maintaining livelihoods, most producers don't have any other choice except for the expenses of the individual products.
Whether it is direct (abolished product) or indirect (support is poor), customers will push customers to competitors (because these behaviors have damaged the expected value attached to the service product). In the short term, they can be released by the revised BUG. New products avoid this trap. In the long run, the only thing to avoid traps may be effective market monopoly for the industry. Eventually, only the only survivor.
In fact, we have repeatedly seen this lack of supported patterns in some market environments very strong competitors, (this model is particularly profound, including people who have survived computers, including personal operating systems, Word processing, general financial procedures, or commercial software). This incorrect motivation comes from the winner of the winner, and even if you are a winner's customer. If it is not a factory mode, what is it? In order to effectively control the real cost of software survival cycle ("effective" in economics and informal occasions), we need a price system based on service contracts, contracts, and buyers and sellers. Therefore, under the free market conditions for the purpose of benefit, we can sustain the final price system that mostly matured software industries.
Why is it possible to grow the status of open source software, not just technology, but also the challenge of mainstream order? The above content gives us some revelations. Software development "free" will push us to the world dominated by service, while There are many vulnerable ways to expose the sales of closed source products.
The concept of "free" is easily misunderstood as other meaning. Reducing product costs will result in growth of the entire basis of the support software industry, rather than lowering. Only when the price of the car is lowered, the needs of the car will rise - this is why it is in the open source world, and the other 5% of the programs payable according to the sales value. In the refrer, there is a loss Not a programmer but the investor who will bet on a closed source strategy without seeing the situation.
4. "Information should be free"
It is often that people who think that open source economy are often confused by another myth. That is "information should be free." This often explains with the digital information product's replication marginal cost, this explanation I suggest that the price seems to be zero.
In fact, you only need to consider the value of the account password such as Tibet, Switzerland bank, or the value of the computer service, so it is easy to see this myth. Even if these confirmation information can be replicated without any cost, The confirmed object cannot be copied. That is, the non-zero edge cost is inherited by those confirmation information.
The main purpose of this myth is to declare the discussion of the economic value of the open source; it is still like the software that we will see later, even if the software is in line with manufacturing products (non-zero) value structures, it is still the case. So we don't have to drill whether the software should be free.
5. Refuting the public tragedy
Questioning mainstream mode to see if we can build another model - "For the forces that support the open source collaboration, it makes a strong economic interpretation.
This problem needs to be examined from two different aspects. One aspect is to explain the individual behavior of those who contribute to the open source; on the other hand, we need to understand the economy of open source projects such as Linux and Apache. power.
Hardin's famous fable tells us that a rural farmers have a public green space. They graze livestock there. But grazing makes publicity degradation, tearing the turf, leaving muddy, it is difficult to recover, if there is no right to allocate grazing Agreement (or agreed) to prevent excessive grazing; all pastors will agree to increase the number of livestock as quickly as possible to extract the greatest profit before the public green space becomes a quagmire.
Most people use the intuition model like this. This is not an open source - they are (for short-use) free knights, rather than (excessive) excess public goods - economic problems correct judgment However, I have heard similar opinions behind most of the opposing opposition.
The public owned tragedy will only have three results. One is a quagmire; one is for the interests of the villagers, mandatory use of certain distribution agreements (communist solutions); the third is public break, villagers Each of the projects, protecting yourself a small grass (private solution).
When people can apply this model to open source cooperation, it is expected to have only very unstable short half-life. Because there is no obvious way to force programmers working on the Internet to perform working time allocation strategies, this The mode is asserted that the public will break, and the result is a rapid decrease in various closed code software and feedback to public workload.
In fact, experience clearly shows the reverse trend. The intensity and depth of open source development (data statistics by Matalab and Freshmeat.Net) are increased stable. Very obvious, these are "public The tragedy "mode cannot describe the development of the situation.
Part of the answer is based on the fact that software use does not reduce its value. In fact, for open source software, when users are corrected and characteristic (code patch), the extensive use of software will also Increase its value. The utility tragedy was subverted, the more grazing, the higher the grass.
Another part of the answer is based on the market value of small patchs made by the public source basis. Suppose I wrote a correction for an annoying bug, and some people think this correction value; how can I get from those people? Money? For this small amount, it is usually the right payment. The routine payment system is so expensive to become a real problem. It is not just difficult to charge than the price, maybe how to pricing. Let us think about it, Assume that there is already theoretically perfect small payment system on the Internet - security, convenient, and don't need more fees. And you wrote a patch called "a correction of the Linux kernel". How much is your price? When the potential purchaser has not seen the patch, how do they determine the value not worth paying?
Our problem is like f.a.hayek's "calculation problem" in the haha mirror - it is like a supermarket, that is, it is necessary to estimate the function value of the patch, and believe that pricing is reasonable to promote transactions.
Unfortunately, the supermarket method has a series of shortcomings, so the authors of the patch have two options: lying on the patch to collect money, or throw it out for free. The second choice will be all. Second It is also possible, but it will prompt others to provide reciprocal giving to solve the problem of the headache above the hacker. The second obvious selfless choice, in this game situation, the competition is actually selfish.
When analyzing this cooperation, the issue of free software is important (they may work in poverty, or without sufficient returns), this is not determined by the number of end users. Open source project The cost of complexity and communication is almost completely and participant in the number of functions; there is no benefit of the end users who have almost not seeing the source code. This will only increase in the project mailing list. There is a frequency in the middle boring problem, but build a list of related common problems, ignores those who obviously don't read FAQ (in fact this is already a general practice), it can easily solve this problem.
The real most important free software development problem of open source software is the grinding cost of submitting patch functions. Possible contributors have a small gain in the prestige (see "open up intelligence" one article), without money compensation, thinking "It's not worth submitted to this revision, because I have to play patch, write modification records, sign named on the FSF task file ..." For this reason, the project with a large contributor (after success) is very strong. With Conversely, there are many projects with many constraints that need to have a contributor from start to the end. This grinding cost is like politics. In short, the free software project itself can explain to you, why, no The organization's Linux culture, the effort to closely organize and centrally managed BSD projects, and can attract the intention of cooperation energy; and why the Free Software Foundation is also weakened at the time of the rise of Linux.
This road is much better. However, this is just after the hacker wrote the patch and announced the incoming explanation after this patch. The other half of the answer we need is to write this patch for the JRH. Not a closed source program for sales returns. It makes an economic interpretation. What business model creates an environment where open source development is prosperous?
6. Reasons for closing the source code
Before you classify the open source business model, we should first consider the closed price. When we enclose the source code, what are we protecting?
For example, you hired someone to write and organize a settlement software for your business, then the open source is more than the open source, the enclosed source code will not help solve the problem. If you want to enclose the source code, The only reasonable reason is that you want to sell this software to others, or don't let your competitors use it.
The more obvious reason is that you are in protecting sales value, but this is meaningless to 95% of the software used internally. So do you have anything else?
The second cause (maintaining competitive advantage) is still to be inspected. If you open the settlement software, it is popular, and it has been improved since the society. Now, your competitors have also started using it, he did not The cost of spending is benefited, and it affects your business. Is this a reason to oppose open source?
It may not be - may not. The real problem is whether you are more than the competition losses brought by those who have been brought to those who have been governed. Many people tend to be pale excuses for such transactions, the method is : (a) avoiding the functionality of the feature obtained from additional development help. (b) Do not think that development costs are lowered, but assume that you are also responsible for these development costs, so use them as The cost of open source (if you choose this) is wrong.
There is also a ridiculous reason. For example, you may mistake your commercial system more secure, not easy to crack or break into. If so, I suggest you find one now. Password experts to diagnose your system. People who really guess are all known to confuse the security of the enclosure, because this is to learn from the painful lesson. Safety is one aspect of reliability; Only those algorithms and code implementations that are completely checked can be believed to be safe. 7. Use value fund-raising model
The difference between the value of value and sales makes us noticed a basic fact that only sales value itself has been threatened from the transition from enclosing the original code to open the original code; the value is not.
If the value is used, not the exchange value, it is indeed the fundamental driving force of the software development; and the development of the open original code is indeed more influential and more efficient than the original code, then we should look forward to looking for an environment. In this environment, light is used value, which is capable of fully promoting the opening of the original code forward.
In fact, such a few environmental models are not difficult to find. In such a model, the full-time developer of the open original code can be implemented by the value of (open original code).
7.1 Aapache case: (value sharing)
If you are serving a business company with an efficient high-reliability web server. Perhaps this server is used for e-commerce services, perhaps as a high-visibility media output device that sells advertisements, perhaps just use to build a portal site. You need a day 7 hours online time, you need speed, and normative.
So what should you do? Some basic strategies here can be used for you:
Buy a private web server so that you are adventurizing the propaganda of the seller consistent with your needs. You can provide a perfect guarantee to provide a perfect guarantee in adventurous. Even if it is assumed to be guaranteed, the web server itself will have problems due to lack of normative services. You can only maintain your servers through the seller's choice of several tools. This way of purchasing private servers is not a very large way!
Do one yourself! To make your own network server is still a modified method that is still not ignored; the web server is not very complicated, of course, it is simpler than the browser. A specially used web server can do functional specials, but very well. If you take this way, you can get the features you need and your own specifications, although you have to pay in the process of its upgrade. Perhaps your company will find this server with such or such problems after you leave or retire.
Participate in the Apache group! The Apache server is written out with a team communication through the Internet - team members are system administrators. They believe that more sensible practices are to write their capacity to write, and improve the code set in a single direction rather than going It takes time to write a completely unrelated code. The result of doing this is that they can play the advantages of "do one" and large-scale mass test code.
It is obvious that the advantages of choosing the Apache team are obvious. There are more obvious in the end, and it can be judged according to the weekly review of Netcraft. Netcraft says the Apache server has been stably market share from its birth. In June 1999, various versions of Apache accounted for 61% of the market share
In general, Apache's story provides a model: software users have discovered this model by supporting open original code plans, they find that they can bring them more and better software with minimal expenses, than others Any method is effective.
7.2 Cisco's cases: Risk booth
Programmers for two Cisco (Network Product Manufacturers) were allocated to write a distributed printing system program code for Cisco's partner network. This project is challenging. This system is to print anything on any printer on this network (while users and printers may only be partitioned or thus separated by thousands of kilometers). When the printer has no paper or other emergency system to be able to guide the task Another nearby printer. The system must also report this burst time to the printer administrator. The two of them have made some very good modifications on UNIX, plus some of the original languages of the package, but the next problem is coming.
The problem is that two programmers are reluctant to stay in Cisco. As a result, the two programmers will leave, and the software will not maintain and "rotten" (that is, it is unable to meet the requirements of changing requirements in practical applications). No one is willing to see such a situation in his own or work, the two programmers also think they have done Cisco companies asking them what they do, other issues are not their work range.
So they ran to their manager to open this print software's original code. They think that Cisco will not only lose anything, not only. Cisco enables compensation for the loss of software original personnel by collaborating the development of users and software development collaborators.
Cisco's story leads to another model: the original code open has made the development of a software is shared by many collaborators and the investment is small. All groups discover the opening of the original code, and the existence of a member of the community that is independent of each other will provide a risk-free development environment, and this environment is commercial value - it can make money to make yourself!
8. Why do you have problems with sales?
Open source makes it difficult to obtain software sales profit directly. Difficulties are not from technology, because the source code and executable code are easy to copy, and the copyright law and the constraints of the license method make it difficult to obtain sales profit by open source software than closed source software.
Real difficulties come from the license itself maintained by the development of open source. Because of three mutually promoted reasons, most open source licenses prohibit the use of rights to users, and modify the rights of software, to avoid direct profits with open source software. In order to better understand these reasons, we must have the social background-hacker culture involved in these licenses (you can access the following Website: http://www.tuxedo.org/~esr/faqs/ Hacker-HOWTO.HTML > Do a discussion.
The reason is not related to the hostility of the market. Although this misunderstanding is widely circulated outside the hacking circle. Do not exclude a small part of the hacker that has been hostile to commercial motivation, but most of the hackers are still willing to cooperate with some profitable Linux integrators (such as Red Hat, SUSE, CALDERA). This also shows that most of the hackers will be willing to cooperate with the merchant as long as they meet their will. In this way, the truly reason why the hackers hostile to obtain direct profits is very interesting.
One of the reasons, the principle of peer. Most open source developers allow others to use their results to obtain benefits, and many open source developers also stipulate that some side (except sometimes the source code developers) will take advantage of the privileged position. As long as the hackers have intended to profit from the software or patch they develop, they are generally willing to cooperate with him to work together.
The second, unexpected consequences. The hackers have found that the commercial application and sales of the software in the license (the practice usually used in order to obtain sales profits) will make interpersonal relationships indifference. One of the special cases is the so-called "pirated disc", which should be encouraged, but now it is considered to be illegal and unethical. In general, the restrictions on users' use, sales, modification, distribution software (and other complex rights in the copyright agreement) will lead to people, and always worry that they will break the law (this fear will use people. The increase in the package is more intensified). This is undoubtedly very bad, therefore simplifies the license, and the restrictions in the license have become a general trend. The third reason is the most critical reason, that is, code sharing. This gift and culture have vivid descriptions in the "Pioneering Intelligence Domain". Some provisions of certain license systems are used to protect intellectual property or limit directly to acquire sales profits, making people unable to realize code sharing, such as Sun Company's JINI & JAVA "Community Resources" license). However, the code sharing is considered to be the last rescue "Straw" ("Pioneering Intelligence Domain", the interpretation of the large section of the large section of the large section of the medium), when the software maintenance is unable to undertake or give up the maintenance of the code (than the side said A very closed license), the code sharing is very critical.
The hacker group has a compromise for the principle of peer, so they can tolerate some privileges of some privileges of the source creators like NETSCAPE NPL (NPL clearly defining products that do not allow non-public source Mozilla code). For the second reason, less is less. It is a reason why Sun's Java and Jini Community license plans to be a widely objected by Sun's Java and Jini Community license.
The above reasons explained the terms in the open source definition. These provisions express the thoughts of hackers from some typical free software copyright agreements such as GPL protocols, BSD protocols, MIT protocols, and Artistic protocols. They (although not intentional, but objective) make direct profits. Extremely difficult.
9. Indirect sales value model
However, there is still a way to explore the market related to software services to obtain indirect sales value. There are five known and two models that are exploring (more new development models that may develop more in the future).
9.1 Failed Leader / Market Positioner
In this mode, an open source software is used to create or maintain a marketplace location for proprietary software that directly generates income. In most common situations, client software for open source, drives sales of server software, or increases access to the portal / advertising income.
Netscape uses this strategy when the source code of the Mozilla browser in 1998 is used. The business income of their browser only accounts for 13% of total revenue, and the market share is still declining after Microsoft started Internet Explorer. IE powerful marketing (and its bundled strategy later became the core problem of anti-Turas case) quickly swallowed the market share of Netscape browsers, causing Microsoft to try to monopolize browser markets and use Microsoft to add HTML "standards" ", Forming a situation in which Netscape drives out of the server market.
Netscape effectively prevents the possibility of Microsoft monopoly browsers by opening the source code of the Netscape browser. They expect open source collaboration to accelerate the development and testing of the browser, and hope to reduce the development speed of Microsoft IE to prevent it from defining HTML standards alone.
This strategy has taken effect. In November 1998, Netscape actually started a market share from IE. When Netscape was acquired by AOL in early 1999, it was obvious that it would be obvious from AOL, this can be obvious from AOL action. AOL first commitment is to continue to support the Mozilla program, although she is still in Alpha test stage. 9.2. Candy Strategy
This mode is for hardware manufacturers (the hardware here includes everything from Ethernet or other external devices until the computer system). Market pressure Forces hardware companies to write and maintain software (from device drivers, configuration tools until the entire operating system level), but software itself is not a profit center. It is an expenditure - usually an important expense.
In this case, open source is a good strategy. Since there is no loss of profit, there is no negative impact. The seller gets a miraculous developer team, which is more fast, flexible, and better reliability through peer inspection. And you can get porting of other systems for free. This approach can also improve our customers' trust in the company because customers can customize code according to their needs.
There are some reasons for the source code that is often proposed by the seller. In order not to mix them with more general questions here, I specifically discussed this issue in Appendix.
The "future benefit" of open source is particularly strong in the confectionery strategy. Hardware products have a limited manufacturing and supportful life cycle. After that, users will take care of themselves. But if they can get the source code of the driver, they can be modified as needed, they are more likely to be happy to become the same company.
A very dramatic example of the confectionery mode is the code of the operating system "darwin" of Apple decided to open their Macosx servers in mid-March 1999.
9.3 Deliver recipes, start a hotel
In this mode, open source software has established a market positioning, which is not for the closure source code software as in the failed leader / market positioner mode, but is for the service.
(I used this model called "sending razor, sales blade", but the association of both software and services is not as close as the razor / blade is.)
This is the model used by Red Hat and other Linux publishers. They sell it is not software code itself, but by combining and testing additional value generated by a turnable operating system, this operating system is guaranteed to have a sales and compatible with the same brand's operating system. Additional elements that make up their value include free installation and providing optional continuous technical support contracts.
The ability to creating the open source code is extremely powerful, especially for companies that are born to serve services. A very educated example is Digital Creations, which is an development of Web site design agencies in 1998, specialized in complex databases and transaction sites. Their main tools, the company's knowledgeability - the pearls on the Crown are an object release system, which has had several names, now known as Zope.
When Digital Creations is looking for venture capital, venture capital is carefully estimated that their expected market share, after their human resources and tools, it is recommended to open Zope's source code for Digital Creations.
From the traditional software industry standard, this seems definitely a crazy move. Conventional business schools think that this core knowledge wealth like Zope is a company's handheld pearl, which is not giving up in any case. However, the risk investor considered issues from two interrelated perspectives, one is Zope's true core assets actually is its brain and technology; the second is Zope as a standard for creating a new market, not just It is a secret weapon to produce more value.
In order to see this, please compare two situations. Under normal circumstances, Zope remains secret weapons for Digital Creations. Let us assume that it is a very effective weapon. As a result, the company can deliver high quality software in a short period of time - but no one knows this secret weapon. It is difficult to meet the customer, but it is difficult to build a customer group. However, the venture investor saw the true wealth of the Zope system open source can be the real wealth of Digital Creations - its technical staff produces huge advertising effects. He expects to use Zope's customers to think that experts like Digital Creations will be more efficient than developing their Zope technology than themselves.
A person in charge of Zope has confirmed their open source policy "" Open the door that is not available in many other ways. " Potential customers do respond to this situation - so Digital Creations rapidly develops.
Another very close example is E-Smith
9.4 Additional Products
In this mode, we sell additional products for open source. At low-end markets, sell cups and T-shirts; in high-end markets, the documents and books specially edited and published.
O'Reilly Group is a good example of an additional product company, and he published a lot of excellent open source software references. O'Reilly actually hires and supports some famous open source hackers (such as Larry Wall and Brain Behlendorf) and raises its reputation in the market.
9.5 Free, sell now
In this mode, we publish the executable files and source code of the enclosed license release, but contains a deadline for the closed clause. For example, we can write a license, allowing free distribution software, non-pay business applications, and guarantee that the software will be protected by the developer after the development of the developer or after the development of the developer will be under GPL protection.
In this mode, customers can guarantee that the product can be customized according to their needs because they can get the source code. The future of the product is also guaranteed - the license guarantees that if the startup company fails, the open source community can still take over the product.
Because sales prices and quantities depends on customers' expected value for products, the startup can enjoy better income than the software issued by the closed source code license. Moreover, because the old code is under GPL protection, it can get a serious inspection, troubleshooting and adding other small features, which can reduce the 75% maintenance burden for the original.
This mode was successfully adopted by Aladdin, which created a popular GHOSTScript program (a PostScript interpreter, which can translate PostScript into many printers in the internal language).
The main disadvantage of this model is that the closed terms tend to suppress early peer inspections and participation in the development of products, and at the time of the participation of everyone needed.
9.6 software free, sales brand
This is also a testistic business model. We open a software technology, reserve test packages or a set of compatibility criteria, then sell a brand certification to users to ensure that they are compatible with this technology and other products with such brands.
(This is the way Sun should treat Java and Jini.)
9.7 software free, sales content
Another testistic business model. Imagine some services such as stock information. Value is neither a client software that is no longer a server, but is to provide objective reliable information. So we open all software, sell content subscriptions. Our market is automatically extended when the hacker transplands the client to a new platform or expands it in different ways. (This is why AOL should open its client software.)
10. When is open, when is closed?
After investigating several commercial models developed by open source software development, we can discuss when to open source, when to close the source code, there is a general problem such as economic significance. First, we must figure out how each strategy is profitable.
10.1 What profitable?
Closed source code allows you to collect profits from secret bits; on the other hand, it prevents other peers to test the code. Open source code creates conditions for other peers, and you can't get profits from secret bits.
It is very good to understand from the secret bits; the traditional software business model is built around it. But until recently, the value of other peers inspection code has not been well understood. However, the Linux operating system makes us more clear understanding of the problem, these understandings should be learned from Internet core software and other software engineering branches a few years ago - the peer inspection of open source code is obtained High reliability and high quality software unique retractable methods.
Therefore, in a competitive market, customers who look for high reliability and high quality software will return those open source software developers to reward, and they explore how to maintain in service, added value, and software-related auxiliary markets. A stable revenue loop. This phenomenon is the reason behind Linux, and Linux has developed a blank blank in 1996 to 17% of the commercial server market in 1998, and it seems to occupy this market within two years (early 1999, IDC Predicting Linux will grow in 2003 than the sum of all other operating systems).
A almost important role of open source code is a means of dissemination of open standards, around it builds a market. The dramatic growth of the Internet benefited from no one to have TCP / IP; no one has the right to block the core agreement of the Internet.
The impact of TCP / IP and Linux successful interconnect networks to the world is obvious, and the open system will ultimately reduce the problem of trust peace. If you can see how the underlying structure works, they will It is reasonable to trust it; people prefer an all-equally underlying structure, rather than a certain side with profitable privileges and can apply control under construction.
However, in order to emphasize the impact of the network in order to explain the importance of software users. No software users discard the open source software under the conditions where quality and functionally similar open source software, and choose to close source code software, and have to make yourself have been happy by a supplier monopoly. The more important the software is more important to consumers, the more prominent this problem - the more important it, the more consumers can be tolerated by the other party.
Finally, the important advantage of open source and trust issues is its bright prospects. If the source code is open, even if the issuer has fallen, the customer can still master some resources. This is especially important for the confectionery strategy because the hardware tends to be shorter life cycle, but the role is more common and converted to the growth value of open source code.
10.2 How do they interact?
When the rewards obtained from the secret bit are high from the open source, the source code should be closed from the economic sense. When the revenue obtained from the open source code is higher than from the secret bit, then there is no doubt that the source code is more meaningful.
From the surface, this is a very ordinary idea. But when we pay attention to open source rewards more difficult to measure and expect, it is often underestimated by the return, which is not so strange. In fact, until early 1998 industry mainstream began to reconsider the premise of complying with the Mozilla issued source code, the return of open source has been considered to be zero. So how do we evaluate the return of open source? Generally, this is a difficult problem, but we can handle it like any other prophecy problem. We can start from observing the success and failure of open source code. Try abstract a model, at least give a qualitative feeling, where open source code produces net income for investors or the largest return of commercial operations. Then we use data to refine this model.
From the analysis of the "Cathedral and Market", we can get open source in (a) Reliability / stability / scalability, and (b) the correctness of design and implementation In addition to using other High investment returns are difficult to verify outside of the peer inspection. (Most important programs in practice are in line with the second standard.
When the software is critical to a consumer, consumers have enhanced his interested in open source to avoid being controlled by a monopoly supplier (and therefore raised the market competitiveness of open source manufacturers). Therefore, another standard (c) When software is a very important asset (for example, the MIS department in many companies), the closed source code will push the user to the open source.
In the application area, we see that the open source underlying software creates the results of trust peace, Over time, more customers will attract more customers, thus winning the enclosed source code underlying software; in this rapid expansion Smaller shares in the market are often better than in closed and sluggish markets. Therefore, for the infrastructure software, the open source method is more than a closed source code that utilizes intellectual property to obtain income.
In fact, potential users know its future development capacity according to the publisher's policy, and they are reluctant to accept a monopoly supplier, because this will mean being constrained in situations; unless there is an overwhelming market Force, otherwise you can choose an open source to choose a way to benefit from the closed code - but it is impossible to choose both. (You can see similar situations in other places. For example, users often refuse to purchase separate source of supply.) The neutrality of this situation can eliminate some: places where the network dominates, open source The code seems to be the correct choice.
We can summarize this logic: When (d) creates a public computing and the underlying structure of public computing and communication, open source software seems to have a bigger return than the closure source code software.
Finally, we noticed that the only or unique service provider is more worried about the only or unique service provider, providing unique or unique services, is more worrying that the competitors will imitate their methods. Therefore, when the (e) core method (or function) is part of public knowledge, the open source code is more likely to win.
The Linux system that implements Internet core software, Apache, and ANSI standards is a typical model of five standards analyzed above. After the end of the 19th year of the Imperial Try, such as DECN, XNS, IPX, etc. Evolution path.
On the other hand, open source does not have much significance on companies with software assets with their own unique creation value (strongly satisfying conditions (E)), the following conditions are not applicable to open source, such as software (a) to fail It is relatively irritated, (b) can be verified in a manner other than a peer test, not (c) key transaction, and is not mainly obtained from (d) networking or universally used value. As an extreme example, in early 1999, a company asked me "Do we open source code?", This company writes a software for cutting mode for the sawmill, which can get the largest sheet from the log. My conclusion is "no". The only condition that they are close to the satisfaction is (c); however, it is a skilled operator to manually determine the cutting mode.
It is worth noting that specific products or techniques that meet these conditions will change over time. From the following cases, we will see this.
All in all, the following conditions should be used to adopt open source mode:
(a) Reliability / Stability / Exchangeability is very critical (b) The correctness of design and implementation cannot be easy to verify the user / her business with methods other than other peer inspections. When the key (d) software is used to create a public computing and communication infrastructure (e) key method (or equivalent function) is part of public engineering knowledge
10.3 Doom: A case
ID Software sells the most popular game Doom's history showing how marketing pressure and product evolution change the number of closed source code software relative to open source.
When DOOM was first released at the end of 1993, its subjective perspective, real-time animation was extremely unique (conditional (E) opposite). Not only because of its kind visual effect, but no one knows how they do this effect on the low processor for a long time. These secrets can get very important benefits. Moreover, the potential income of open source code is very low. As a separate game, this software (a) has a small price of its fault, (b) is not very difficult to verify, (c) is not critical to any user, (d) does not benefit from The internet. So doom has become an enclosed source code is economically reasonable.
However, the market around DOOM is not stationary. Competitors invented the equivalent function of its animation technology, other "subjective shooting" games such as Duke Nukem, etc. began. When these games erodes the market share of Doom, the revenue of the secret bit begins to decline.
On the other hand, the efforts of extended market share have brought new technical challenges - better reliability, more game features, larger user groups, and cross-platform. With the "DeathMatch" multiplayer mode and DOOM game service, the market began to show the dependence on the network. All these needs require ID companies to spend more energy in the following versions of the game.
All of these trends have improved the return of open source. At a point of return curve, open source code becomes ID company economically reasonable choice, they can benefit from the second market such as game extension. At some time after this, things did happen. At the end of 1997, the complete source code of DOOM was publicly released.
10.4. I know when to let go
Doom is an interesting case because it is neither an operating system is not a communication / network software; therefore it is far from the usual obvious example of open source code. Indeed, the lifecycle of Doom, including intersections, can be used as a typical application of today's code ecology - in this ecological environment, communication and distribution computing software requires higher robust / reliability / expandability, only Verify through peer tests and often transcend the boundaries between the technical environment and competitors (including trust peace, etc.).
Doom evolved from a single game to DeathMatch mode. Network calculation is increasingly important. The same trend can be seen from the most important business app, such as the ERP system. Business networks link suppliers and customers together - of course, they are included in the architecture of the entire web. This situation can be seen everywhere, and the return of open source code has increased steadily. If the current trend continues, the core challenge of the next century software technology and product management will be to know when to let go - when the closed source code is changed to an open source architecture, so that the benefits of the peer inspection, and Get higher returns from services and other second markets.
Everyone is obviously not to be too far from the intersection in any direction. In addition to this, wait for a long time to face serious risks - you may be flattened by a competitor on the same market to open source code.
The reason why this problem is that the user group and expert group that can be attracted to an open source collaborator of a type of product is limited, and these people are difficult to transfer. If the two functions are basically the same, the source is open, then the first open source is more likely to attract more users and more the most passionate cooperative developers; after opening, they have to eat leftovers. . The attracted person is difficult to transfer because the user is already familiar with the software, and the developer has invested a lot of time on the code.
11. Open original code business operation
In the community of open original code, it is usually organized to organize its own business activities in a way that tends to enhance open original production benefits. Especially in the world of Linux, there is a fact that there is an important economic significance, that is, there are many distributors who compete with each other, and they form a separate level, independent level.
Developers write original code, and these original code can be downloaded on the interconnection. Each publisher selects some from these downloadable original code and integrates, packages, and registers trademarks, and finally buy it to customers. Users can choose the publisher's product, or use the original code directly from the developer's website to add the release.
This differentiation comes out of the layers formed by the publisher is to create a very easy to change, and can continue to improve the intrinsic market. Developers compete with each other in order to attract more publishers and customers. The publisher makes more money from the user, and competes with each other to choose the original code strategy and their added value to the software.
The first feature in the intrinsic market structure is that there is no original code in the network is indispensable. Developers may be closed, even if their part of the underlying code is not directly used by other developers, competition leads to more attention will tend to generate a functional alternative product as soon as possible. The publisher may be bankrupt without destroying or modifying an open original code. As a whole, the commercial system of the entire open original code is compared with the publishers of any independent enclosed operating system, and there is a faster market demand, and in inhibiting huge fluctuations and self-innovation. Have more stronger ability.
Another important feature of open source is to reduce costs by dividing costs. Developers are reluctant to experience the routine-like stress in the traditional enclosed original code project, but like this: there is no form that is not essential from the market; the decentralized payment form; did not ask them not suitable, Outdated language or development environment for mandatory commands; no characteristics of highlighting products and protecting intellectual property rights require a new, incompatible manner to redesign "tire" command; and most importantly no project is completed The deadline constraint. In this way, the company will not launch a 1.0 version in the product, just as Demarco and Lister are commented in the discussion of "I have finished calling me" management model (see " The development team is like product "one article), which is usually not only good for quality, but also actually contributes to a real research results to spread at the fastest speed.
On the other hand, publishers can engage in things they can complete. In this way, they can focus on systematic integration, packaging, quality assurance, and services, without having to consider the problem you need, and to develop their competitiveness. By being an indispensable part of an open original code business model, it is more honest with the continuous information feedback and supervision of users, whether the publisher or developers will be more honest.
12. Successful replication
The common tragedy may not be in the development of the open original code business model, but this does not mean that there is no reason to suspect whether the current situation in the open original code community can continue. Will major participants betray a common cooperation with the further increase of risks?
This issue can be proposed from several different levels. The story of our "successful" is based on such an argument, that is, the individual is difficult to quantify the contribution value of the open original code. But this argument is not much more influence on companies that have part of Linux's income that have been connected to open original code. Moreover, their contribution value has been quantified. But is this cooperative role now?
Research on this issue will lead to our interesting thinking about some problems, such as the economic situation of open original code software in the real world, and what is the model in the real software service industry in the future software industry.
From a practical perspective, this problem for an existing open original code community can usually be proposed in two different ways. Is Linux split? The other is the first opposite, Linux will develop into a dominant product, similar to monopoly products?
When it is hints that Linux will split, we cannot think of the history of UNIX version split in the 1980s, and many people researched whether history was repeated. Despite the discussion of open standards, although there are many alliances, collaboration and contracts, UNIX ownership is divided. It turns out that the seller has enabled or changing operating system equipment to make their products and the unsatisfactory wishes, constantly reducing the entry barriers to independent software developers, and reduces the total fixed business relationship with customers. Cost, to increase the interest of UNIX's entire market share to be more intensive.
However, the above situation is unlikely to happen to Linux, which is based on a very simple reason, that is, all developers of Linux are restricted based on open protocols such as foundations for development and other operations. And in fact, it is unlikely that any of them is unlikely to maintain their products, because the original code of Linux can be efficiently developed to share the original code with all the publishers. . Any publisher uses a new feature, and some of their competitors can clone it for free.
Because all publishers know this, no one has thought of a plan to implement a conspiracy, one and a policy that causes UNIX standard splitting. Instead, Linux's publishers are forced to compete with a way that is actually beneficial to customers and the entire market. That is, they must compete in service, technical support, and actually make the installation and use of all convenient design.
The common open original code also removes the possibility of monopoly. When people in the Linux community are worried about this problem, they often complain that "Red Hat" name, and "Red Hat" is the largest of Linux is also the most successful publisher, which has 50 US markets. 90 % Share. But there is also a thing that is worthy of persistent, that is, in May 6.0 of the long-awaited Red Hat in May 1999, download CD image from Red Hat's FTP site. A library publisher and many other CD software issuers have begun to sell more than Red Hat, and in fact, in this time, the Red Hat's CD-ROMS has not yet real batch shipping sales. . However, Red Hat did not invade this thing, because they know that they did not know any of their binary data in their products. Because social guidelines in the Linux community are not allowed to do so. In the later days, John Gilmore has appeared John Gilmore, that is, the people on the interconnect will interpret the interconnection network inspection system as a procedure for it and some routine. Based on this, the hackers responsible for Linux have cleverly explain attempting to control the original code as being the destruction of them and some routine official procedures. For Red Hat, if they oppose their new products, they will seriously enhance their ability to attract developers in the future.
Perhaps, in the form of a combination of Linux community guidelines, the software license system of Linux community guidelines is actively proactively, and the Red Hat is stopped from the monopoly based on open source products. The only brand, service, and technical support relationships with users who voluntarily pay them. This will not make an overwhelming monopoly situation have too much likelihood.
13. Open R & D and re-development
Another reason why investors investment in open source world is to change him. Developers have gradually felt that they can get compensation from their things that they want to do, rather than using their own formal work, and maintain their hobbies for open source sports. Companies such as Red Hat, O'Reilly Associates and VA Linux System are exploring how much investment needs to build semi-independent R & D institutions through hiring and maintaining stable and dry open source programmers.
This approach is only economically viable when the income brought by the company's income brought about by rapid expansion of the market is economically feasible. O'Reilly is capable of burdening Perl and Apache's main authors to accomplish their work because efforts can be sold and perl and apache; VA Linux system can make the laboratory have enough funds for the source of funds. Linux's prosperity, they can sell more workstations and servers; Red Hat can bear his senior R & D laboratory is also due to the value of the company's Linux products and attract more users.
In the influence of intellectual property rights such as patents, trade secrets as the culture of the company's hand, this idea (open source) is unable to explain for the strategist of traditional software industries (although the free software market is in fact constantly Expanded). Why do you spend money to do the results that you can make your competitors enjoy free?
It seems that there can be two reasonable explanations. One is the return of huge market share from open research as these companies continue to stay in their market. Through open research and development, I changed the profits of "Tomorrow". This seems to have some days of night, but what is interesting is that it is not true, why do companies do not hesitate tolerate freedom?
In this capital home, it is desperately staring at the world of investment risk assessment. Although venture capital analysis is necessary, this is not a good explanation of the star effect, because it actually investigally investigating himself is not very clear. If you are asked, they will tell you what they do, they think that they are considered to be right. The president of the three companies mentioned earlier, so it can be said that the conclusion that I said is definitely not a lie. In fact, I still have to do a section in Va Linux Systems at the end of 1998, so I can put some "correct" suggestions, I found that the company has basically no objection to what I did. Economists will ask, so how to calculate these work? If we have accepted the "correct thing" said that it is not empty, we will think that "Do the right thing" will bring any benefits? The answer to this question is neither surprising and not difficult. In fact, in other industries, the big public is selfless, actually in order to win a good name to the company.
Strive for famous gas and think this is an intangible asset that can be rewarded in the future market. This is not a new thing. These companies' behaviors show that they are building reputation, this is a very high value. They clearly hope that they can do their high-priced people to do projects, not to protect directly from it, even if they are very lacking in stock preparation, the capital is very lacking. And at least until now, this approach has begun to get a return from the market.
The heads of these companies are very clear and credible for the company. Volunteers in the customer group not only help them do research and development, but also informal market partners, these are their backs. The relationship between the company and the user is very intimate, which is usually based on the private relationship between the company or external mutual trust.
These phenomena enhance the understanding of the inferior me before. The relationship between these companies and their customers and developers are completely different from the relationship between the RED HAT, VA and O'REILLY and their customers and developers. This is a very interesting special model and is a knowledge-intensive service industry. In addition to the technical industry, we can also find shadow of this model from the legal, clinical medicine and colleges.
In fact, we can see that the open source company hires excellent hackers and universities to hire a well-known professor. In the implementation method, both have some investment methods of the aristocrats in front of the industrial revolution, and the similarities of some aspects are obvious.
14. Thus and each
Fund support (of course, it is also necessary to profit from the profit) The market mechanism for the source code open development is still in the rapid development. The business models described herein are not final conclusions. Investors are still constantly summing up experience from the results of software industry reform. This new model is service-oriented rather than emphasizing protection of intellectual property rights. They will be in an appropriate
The software industry will bring benefits to the original people only by investment in 5% market value; the service industry is not as good as manufacturing, but the doctor or lawyer will tell you, the service industry The returns obtained by entrepreneurs). However, when software users can get many benefits from free software products and save money, they can get more profits from investment. A similar example is a huge impact of the traditional voice telephone network to the current interconnection network.
Commitments to saving expenses and better use are creating a huge market opportunity, and many companies and venture investors have begun to explore this market. When the first draft of this article is completed, a very famous venture capital institution in Silicon Valley began to note. They invested a service company with 24 * 7 Linux technology support, which is generally expected to be before 1999. There are several Linux manufacturers and some stocks related to free software, and their financing should be very successful.
Another interesting direction is systematic to create an outsourcing market on a free software development. Sourcexchange company
15. Conclusion: After free software change
After the transition to the free software, what will the entire software industry look like?
In order to answer this question, it is necessary to classify the software to categorize the software according to the services provided by the software, and the service reflects the opening of the software. This division is closely related to the marketing level of software service services. The essence of this proposed method is just similar to the three nouns of our daily: application (basic without commercial services, no or lack of open technical standards), component (service commercial, standardized), middleware (There is a need for some commercial services, there is technical standard but not perfect). Current (1999) Typical examples of the above three software are word processing software (applications), TCP / IP protocol package (components), and database engines (middleware).
The previous analysis of the allocation method showed us that three software forms of components, applications, and middleware will transition to the free software system in different ways, and their respective freedoms in combination with closed software. It also needs to be pointed out that in a field of free software in a field of software, it is also very strong whether the network influence there is strong. The negative impact of software enterprises has a negative impact and software products are still a business. The impact of sensitive capital resources and other factors.
If it is not limited to a particular area, we can boldly make boldly made the following predictions from the overall perspective of the software industry:
Component products such as Internet, Internet, Operating System, and other underlying communication software needed to cross each other in competition, will gradually be open, these software will be based on software issuers or other service institutions such as Red Hat or other service institutions will be with users. Groups are jointly maintained.
On the other hand, the application type software will continue to be closed. This software is usually very advanced for their unapproved algorithms that use very high or use, so that users are still willing to spend money to buy these enclosed sources software, and this means that this software reliability requires very bottom, and It may cause the risk of industry monopoly to be tolerated. This phenomenon is most likely in the field of vertical market in a relatively small network impact. One of the Lumber-Mill we mentioned earlier is this product, the most beautiful software product in 1999 - biomolecule structure identification software is also here.
Middleware, like database tools, development tools, or other high-end application protocol packages for specific areas will be a free and closed convergence. These middleware software products will gradually be closed or open may depend on the software's bankruptcy risk, the higher the cost, the more costs required to open the market, will be more open.
Anyway, in order to depict a complete blueprint, we should still see whether the app is still an middleware, this is a static division. In the "when" will open "section, we have analyzed any software products will have passed a life cycle from reason to enlarge the ingredients from reason, which is the same for the entire software industry. With the popularity and standardization of key technologies, as the proportion of commercialization is increasing in software industry, the application will gradually convert into middleware, such as after separating the database front interface and database engine, database The interface becomes a middleware. When the services required for the middleware products are getting more and more, they will gradually turn into an open source, and the change we see today is this example.
We can expect that in the future, with the powerful competitiveness of free software, the final fate of a certain software will not be demanding that is part of the open component system. Although this is indeed a bad news for software companies that intend to earn profits from closed software, the software industry is still an industry, then new high-level applications will continue to open, privatized intelligence. Resource monopoly a software will only have a limited life cycle, and will eventually be converted to free software.
Finally, we have to see that this change from closed to openness is still mainly used by users from software products to continue to develop. More and more high quality software will be created and used for long-term use, rather than being hidden by some people in the secret room. This miracle is not appropriate enough with Ceridwen's magic pot, because foods that have changed from the magic pot will gradually rot, and the software in the free software world will be an endless treasure. In the free software you have the most free freedom, whether you are intending to provide business services or intend to contribute to him, the free software world will provide an accumulated and exhausted precious wealth to everyone.
16. References and Acknowledgments
[Catb] Cathedral and market
[HTN] open up intelligent domain
[DL] de Marco and Lister, Peopleware Complete Productive Projects and Teams (New York; Dorset House, 1987; ISBN 0-932633-05-6)
[SH] Shawn Hargreaves written on how to combine open source and game makeup Playing the open source game
In the process of completing this article, through several fierce discussions with David D. Friendman helped me further refine how to strengthen the "turning civilians" of open source groups cooperation. Thank Marshall Van Alstyne to point out the exact meaning of "hot information products", I owe him a human condition. The Ray Ontko organized by Indiana gave me a lot of beneficial criticism. There are many enthusiastic listeners who have issued a speech in June this year, and if you are a member of the audience, you will understand who I refer to.
After I published this article, I also received many materials about free software development models, which continued to enrich the content of this article. Lloyd Wood pointed out the importance of "future benefits" free software development model; Doug Dante reminds me of "free" in the future "this business model; Lionel Oliviera Gresse helps me give a better name to a business operation mode; Stephen TurnBull gave me a good thing for ignoring the free knight.
17. Appendix: Why is the hardware manufacturer of the driver source code Waste investors' money
Peripheral developers, icon, net card, hard drive, or graphics card manufacturer, their traditional practice is to enclose the source code of the driver. But this phenomenon has now changed, such as Adaptec and Cyclades have become accustomed to disclosing their various board driver source code and corresponding documentation. However, if you want to open the source code to become a universal practice, there is still a lot of difficulties. In this appendix we are intending to clarify some of the wrong ideas that still maintain a closed source code system in the business sector. Assume that you are a hardware manufacturer, you may worry that the opening of the driver code will leak your hardware how to work many important secrets, so that your competitors can make you unfair by analyzing your source code. Competitive environment. This idea will only stand in the age of three and five years. However, it will have to spend a major part of the entire product update cycle even if the source code is open, your competitors will have to spend a major part of the entire product update cycle. To ponder what you have published, because the product update is now shortened, your competitors will not have enough time to think about and innovate their own products. So they go to study your open source, I have actually drilled into your circle.
Anyway, the secret in today's code will not be hidden for a long time. Hardware drivers are not as complicated by operating systems or applications. They are generally small, it is easy to be compiled and imitated, which can be customary, and actually is often often used. These people get it. It can be said without exaggeration, there are thousands of passionate programs for Linux or FreeBSD work in the world, they are willing to write drivers for any new board. Since many kinds of hardware devices have relatively simple and standardized interface specifications, such as common disk controllers or network cards, enthusiastic hackers can be quickly available even without documentation, no need to refine existing drivers. Write the correct driver, and often get fast than the original manufacturer.
Even if you encounter complex devices such as graphics cards, it is difficult to argore with anti-compilation tools. This job does not require a lot of effort, it is difficult to say whether it is illegal, and under the joint efforts of the global programmers, it is already possible to do any legally reverse engineering in legal laws. From the Metalab website, check the hardware type list that Linux core and device driver library can support, you will immediately understand the non-virtual, Metalab's URL is:
Conservatives in your driver still have a temptation from a short-term effect, but from the perspective of long-term strategy, it is not advisable, especially when your competitors have opened the source code. If you have to stubbornly enclose your source code, you can only burn those code into the ROM on the board, and only publicly access the interface. So quickly open your source code, quickly expand the market, you have to believe that you have the ability to think through your own constant thinking and innovation to attract more potential user groups that belong to your competitors.
Persistence to the closed route is a dead end, your secret will inevitably expose, you will not be able to get the help of the free programmer, and there is no stupid competitor to spend time to imitate your design. More importantly, if you have the idea of opening up to open, you can get a broader development space, but you have missed it. Because your equipment is too conservative, missing information and solid self-sealing, and you can't recognize your own mistakes, therefore, most of the network administrators and more than 17% of the network administrators and more than 17% of the business data centers will take your hardware equipment from The deletion in their purchase list, and turn your gaze to other open hardware vendors.
18. This document revised record
What you see now is the 1.14 version of this document.
In the following list, some minor revisions and printing are no longer listed.
May 20, 1999, 1.1 version - draft
June 18, 1999, 1.2 - The first version of the first version of the private exchange, June 24, 1999, 1.5 version - the first version of the foreign publication
June 24, 1999, 1.6 - made some small changes, gave the definition of 'Hacker'.
June 24, 1999, 1.7 - clarified some standards
On June 24, 1999, version 1.9 - Added discussions about "future benefits", "free free" development model and chapter on closed costs
On June 24, 1999, 1.10 version - took a better title to the "blade" mode
On June 25, 1999, 1.13 version-corrected the 13% income of Netscape, added analysis of the free knight, and corrected the list of closed network protocols.
June 25, 1999, 1.14 - Added an example of E-Smith
July 9, 1999, 1.15 - Updated the content about the hardware drive appendix, and gave "hot goods" with a better explanation with the help of Rich Morin.
