[ I send this report on behalf of Hong Feng (hongfeng@gnu.org) who is building FSF China and with his permission. I see as very important that Free Software Foundation friends that are expanding the Free Software Foundation in Germany, Sweden, Italy, Spain, Portugal, United Kingdom, France and Japan are aware of the work that is being done in China.
FSF China does not yet have a web site but you'll find a lot of content related to current Free Software activities in http://www.rons.net.cn/. ]
FSF-CHINA Activity Report
THE FSF-CHINA, or in full name, "Free Software Foundation, China Academy" was planned to establish on May 28, 2000 in Wuhan, China, by Dr. Richard Stallman, president of FSF, Inc. as well as the founder of the GNU Project, and Hong Feng, the president and CEO of RON's Datacom Co., Ltd.
Since then, we have made a lot of feasibility study, mainly on the legality in China to set up FSF-CHINA. By the beginning of March 2001, our lawyer informed us it is legal to adopt the FSF by-law terms as FSF-CHINA's. So we have quickly draft the by-laws of FSF-CHINA in Chinese, and started the registration process.
On March 05, 2001, I launched the MNM Project, stands for "MNM's Not Millions", it includes three sub-projects: to establish the FSF, China Academy; to publish hundreds of free books; and organize hundreds of thousands programmers and engineers to support free software movement in China. The mission of MNM Project is to support the free software community in China.
Many people asked me a question about the relation among FSF-CHINA, RONSNET, and Ron's Datacom Co., Ltd. Here I would give a clear explanation. FSF-CHINA is a non-profitable organization, which systematically studies the free software philosophy, technologies, also it leads a group of free software R&D projects (see section below). RONSNET is a registered logo of RON's Datacom Co., Ltd. in China for the web site "www.rons.net.cn", with the only one mission to support the free software community in China, and RONSNET is supported by Ron's Datacom Co., Ltd. by money and volunteer human labors. FSF-CHINA is completely independent with RON's Datacom Co., Ltd., though RON's Datacom Co, Ltd. has donated money and other resources to FSF-CHINA. Eventually, FSF-CHINA will run another web site with domain "fsf.ac.cn" at the time of necessary.
FSF-CHINA is looking for members who are qualified to enter into the board of directory. At the present, there are three people, Hong Feng, Yan Feng (who is an assistant professor in a law department in a university), and Zheng YongGang from Shanghai, whom RMS accepted. RMS is the honorable president of FSF-CHINA.
As there are so many misunderstandings in China about free software and its philosophy, so FSF-CHINA has done a lot of works to broadcast the truth about it, including why free software fight against proprietary software mechanism, what is the difference between free software and open source software. We have printed RMS's article "GNU Project and Free Software Movement" in thousands paper copies (in English-Chinese bilingual form), and distributed it for free. Our effort now has got some returns -- now most mass media in China have realized these difference, and the journalists become careful when they use the related terms.
We also arranged a lot of speeches about free software. Besides RMS's a chain of speeches given in May-June, 2000, we also invited Mr. Robert Chassell visited China last August, and he gave speeches in Beijing and Xi'an.
Other Activities ----------------
1. Chain schools.
We have started the work to set up a chain school in China, which has more than 10 sites under preparation. The training courses include how to use free software tools, like GNU Emacs, GCC, GDB, CVS, etc. I have put this as the first step till our "Hackerdom Training" course series, and eventually we will hold a sort of exam, the students who passed the exam could acquire a "Hackerdom" certificate or a diploma. I believe this is a right way to organize the "hundreds of thousands of programmers" in China to develop more free software, which is a part of MNM Project goals.
We will welcome other training organizations to join the chain schools, as long as they agree to teach the free contents from us. However, if they want to hold the exam of "Hackerdom" certificate in the future, then they need to sign a franchise agreement with us, to get the permission to place RONSNET logo, and get the technical support from RONSNET, including teacher's training, textbooks with discount, web-based courseware, answers for students from professors, etc. (a little bit like McDonald's fast food chain restaurants), 10% of the tuition per student will be paid as a collection to support FSF-CHINA.
(FSF-CHINA accepts other incomes or donation, and use the money to develop more free software or free documentations.)
2. Hospital for Free Software Companies.
June 21, 2001, We have set up a hospital on RONSNET for free software companies. To run a business is not simple, to run a free software company in China at present may be more complicated, not every programmer has the talent to run a company to be successful, so they need help. If a company is a free software company, and want to get help, they could ask help from RONSNET Hospital; if a programmer has a good idea, and s/he is going to set up a firm to do free software business, we welcome him/her to make an exam in our hospital before you doing that.
RONSNET hospital is helping Center for Free Software companies or individuals in China, the service includes but not subject to the following items:
* Understanding Free Software * Business Planning * Organizing a Company * Leadership & Vision * Directory Board * Sales & Marketing * Growing & Managing a Business * Going Global Operations * Tracking & Controlling Costs * Forecasting Budgets * Business Financing * Analyzing Financial Statements * Managing Investment
We will invite some experts for technology and management as doctors in this hospital. 10% of the diagnose and exam income will be transfered to FSF-CHINA to support its R&D.
3. Publishing --------------
We have started the translation of GNU manuals from English into Chinese, till the present, Programming in Emacs Lisp: An Introduction was published, other main manuals are still under translating or proofreading, or final editing. We are going to publish the Chinese manual on RONSNET, and if a local publisher wants to print them in paper copies, we welcome them do that, and hope they could donate us some money, the donation will be fully transfered to FSF-CHINA.
Also we are preparing the Free Software Magazine. On RONSNET, we have opened a link dedicated to this online magazine, when the articles accumulated up, we are going to publish it in paper copies.
I have discussed with Linux Journal, for the possibility to exchange the articles, and I got a positive result. Also I am talking with the former president of SuSE Inc in States, who has a lot of experience of the journal /magazine circulation services, he could help us for distribute the FSM world widely. At the present, we have to invest money to FSM, as we have no ads incomes yet. But I hope the time of ads incomes appears the sooner, the better.
Also I have talked to a lot of authors, so that they could FDL their works, and we could collect them into our MNM free book catalog. I think I have made a limited progress, some authors agreed to do so, and some books planned to publish as proprietary book were turned into free book after my wording (for example, the Zope Book from www.zope.org is a free book now.), RMS helped a lot on explanation to the authors about the terms of GNU FDL on this work.
Also we are looking for authors to create new free books. This will take more time, but we will keep on doing it. When China joins WTO by the end of this year, it might be possible for us to register a publishing house, by then we could print the books and sell them directly, all the incomes will be transfered to FSF-CHINA.
FSF-CHINA Projects Overview ---------------------------
1. Free Chinese Fonts
As there is no free Chinese font to use, so I started a project to design a set of free fonts, both in Postscript and bitmap. Till the beginning of June, we have finished 30% of the 27,484 Chinese characters, and we expected it could put into use by the end of this year. Adobe expressed they could help us to pack them in CIDs.
RON's Datacom Co., Ltd. financed the development. And all the fonts are created by free software. The fonts will be free to use by any GPLed software, due to many hardware manufacturers and open source vendors are hostile for free software community, so the fonts are usable by them, but it requires additional agreement. Proprietary software must pay us an amount to use the free font in the software, as we dont want to see our effort are abused by the proprietary software publishers. As our Postscript fonts are coded, so it could apply for copyright.
2. POD with free software
When the fonts are ready, we will start the successor project of POD service network, means the print-on-demand services, we will set up a chain of POD centers national wide, or world wide, help the centers to install our free fonts on a devices connect to Internet, so that anyone over the Internet could send his/her files in PS, PDF, DVI or SVG format to the POD center, and ask the POD center print it our, bind it, pack it, and delivery it to the user. The quantity of the copies could be any number greater and equal to 1.
POD is not a new concept, but all the POD equipments now are proprietary, we need to change this situation, so that everyone could have the freedom to share the information with the benefits of paper copies.
To reach this freedom, there are a lot of technical problems to work out, most of them are coming from hardware side, not from software side. We will seek the hardware manufacturers to cooperate. In this July, I will visit Russia to talk with a laser R&D institute, to solve the drill problem (which required after piling and before binding the pages). laser beams could cut metals, so I think it works with paper, but need to control the time interval of the laser pulse, so as to avoid the paper burned. Once the laser beams could be controlled by free software, then we will move a huge step ahead.
3. MNM Office
As SGML/XML documents more and more appear, we started a project of new office suite, that is not the same office suite like GNOME office, but a new one. RMS hopes to develop it with GTK+, and I agreed. It has a couple of componenets, it will address to the enterprises customers or powerful individual users.
4. Meta-kernel and Scheme Machine
There was an idea of Lisp machine in 1970s-1980s, i.e. the programming language works as the operating system. After some researches, we think this idea could be coming back on stack based chip.
Three months ago, I finished the proofreading on the book (which you have received), I started my hacking on Scheme, as RMS told me Lisp is the most powerful programming language, and Scheme is a modern dialect of Scheme. (Scheme has a dozen of implementation since 1975, GNU Project has a Scheme implementation called Guile)
As Pansystems Workshop has years of research since 1976, so there are many results on system science. One of them is about the new approach on how to design non-Von Neumann computer. Von Neumann architecture came from the early thought of Turing, roughly speaking, that is a way of "Tower of Babel", means, computer approaches infinity very quickly, and it need more and more memory when more applications are under computing. Also it needs an OS as a "business manager" to administrate the resources on the computer.
When it comes to some embedded system, I think this approach does not work very well. These embedded system do not have the need to run an OS on it.
After our research from Pansystems theory, we drew a conclusion, that we could just put the method to operate the relation into the embedded system, and try to implement the method in a set of basic software procedures. These software procedures like the basic bricks, which could be revoked with each other, in recursive and dynamic hierarchical way.
The prototype came from a Pansystems expression, I could write it in TeX here: $ B \subset A ^n \times W $, where B is software system, or the set of software procedures, A is the hardware set, W is short for weight, which is a set of relation built upon the hardware set's direct product. In theory, we could also construct the $W ^m$, which is a direct product of weight. But in practical, I think just W is enough, and easy to implement on hardware.
The best chip architecture to implement the idea is neither register based CISC, nor the RISC, but stack-based chips. As the stack is a very ideal data structure to keep the software procedures. Also it is easy to implement by the VLSI hardware technology today. As you know, Scheme or Lisp could be the best programming language to handle the recursive list, which could be represented by stack in hardware level.
I have worked for a time looking for the stack-based chip, and the key requirement is about the low power consumption (some embedded systems like PDA has very high requirement on the power consumption).
When the chip is eventually selected, then our team will start to port a Scheme implementation, says, guile onto the chip from scratch, and finally get a Scheme machine. It is like the old idea of Lisp Machine, which the programming language is stand for an OS. I would call our Scheme porting as a "meta-kernel". Our Pansystems research indicated, there are 33 operators could be implemented to handle various relation transformation. So based on these 33 operators, we could then build the software procedures, which could be handled by the stack-based chip.
When the meta-kernel is ready, then by designing and organizing the software procedures (that's the work about the W in the above formular), applications like word processing, spreadsheeting, XML parsing, networking etc. could all be processed by the meta-kernel. If the algorithms are well designed, then the stack-based meta-kernel could be expected to reach high computing speed, which could meet the requirement of some real-time applications on embedded systems.
Actually , the stack-based computing is not a new conception, PostScript is forth-like stack based programming language, and PostScript based printers are working world widely already. But the difference of meta-kernel with the PostScript technology is: PostScript is a software implementation, not at hardware level, also Postscript application is just one program processed on the computer, while on the meta-kenerl, all applications will be transformed and processed by the meta-kernel.
This project is open, anyone who are interested in it could join. And we are looking for company to finance it.
Administration of FSF-CHINA ---------------------------
When the board of director has 3-5 people, we will hold the first meeting of directory board. It is anticipated to hold in the second half of this year.
1. Donations
RON's Datacom Co., Ltd. has donated RMB500,000 cash for FSF-CHINA (ca USD60,240) since last May. As FSF-CHINA is not tax deductible yet, so we need to work hard on this, so that more companies like RON's Datacom could make the donations. The money was used for covering misc costs to promote FSF-CHINA and support its projects. I have asked an independent accountant to record all the expenses.
In August-September, when the registration process finished, we will open an account officially to accept the donations and equipment. All companies and individuals which donated to FSF-CHINA, will be published on RONSNET, as long as they agree to list them on.
2. Hardware
RONSNET donated a server (Zenit): K7, 128MB RAM, 9GB HDD which could be served www.fsf.ac.cn soon. ----------------------------------------------------------------------