A long time ago in the now forgotten 1980s when Japanese electronics manufacturers could clearly see the benefits of creating an open, real-time, total computer architecture for computerizing human society in the 21st century, there were actually plans to create new hardware devices based on TRON developed specifications, in particular microprocessors based on the world's first open microprocessor architecture. In addition to the TRON VLSI CPU subproject, there was the µBTRON LAN bus, a TRON-designed ergonomic keyboard, and even a TRON Intelligent House with TRON-specification fixtures based on a unified Human-Machine Interface (HMI). In other words, the TRON Project from its inception has called for the parallel development of hardware along with the operating system software that it has become famous for. Sadly, Japanese electronics manufacturers were unable to maintain their faith in the TRON concept, and the hardware development plans slowly fell by the wayside one by one.
Well, the parallel development of hardware and software is now back in style in the TRON Project. Having played around with just about every operating system developed in the U.S. and discovering there is little profit in it for them--this lack of profitability, incidentally, extends even to Microsoft Corporation's "strategic partner" in the Japanese PC market, NEC Corporation--and that nothing in the U.S. is as suitable for embedded use as ITRON and BTRON, certain Japanese electronics manufacturers have finally come to the realization that further developing the highly scalable and royalty-free TRON Architecture is the best way to develop a profitable business model in the "Post PC Age." However, in order to further develop ITRON and BTRON for the vast array of embedded devices that will underpin the computerized society of the 21st century, one thing is necessary that has been lacking up to now--a standardized development environment that includes a reference board for prototyping these new devices.
That problem has now been solved through the appearance of T-Engine, which was officially announced to the press on December 12, and later again at TRON SHOW 2002. As he began his keynote lecture at TRON SHOW 2002, Prof. Ken Sakamura proudly held up a small yellow box (the outlines of it can been seen in the image above) in the palm of his hand and introduced this new open development environment to a standing room only audience. Although he proudly pointed out the achievements of the ITRON, particularly its astounding success in the cell-phone application field, he admitted that ITRON's development environment has been bad and that there is a shortage of programmers who can program for it. Prof. Sakamura being Prof. Sakamura, he threw out another of his famous "exaggerations," saying that he wants to create software that lasts 100 years, since the life cycle of software is too short. In non-exaggerated terms, that means he wants software to last considerably longer than it currently does.
In regard to software, T-Engine is designed to fill another gap in TRON development--the smooth distribution of middleware, which has particularly been a problem for the BTRON subarchitecture that is currently being developed on the less-than-fully-open IBM-PC/AT-compatible platform [1]. As TRON-based devices increase their functionality in response to demand, various types of middleware have to be distributed from one platform to another. A standardized development platform is essential for that purpose. In fact, to emphasize how much middleware is immediately available for T-Engine, the wall of the hallway leading to the exhibits area had a list of approximately 150 pieces of it ready to go, including parts of the BTRON3-specification operating system. And just to show that this isn't all talk and no implementation, the exhibits area was filled with T-Engine prototypes based on ITRON, JTRON, and BTRON plus various types of middleware. These prototype implementations based on T-Engine included:
|
Now, to the non-specialist TRON Web visitor reading this, the appearance of T-Engine may seem like no big deal. It certainly doesn't sound revolutionary. However, in fact, it is revolutionary--it's as revolutionary as the appearance of Macintosh in 1984, but with a couple of twists. The Macintosh was able to able to offer features such as plug-and-play a decade before it appeared on the IBM-PC/AT-compatible architecture because it was based on a "unified hardware and software model." The "unified hardware and software model" embodied in T-Engine will likewise enable the development of new functions in the future ahead of other architectures; and since it is open and royalty free, all hardware and software makers can develop on top of it without the payment of profit crushing royalties. For example, the BTRON3-specification Cho Kanji 4 operating system has already been successfully ported to T-Engine. No foreign PC manufacturer has been able to put that much functionality into such a small box!
However, as with everything TRON, T-Engine also comes with subsets, since the architecture is aimed at the computerization of human society in the 21st century--really! Prof. Sakamura even mentioned in his keynote lecture that the TRON Project has been misunderstood as challenging Microsoft--no doubt to the delight of those external forces that would like to interfere with and ultimately control the development of computer technology in Japan--but in fact the TRON Architecture is based on different underlying models. From the beginning, the "TRON total architecture" has been divided into a group of "subarchitectures," which in turn are divided into hierarchical "subsets." T-Engine is no exception to this model. Thus there are two more subsets of the T-Engine. One is µT-Engine [pronounced "micro T-Engine"], and the other is pT-Engine [pronounced "pico T-Engine"], which are intended for smaller and still smaller devices linked to TRON-based networks. Here's an outline of the subsets:
Subset | Applications | Specification Outline [2] |
T-Engine | Network devices with a highly developed user I/F, eg., pocket PCs, PDAs, cell-phones, etc. (these are referred to as "communication machines" and they are used to control "intelligent objects" based on the two subsets below) |
CPU - 32-bit with Memory Management Unit (MMU) RAM - 16 MB/32 MB (increase possible with expansion board) Flash memory - 4 MB (increase possible with expansion board) Serial I/O - 384 Kbps or more possible PCMCIA - Type II 1 slot USB host - Type A connector, 1 channel eTRON card I/F - (SIM card connector) LCD display panel I/F Touch panel I/F Sound CODEC - Input 1 channel, output 2 channels Extension bus I/F Calendar clock External measurements - 75 mm x 120 mm |
µT-Engine | Network devices with a simpler user I/F, such as advanced home appliances, e.g., air-conditioners, microwave ovens, video recorders, video intercoms., etc. |
CPU - 32-bit, MMU not required RAM - 2 MB/4 MB (increase possible with expansion board) Flash memory - 4 MB (increase possible with expansion board) Serial I/O - 384 Kbps or more possible CF card slot - Type II, 1 slot SD card slot - 1 slot eTRON card I/F - (SIM card connector) Expansion bus I/F Calendar clock External measurements - 60 mm x 85 mm |
pT-Engine | Smallest devices that constitute "intelligent objects," e.g., lighting equipment, switches, sensors, locks, bulbs, etc. | Still in development (the specification will most likely be based on 8- and 16-bit microcontrollers) |
There are a couple things to notice in the above table. First, the T-Engine specification does not call for the use of a particular manufacturer's microprocessor. As long as it's a 32-bit CPU and has an MMU, any manufacturer's chip is acceptable, although obviously low power designs--a Japanese specialty--will predominate on this platform. So T-Engine is a platform that is going to encourage competition among hardware manufacturers. Second, both T-Engine and µT-Engine have security specified from the start in the form of eTRON, and it goes without saying that pT-Engine will also be equipped with this same security standard. Accordingly, all nodes in TRON networks will be covered by a security system to prevent mischief by malicious programmers. It should also be noted that the µT-Engine subset is not all talk, either. It has already been implemented with a miniature camera and a TCP/IP stack, which enabled it to be used throughout TRON show site to take pictures and send them through the Internet.
Prof. Sakamura made many other interesting comments in his keynote lecture. For example, now that T-Engine and eTRON have been created, he believes that the TRON Architecture is almost complete. If that's the case, then the TRON Project will move in earnest from the basic technologies development phase into the applications development phase, which means it's time to start computerizing human society--at least in Japan, anyway. He also said he believes that Japan, which is the most industrialized country is Asia has a duty to contribute to the development of Asia, and that he has written a book describing the Japanese model for the development of an information society (Joohoo bunmei nihon-no moderu [The Japanese model of information civilization]). In addition, he pointed out that the TRON Association plans to ask Japanese developers of personal word processors, which were a hit product in the 1980s, to release their data formats so that users can convert their data into current formats.
Special Presentation: Protecting Digital Content with a Unique ID
The U.S. government did not allow the Internet to be opened to the public so that ordinary people could have access to a low-cost, global computer network with lots of freebie content. Rather, it went along with the idea of opening the Internet to the public so that companies--in particular, American companies--could make lots of money by selling goods and services through it. Ironically, massive over investment on Internet infrastructure and related startups in the U.S. led to huge losses and a recession there, while in Japan it led to the greatest economic success story of the 1990s--the development and commercialization of NTT DoCoMo Inc.'s highly profitable i-mode wireless network. Moreover, since i-mode was the first Internet experience of a large number of Japanese, many people in Japan are used to the idea of paying for goods and services on line. With that in mind, it should come as no surprise that NTT has developed and is ready to deploy a new scheme to protect copyrighted digital content over broadband networks.
This scheme was introduced in a special presentation at TRON SHOW 2002 by the Content ID Forum. The easiest way to understand what the Content ID Forum is doing is to think of the paper money we use in daily life. Each currency note has a unique number, a watermark, and other security features to prove its authenticity and allow specialists, i.e., bank personnel, to detect counterfeit notes. That's exactly what the Content ID Forum is doing with copyrighted digital content, particularly data in the JPEG and MPEG formats. Each item of copyrighted digital content registered with the forum is given a unique identification code that includes an ID center management number and content, copyright, contract, and distribution attributes. This unique identification code, which cannot be erased and withstands compression and decompression, is read using a ContentID Service Gateway that runs on an individual's network terminal and confirms content usage via the contents service gateway. The contents service gateway, which follows global standards for content identification, accesses a content management identification center with information about basic value added services, plus expanded services.
During the presentation, a lot of emphasis was placed on new possibilities, such as easily locating the content owner and acquiring further product information, but the main reason this system is being developed seems to be to protect video- and audio-on-demand services that will be streamed through open broadband networks currently under development in Japan. As a result of harsh criticisms from abroad about the slowness of Internet connections in Japan, the Japanese government announced an eJapan program that aims at creating the highest speed broadband networks on the planet, which, needless to say, won't be very valuable without high-quality content. However, many Japanese content providers, such karaoke provider Yamaha Corp., currently do not trust the Internet due to security concerns, and thus stream their pay per view content via private networks. Although there was no discussion of the question during the presentation, it also seems likely that this system will be merged with the eTRON security architecture developed as part of the TRON Architecture. In addition, there is a possibility that it will come into use on a global basis, since the Content ID Forum has links to powerful content protection bodies overseas (click here for a list).
ITRON: Giving Linux a Helping Hand
In the afternoon on the first day of TRON SHOW 2002, various companies participating in the ITRON subproject announced their latest products, which mainly consisted of µITRON4.0-specification kernels and related development environments and middleware. From the point of view of foreigners, however, the biggest news of the day was probably that Elmic Systems Inc. has merged ITRON with Linux to create a yet another TRON-based hybrid operating system. The purpose of this hybrid operating system, as in the case of previous ITRON-based hybrid operating systems, is to speed up Linux. In spite of the fact that there are people who believe that Linux is going to drive ITRON from the market, since like ITRON Linux is also royalty free, the fact of the matter is that Linux has a task switching time measured in tens of milliseconds in comparison to ITRON's task switching time of only one of two microseconds. Thus the two operating systems are not, and should not, become direct competitors of each other.
Elmic Systems' new hybrid operating system,"Accel-Linux," is based on an architecture in which Linux runs on top of ITRON as an ITRON task, and in which the ITRON tasks have the highest priority. Naturally, that means that all scheduling is carried out by the ITRON kernel, and that the ITRON kernel can override a Linux interrupt. The two operating systems can, however, communicate. Elmic Systems has built an interface between the two operating systems that uses streaming, which is good for small chunks of data, and shared memory. Elmic Systems believes that this architecture is the best architecture for obtaining the best results from both operating systems. As for application fields, the company said it is aiming the new operating system at developers of mobile and Internet devices, who will be able to take advantage of free source code from thousands of Linux hackers while developing on an industry-standard ITRON based. The big question, of course, is what the power and memory requirements will be.
T-Engine: Standard Development Platform for Next Generation Network Devices
On the second day of TRON SHOW 2002, the main event was a technical session on T-Engine that lasted most of the day. This technical session began with an introduction to T-Engine by Prof. Ken Sakamura, and as usual it was a standing room only event in the theater. Prof. Sakamura said that as we move into the age of ubiquitous computing, we are going to have to develop new algorithms that allow cooperative operation among networked devices, which is actually very difficult. To make matters worse, although the level of finish of the latest ITRON kernel, µITRON4.0, is very high, development efficiency has been very low due to the lack of standardized hardware specifications. That in turn has made it difficult to distribute middleware from one company's platform to another. For those reasons, the TRON Project has launched the T-Engine project, which is an open, real-time development environment aimed squarely at developing the ubiquitous computing environment of the 21st century.
One of the main goals of this platform is implementing strong network security, and for that purpose, the T-Engine specification calls for the incorporation of an eTRON security chip. In fact, if the eTRON chip is not on board T-Engine, T-Engine will not work. The basic software for T-Engine is T-Kernel, which is nothing other than a ITRON-based kernel with support for eTRON. T-Engine also boasts a standardized object format to enable the smooth distribution of middleware, since large amounts of middleware allow development time to be reduced. There is also a T-Monitor for handling debugging, which is also something that can slow development of a product. Prof. Sakamura said the first T-Engine products will hit the market in the second quarter of 2002. He was slightly dismayed that more companies are not participating in the T-Engine project, something he attributed to the existence of lots of weak hearted people in Japan. He also opined--in perfect Sakamura form--that the way to solve the IT slump was to raise prices!
Prof. Sakamura's presentation was followed by nine different presentations on various aspects of T-Engine. A spokesman from the first developer of an example of T-Engine, Mr. Osamu Akihira of Hitachi Ltd., showed a huge development board from 1996 plus a half-size development board T-Engine was designed to replace, so one aspect of the T-Engine project is definitely making things compact. Another is propagating MPU families. During the second technical session on the following day, both the spokesman from Hitachi and Mr. Akihiro Uegaki from Mitsubishi Electric Corporation, which developed the first example of µT-Engine, admitted to the audience that their main goal in developing these T-Engine implementations is selling their firms' microprocessors. They also pointed out that these development boards are so small, that an engineer could even debug a system while riding on the Shinkansen, Japan's high-speed railway, to a customer's location.
Hitachi's 75 mm x 120 mm T-Engine development environment is based on a 32-bit SH7727 microprocessor running at 96 MHz [3]. It can be equipped with 16 or 32 megabytes of SDRAM, and it has 4 megabytes of Flash memory. It also has a 320 x 240 pixel STN color LCD screen with a tablet function, the standard array of input-output interfaces mentioned above, and it is powered via an AC adaptor or 3.3-volt lithium-ion batteries. What's most remarkable, however, is the expansion board line-up for Hitachi's T-Engine. In addition to a compact Flash board, there is also a LAN board (100Base-T/10Base-T), and an display board for hooking T-Engine up to a high-resolution CRT. In addition, at the request of Prof. Sakamura who likes smart design, there is a special case to put the board into, although that makes it look like a child's handheld game machine to untrained eye. The most interesting demonstrations based on T-Engine were the audio and visual ones, i.e., musical applications and MPEG4 motion pictures. It will be very interesting to see what T-Engine implementations appear in the future.
Mitsubishi Electric's 60 mm x 85 mm µT-Engine development environment uses a 32-bit M32104 microprocessor that runs at 216 MHz, giving it a maximum speed of 243 MIPS. It is loaded with 16 megabytes of RAM and 8 megabytes of Flash memory, and it comes equipped with an eTRON connector, a compact Flash interface, a multimedia card interface, a debugging interface, and an expansion bus. Mitsubishi's µT-Engine implementation has neither an LCD screen nor a human-machine interface device attached to to it, but it is equipped with a color camera that was used to take still pictures of the show site and transmit them using Mitsubishi-developed middleware, an on-board TCP/IP stack, and a 100Base-TX/10-Base-T LAN board. What Mitsubishi was trying to demonstrate with its µT-Engine implementation was that it is now possible to control a variety of equipment types via the Internet merely by transmitting the pictures to a standard personal computer or a PDA.
The software architecture that runs on top of T-Engine is pretty straightforward. On top of the standardized T-Engine hardware platform, there is a µITRON-specification kernel called T-Kernel, which differs from previous µITRON-specification kernels mainly in that it supports an eTRON security chip. Another feature that differs from the past is that there is a standard monitor for debugging called T-Monitor. These two parts, kernel and the monitor, are the most important parts, according to Akira Matsui of Personal Media. On top of the kernel and the monitor, various types of middleware modules, of which there are over 100 at present, are loaded according to what is needed for a particular application. The final layer is the application layer to which the applications of the BTRON3-specification operation system can easily be ported. To make sure that things are kept open and royalty free, the standard development environment for the T-Engine software architecture is the GNU-based environment.
However, that does not means that software companies cannot make money in the T-Engine project by marketing development environments. The Japanese software company eSol Co. Ltd. gave a presentation on its eBinder development environment for T-Engine, which it was offering at a discount price, i.e., 990,000 yen instead of the standard 1,980,000 yen. In addition, there were presentations on T-Engine development environments by Yamaha Corporation, Hitachi ULSI Systems Co. Ltd., and Personal Media Corporation. Aplix Corporation also introduced a Java-based development environment for T-Engine. So there are lots of choices for system developers who are planning to create software applications on the basis of T-Engine. This is quite a change from the old days in the TRON Project when there were minor differences among the various implementations of ITRON-specification kernels and hence developments environments and tools.
Since there is all the above mentioned standardization of both hardware and software, there is a huge payoff for the potential T-Engine developer--fast development time, and hence greater profits. T-Engine's modular hardware and software design means that the computer architectural equivalent of an "Erector Set" is being offered to developers. That is extremely significant. It's probably as significant as the fact that this is all based on an open architecture that is royalty free, something that enhances profitability even more. The various pieces are designed to be snapped together and thus allow the developing firm to focus its attention on the portion of the system that is truly unique, the system part with the greatest added value. It's all so logical and simple, it leads one to wonder why the leading members of the TRON Association didn't try this sooner. Isn't it that in addition to necessity, lack of profitability is also the mother of invention?
|
|
Note: the above images are high-resolution images |
BTRON: Character Searches Get Even Easier, Mozilla Appears on the OS
Just prior to TRON SHOW 2002, Personal Media Corporation announced a new version of its BTRON3-specification operating system. The new version, Cho Kanji 4, drew a lot of attention because it is equipped with even more advanced character search features, and also because it is the first version of the operating system with a standard Web browser, the open source Mozilla browser (click here for a screen shot) that traces its lineage to the Netscape browser. These are interesting developments, because on the one hand Personal Media is pursuing an independent development path aimed at implementing the best Japanese language processing functions on the planet, while, on the other hand, it is porting standard software to BTRON. It should be pointed out that Mozilla for Cho Kanji 4 does not have a BTRON-like interface, and at 14 megabytes it's a huge application by BTRON standards. It does, however, offer a cookie function, compatibility with frames, JavaScript, and SSL, features lacking in the much more compact BTRON Basic Browser.
The new character search function, which Personal Media calls ikeiji yuragi kensaku [variant character swing search], is one of those things that is difficult to put a name on, but it's easy to understand if one sees it in operation. The main goal of this character search function seems to be finding variants of kanji, or Chinese characters--and there are lots of variants used in Japanese personal and place names, so this is an extremely important tool in Japan--inside the enormous Cho Kanji character set, but it can also be used with the characters of other scripts. For example, by inputting a circle and the numeral one into the character search panel, a circled Arabic numeral one, the Greek letter for 'f', etc., are output. As for another example, inputting a question mark leads to everything with a curlicue in it being output, so it sort of does with characters what a thesaurus does with words. On top of that, Personal Media has improved the search function so that the entire real object/virtual object network inside the hard disk can be searched.
These powerful character search functions were demonstrated in a presentation by Mr. Hiroyuki Oshima of Personal Media. He also demonstrated the various improvements to Cho Kanji 4, plus some applications that had previously been released. For example, he showed how easy it is to create Web pages using the Cho Kanji Web Server and Web Converter applications, which can handle text and graphic images. A BTRON user employs the standard text and graphic tools to create a document, and then merely by clicking the "Convert to HTML" item on a panel, the document is automatically converted into a Web page. No HTML to play around with, and no special editors required. He also pointed out that Cho Kanji 4 can operate on either MS Windows or the Macintosh operating system using Connectix Corp.'s Virtual PC 4.0 [4]. In addition, Cho Kanji Mail, a heretofore separately marketed e-mailer that can handle the entire Cho Kanji character set, has been made a standard system application in Cho Kanji 4.
Personal Media, along with Hitachi, had the two largest booths at TRON SHOW 2002, and both of them attracted lots of attention from visitors to the show. The Mozilla browser for Cho Kanji 4 seemed to draw the most attention--no doubt because of dissatisfaction with the BTRON Basic Browser--but there was also an interesting "motion picture" demonstration based on the MicroScript programming language, since it has been modified to handle new image data formats, i.e., the JPEG, PNG, and BMP formats. The MicroScript interpreter works so fast that a JPEG stack looks like a short MPEG movie, which is a handy feature for people who have to create presentation materials with a BTRON-specification computer. The Personal Media booth was also doing a brisk business. The company was conducting selling its Cho Kanji 4 operating system for 5,000 yen off, plus TRON publications. Prof. Sakamura's new book mentioned above was completely sold out, so someone must be interested in the TRON philosophy.
Having Fun with T-Engine
As strange as it sounds in an age when people are forced to use computers at school and at work, Prof. Ken Sakamura, the TRON Project Leader, wants people to have fun with T-Engine, and thus an hour and a half technical session was dedicated to that topic on the morning of the third and final day of TRON SHOW 2002. Some young people might not realize it, but the personal computer revolution began from groups of hobbyists whose main goal was to have fun with a new invention called "microprocessor." Apple Computer Inc. is an outgrowth of the Homebrew Computer Club in Silicon Valley, and Microsoft Corporation began while Bill Gates and Paul Allen went to New Mexico to play around with a hobby computer called the Altair. Over in Japan, a young electronics junkie named Ken Sakamura was so impressed by the microprocessor that he talked his countrymen into launching a technology project to make it the core of a new computer architecture called The Real-time Operating system Nucleus (TRON).
Prof. Sakamura began the technical session by pointing out that it was really a difficult job building the TRON Intelligent House back in the late 1980s because of the level of the technology at the time was so low. With T-Engine and ancillary technologies, however, it would be much easier to create a TRON Intelligent House today, and in that sense T-Engine brings us "back to the future," Prof. Sakamura said. He also pointed out that it is easy to attach many different things to T-Engine and play around with them, thus making it easy to have fun with T-Engine. Nevertheless, T-Engine will not hit the market until the second quarter of 2002, since documentation has to be created to support the new development environment. Hitachi's T-Engine implementation is five times faster than a cell-phone, and it can be fitted with either a cell-phone or a PDA screen. Mitsubishi's µT-Engine development environment, on the other hand, has been clocked transferring data at 25 Mbps per second, Prof. Sakamura said.
There were some interesting comments from the other participants in the technical session. A spokesman from Aplix Corporation, which has won acclaim with its portable Java on ITRON (JTRON) implementation JBlend that starts up in a mere two seconds, noted that applications can easily be moved from cell-phones to T-Engine. In other words, the "write once run everywhere" goal of that Java's creators had in mind when they designed the language has finally become a reality thanks to Japanese technology. In the world of PCs, that is not that case, the spokesman pointed out. He also noted that Aplix has had PDAs and other portable devices in mind from the beginning. JBlend, for example, has APIs for a vibrator function, which some portable device users need. He also said T-Engine is easy to work on and easy to develop on. For those who do not know, Aplix provides a JTRON product that is used by all the major cell-phone providers in Japan, thus making the company a key ally of Sun Microsystems Inc.
Personal Media Corporation's Akira Matsui said Cho Kanji for T-Engine was developed in a short period of time, which made it possible to demonstrate the new version of the BTRON3-specification operating system at TRON SHOW 2002. He demonstrated a video capture via MicroScript plus a character search to show how fast T-Engine with Cho Kanji is. In fact, he said, it is faster than "1B Note," a combination of a Matsushita Electric Industrial Co. developed laptop based on an Intel Corp. 16 MHz 80386SX microprocessor plus the BTRON1-specification operating system that became the first commercialization of BTRON marketed to the general public in 1991. He also pointed out that it is possible to attach a keyboard to T-Engine and use it like a personal computer. In addition, T-Engine with Cho Kanji can utilize dataware designed to run on the Cho Kanji series, i.e., the Kojien and K'ang Hsi dictionaries, which is not to mention Personal Media's popular Tompa hieroglyphic character font that has even appeared on Japanese television.
The most unusual demonstration of the technical session was a musical performance by representatives of Yamaha Corporation's using the company's new "EZ-EG" electronic guitar (click here for the Japanese language Web site), which can play itself and/or teach the user by lighting up the cords, and a mini electronic keyboard that have not been put on sale yet. The company had two musicians get up on the stage and give a short performance using these new ITRON-based musical instruments. These can be connected to T-Engine via a Musical Instruments Digital Interface (MIDI) input/output jack, thus making it possible to create new digital music. However, creating complicated digital music on a computer system like T-Engine also requires a knowledge of multitask programming, since you have to know what to do and where to do it inside a byte stream. Yamaha said it is planning to offer a super real-time programming course for that purpose. Prof. Sakamura said that he would like to see a T-Engine orchestra come into existence in the future.
Exhibits at TRON SHOW 2002
The real reason to go to a TRON show these days is to get a look at the latest technologies that have been developed on top of the world's first and only total computer architecture--and lots of people did just that. According to the reliable sources, attendance this year was in the thousands, which is a big jump from the mere hundreds who attended in previous years. They certainly got a good view of what is to come in the future in Japan. As they entered the site of TRON SHOW 2002, they got to look at a whole series of prototype gadgets based on T-Engine. Inside the exhibits area, the largest booth was by Personal Media, whose BTRON3-specification Cho Kanji 4 operating system drew a lot of attention. The second largest booth was Hitachi's T-Engine demo booth, which had quite a variety demonstrations, and the third largest booth was Aplix Corporation's booth, where there were various demonstrations of its JTRON-based JBlend language, particularly in cell-phones.
There were a total of 25 booths at TRON SHOW 2002, which is about average for a TRON show. Although that may seem very low by American or European technology show standards, a TRON show cannot be judged by foreign standards. That's because the TRON Project is not about developing a general-purpose architecture on top of which thousands of applications are supposed to run and hundreds of peripherals are supposed to be connected. Rather, it is about developing a series of basic operating systems that can be embedded in innumerable mono-functional devices that can be connected together for the benefit of humans. In that regard, Oki Electric Industry Industrial Co. Ltd. demonstrated its Bluetooth-based transceiver LSI chips that will probably be incorporated into TRON-based devices in the future. In addition, other companies marketed software to work in conjunction with embedded devices. Although there is little fanfare involved, this is the future of computing in Japan.
The second event on the final day of TRON SHOW 2002 was TRON Electronic Prosthetics Symposium (TEPS) 2002, which is popularly known as the "Enableware symposium." The theme of TEPS 2002 was the same as the previous year, the use of cell-phones to help the disabled. Prof. Ken Sakamura, who led off the event with a keynote lecture, said that Enableware communication will be realized with the next generation of cell-phone technologies, i.e., third generation, or 3G, technologies, that were commercialized in Japan in October 2001. Returning to a point he made earlier in the show, he noted that the appearance of the Internet, the completion of the TRON operating systems, and the implementation of the eTRON security architecture have finally put in place the basis for creating the ubiquitous computing he has been talking about for the last 20 years. Giving a hint of what to expect in the future, he said the cell-phone is the device for interfacing between humans and the innumerable computers of the ubiquitous computing environment, i.e., it will become what Americans call a "hub."
There were two representatives of Japanese industry who gave presentations at TEPS 2002. Mr. Hirotaka Nakano, Multimedia Laboratories NTT DoCoMo Inc., and Mr. Hirohiko Ohkubo, Mitsubishi Precision Co. Ltd. In addition, members of the Japanese disabled community spoke, as did their teachers.
Mr. Nakano gave a presentation on NTT DoCoMo's new 3G service, FOMA, or "Freedom of Mobile multimedia Access." The main advantage of FOMA is that it will offer higher throughput anywhere in the world, although in 2002 it will be limited to major cities in Japan. It is currently offered via three devices: a basic cell-phone, a video capable cell-phone compatible with NTT's ISDN video-phones ["Mopetto"], and a notebook PC adaptor. User fees are still high, 0.2 yen for a 128 byte packet, but they will be reduced to one-tenth over time. The new 3G cell-phones have an "i-appli," i.e., Java, architecture that makes them makes them application players, and a User Identification Module (UIM) on a chip. One of the long-range goals of NTT DoCoMo is to seamlessly connect with FOMA appliances using a plug and play feature; another is to create a new communications style in which cell-phones are used to strengthen humans ("i-borg"), something one of the visually disabled persons at TEPS 2002 actually attempted.
Mr. Ohkubo described a new infrared technology his firm developed that allows signs to talk with people via an electronic handset. Mitsubishi Precision developed this technology, he said, as a result of the establishment of the barrier-free concept for public facilities. The technology is directional, which means the "talking" stops if the handset is taken away from the target sign, and that is good for visually disabled who navigate using it. At each target sign, there is an "electronic label" that does the talking with the handset. This technology, which has won a best design award, has been tested extensively both in Japan and the U.S. In Japan, it was tested at a hotel, Shinkansen train stations, and city ward offices. In the U.S., San Francisco to be exact, it was tested at a library, a convention center, subway stations, public telephones, and pedestrian crossings. As a result of the tests, there was a call to make it smaller, and many users in Japan want it incorporated into cell-phones, i.e., the "hub" of the future in Japan.
Mr. Sadao Hasegawa, a visually disabled person from the Japan Braille Library, actually tried out one of the video capable 3G cell-phones mentioned above, NTT DoCoMo's FOMA P2101V (for a Japanese language description with lots of pictures, click here), although he didn't use it as the designers intended. The two color image formats supported by the P2101V (Common Image Format [CIF] 352 x 288 pixels; Quarter Common Intermediate Format [QCIF] 176 x 144 pixels) are mainly intended to enable video-telephone conferences at arm's length from the speaker's face. Mr. Hasegawa, who explained that in ordinary life the visually disabled need people to tell them where things are, purchased two P2101Vs and conducted a series of experiments in which he plugged in a headset to the cell-phone, pointed its video camera away from himself, and then asked a visually able person using the other cell-phone to relay information to him about what he or she could see via the headset's video camera.
The experiments that Mr. Hasegawa conducted covered a wide range of activities. He used the cell-phone's video camera to have someone scan newspaper headlines for him, to help him navigate while walking, to use ATMs inside convenience stores, and even to buy food, such as rice balls. As a result of his experiments, Mr. Hasegawa came to the following conclusions: (1) the resolution of the video camera is too low, (2) the field of vision is too narrow, and (3) the video camera function cannot be used in the dark. He recommended that an on-line "teleguide," i.e., a human who acts like a guide dog via a video connection, system be established as soon as possible. He also said that future cell-phones should be equipped with a Global Positioning System (GPS) function, a voice output function, and an infrared system, such as the one mentioned above. Mr. Nakano of NTT DoCoMo was stunned to see the visually disabled using the 3G cell-phone's video camera in the way it was used.
Another visually disabled person, a young student from Tokyo Metropolitan Hachioji School for the Blind named Jun Shionoya, gave a presentation on Fujitsu Ltd.'s F671i cell-phone (click here for a Japanese language description and lots of pictures), which has an e-mail voice readout function that he uses for receiving e-mail. Needless to say, the designers of this function never had the disabled when they designed it. That's why the user manual's instructions weren't written with the visually disabled in mind, which he complained about. Jun Shionoya's teacher, Mr. Yoshitake Misaki, took Jun up on stage stage and also had him help with his own presentation. Mr. Misaki, who used the BTRON's machine's SmartPoint software to give his presentation, spoke about his experiences helping the elderly, including an 80 year old blind woman, learn how to use computers for various goals in a special class. One woman wanted to write haiku again; another wanted to send e-mail to her grandchildren.
Mr. Hiroshi Kawamura of the Japanese Society for Rehabilitation of Persons with Disabilities spoke on the subject of digital television and what it can do for the disabled. Since a digital television is essentially a personal computer, he said, it can be equipped with special software to help the disabled. Software development along these lines is proceeding using Synchronized Multimedia Integrated Language (SMIL), an eXtensible Markup Language (XML) document type that specifies timing commands for executing audio, video, and graphical elements either sequentially or in parallel, and Digital Audio-based Information SYstem (DAISY), hardware and software standards for recording, storing, transferring, and reading digital audio books. The former is expected to be used to provide sign language for the hearing impaired. Another possibility is real-time caption-to-Braille conversion so that both the hearing and visually impaired could monitor a news broadcast.
During the panel discussion, which was the last event of TEPS 2002, Prof. Sakamura noted that the e-mail voice readout function in cell-phones was not intended for the disabled, and thus the manual instructions are not written with the disabled in mind. Mr. Nakano replied that he would look into the matter. He also said he didn't expect to see disabled people using the FOMA video camera function in the manner he saw them using it. Mr. Hasegawa talked of his early experiences with a Fujitsu 8-bit personal computer, and noted that the company supplied him with some free sets to help the visually disabled. He also said that he would like the an infrared function for maneuvering and a video camera with higher resolution installed in future cell-phones. Mr. Nakano said that the video camera resolution would rise in the future, but as the resolution rises so would the cost of transmitting the higher resolution data to a third party for interpretation.
Mr. Kawamura said he felt that there were a lot of good presentations at TEPS 2002, and he noted that there are lots of good programmers in India who could help out with software development, since there are also a lot of disabled people in India. He also said that a lot of peripherals will need to be developed, and thus there is a need to open the cell-phone architecture. Mr. Nakano countered that his company needs to know who wants to do want before opening the architecture. Prof. Sakamura said that a preparation period would be needed before opening up the specifications, while noting that T-Engine will help the hardware development efforts. Mr. Nakano also pointed out that his company has development efforts under way overseas also, and that as a result the company is expecting to receive lots of requests to do various new things using his firm's cell-phone as a platform.
Mr. Ohkubo noted that many testers of his company's infrared talking sign technology want it installed into cell-phones, and he asked Mr. Nakano to think about incorporating it into NTT DoCoMo's cell-phones in the future. Prof. Sakamura noted that some people also want the Bluetooth short-range wireless LAN technology incorporated into cell-phones in the future. Giving a glimpse of the future of networked computing in Japan, Prof. Sakamura also said that he wants to do things that cannot be done with personal computers at present. Mr. Misaki, with Jun Shionoya's manual problems in mind, said he would like cell-phone manufacturers to prepare voice manuals for the disabled in addition to Braille manuals. Mr. Hasegawa said he believes that the video camera function in the new video capable FOMA cell-phones could also be used to help the hearing impaired by transmitting sign language from one location to another. Unfortunately, the hearing impaired have yet to use FOMA.
____________________
[1] Although it began as an open hardware standard, the IBM-PC/AT-compatible personal computer architecture in recent years has become less and less open through the introduction of specialized hardware devices, i.e., Application Specific Integrated Circuits (ASICs), some of which incorporate "intelligence." As a result, it is nearly impossible to reverse engineer drivers, which is why there aren't that as many peripherals for the BTRON3-specification Cho Kanji operating system as there should be.
[2] CPU is a Central Processing Unit, MMU is a Memory Management Unit, RAM is Random Access Memory, I/O is Input/Output, I/F is InterFace, SIM is a Single In-line Module, LCD is a Liquid Crystal Display, and CODEC is COding DECoding.
[3] For people interested in the history of TRON-based products, the microprocessor clock speed of Hitachi's T-Engine is 5.3 times faster than µBTRON-based BrainPad TiPO, a powerful PDA developed by Seiko Instruments Inc. that is no longer in production. It should also be pointed that Hitachi has jointly developed with Xybernaut Corporation and put on the market a Windows CE-based device similar to T-Engine (click here and here, for details), so let it not be said that the T-Engine project is another clever Japanese "plot" against the American PC industry.
[4] Connectix Corp.'s Virtual PC 4.0 does not support Max OS X, but the recently announced version 5.0 does. Virtual PC allows for Cho Kanji 4 to be run on either of MS Windows or Mac OS as an application, which is very convenient for people who want to move data from one operating system to the other, in particular people who want to use Windows- or Mac-based graphics tools and Cho Kanji's unabridged kanji character set.