-- It's been a long time since we had an "Asking the Project Leader" interview. During that time, BTRON has evolved, and "Cho Kanji" [B-right/V R2], which can handle 130,000 characters, has at last appeared.
To cope with the implementation of "Cho Kanji," the BTRON OS that can handle 130,000 characters, we are preparing to set up a server called the "TRON Moji Tooroku Sentaa [TRON Character Resource Center]." I believe that it will be set up by about the time this issue of TRONWARE appears. This is because even though we are talking about 130,000 characters, there will be characters lacking, and, as I have said many times, even we don't know how many kanji [Chinese characters] there are. Thus we thought when there are characters lacking, we would need a center where we could register those characters that are lacking, using as a base the 130,000 characters that have been implemented at present. Some things that have already been requested are Japanese Jindai and Hotsuma characters. In addition, a request has come in asking us to include Maya characters. Also, I'm not joking, but I would like to register the Klingon characters of Star Trek. Besides these, there are many characters that people would like to include.
-- On what basis will you include new characters into the present character set of 130,000 characters?
In the TRON Character Resource Center, we do not have the slightest intention toward autocratically "dividing" things, saying we'll include this character but not this one, and so on. When there is a desire to register new characters, we intend to adopt a policy that in principle will recognize all registrations if: (1) they are clear in terms of rights, and (2) their source is clear.
-- What will you do for the concrete design of fonts?
I believe this is a matter of case by case. It's all right if someone provides us with a font. In cases when someone tells us there is no font but this character exists, there will also be occasions when we will create a font if we think it is necessary. There will be various cases. Of course, providing us with a font and allowing us to make it free would be best. However, in the case of people who make fonts a business, there will probably also be occasions when someone will provide a font but for a fee. There will be many cases, so I cannot make any sweeping statements. Whatever the case, we are expecting that the fonts in circulation will increase.
-- At present, the system only has the Mincho font for the 130,000 characters, right?
It appears that Personal Media is providing a Mincho-like font. Since it wouldn't be acceptable for the maker providing the operating system to provide only the character code framework without any fonts, what they're doing is providing the minimum fonts. From here on, it's all right with us if people who make fonts participate in the new world of BTRON, use this "Cho Kanji" as a platform, and provide us with fonts. It's also all right with us if they make a business out of it. At any rate, the fact is that we are expecting participation by individuals concerned with characters to increase. This is one platform, so in regard to starting a new business or industry on top of it, although people are saying it's because we have released the specifications for free, we have no intention of saying let's make everything free in regard to what people on the periphery have done. If that's all it is, I hope they will allow a new market to be born on top of this.
-- Listening to what you have just said, I get the feeling that there are also people with a misunderstanding of multicharacters.
Right. It seems there are people who misunderstand. They have the misunderstanding that we will one-sidedly decide everything in regard multicharacters/multi-kanji, that we'll end up providing even the various fonts, and thus the business of creating and distributing fonts will disappear. I realize there are individuals with a sense of resistance. On the contrary, for people and companies doing character design, the playing field will broaden. If there is no common ground of this kind to support multicharacters, won't it be difficult to do business even if one intends to? When the platforms are all specialized and disjointed, when even the user defined character codes differ from each other, even if you try to do business, you'll have a hard time trying to cope.
-- I understand what you are saying, but I believe there is also a call from small and medium printers saying that even if you put forward 130,000 characters, it will not by any means be possible to have a complete set.
Not at all. Aren't they mistaken? They seem to think that one must buy a complete set of all the characters. But in this playing field of BTRON-based "Cho Kanji," first of all there are fonts that have already been loaded. Even when print quality fonts are needed, it doesn't mean you have to get a complete set of all of them. In reality, when old-style fonts have been lacking, people have created them. Moreover, if you have created needed fonts, it's all right to attach a price to them and sell them. That people with doubts as to "whether 130,000 characters are needed" say such things leads me to wonder whether they aren't contradicting themselves. If someone writes a book, it is almost impossible that all the pages will be completely filled up with characters that are not ordinarily used. In all likelihood, when characters that are lacking appear, if one acquires and employs a form for making them, then they can use them from the next time. If you try doing this at a printer's shop, they'll make them for you; and if you put them down in the registry at the TRON Character Resource Center, the center will be distributing information about them on the network. Then if someone appears who needs them and an opportunity to sell them is created, it will come about that you can sell fonts in excess of your own demand. Exactly because there is a scheme such as a center for registering characters, the sphere of business will also expand.
-- I also think that's so. You said that the registration center will open soon. Will people be able to search through what sorts of characters are supported with the Cho Kanji operating system from the network?
Of course, they will be able to. With BTRON-based "Cho Kanji" in its present form, it will be possible to search via the Internet what characters are supported, and what characters are not supported. When a certain character does not exist, if the previously mentioned conditions are fulfilled, and if one one sends the character to us at the center, first it will be temporarily registered, and it will be possible to look at it unconditionally. At the stage at which it is officially recognized that it has met the conditions, the character is then registered at the center. As I previously mentioned, this center will soon come into existence.
-- That's epochal, isn't it. By the way, among the 130,000 characters of "Cho Kanji," there are several of the world's kanji character sets. For example, there are the Konjaku Mojikyo and the Dai Kan-Wa Jiten character sets, JIS levels 1 and 2, and JIS auxiliary kanji, aren't there. That they have been entered without modification also means that there are overlapped characters, doesn't it.
That's right. It's just as you say. That's the essence of it, but, as I have been saying, we haven't made any divisions on our part. If it came down to making divisions, then some one person would have to look at everything and make a judgment as to whether this character should or should not be registered. Moreover, when you set out to do this, when it comes to 100,000 or 200,000 characters for a single person, I believe that person will reach his/her limits.
-- I agree.
I often mention this, but to the extent that even Prof. Akiho Yamaguchi, who is said to be a kanji expert and is working on the GT Project, jokingly says, "even I don't know kanji," the fact is that there are lots of kanji. When even a professor who has specialized in kanji humbly says he doesn't know them, I believe that a person who hasn't specialized in kanji will not necessarily know them. Thus we believe that we should include all these large character sets without modification.
-- After you include them, then what will you do?
Then, if there is a table of correspondences of the type that points out which characters are the same among the large character sets, it will be convenient. Although there are things that we put out for free, if someone decided to try to create a dictionary of correspondence tables, it could be done.
-- If someone did that, it would be possible that something akin to a sort of giant code converter would come into existence, right?
I have an image of the android C3PO that appears in Star Wars. I would like to create a super converter for the character codes in the world. Through the cooperation of various people, I would definitely like to include large character sets from throughout the world and build on top of them a system that would serve as a hub.
-- Now that you mention it, the tables of Unicode character listings that have appeared on "Cho Kanji" are splendid, aren't they? Ironically, a Unicode unification chart cannot be created with a Unicode machine (Fig. 1).
That's right. If you showed this to people, the first people to tell you they wanted it would be the people developing Unicode-based systems. Seriously speaking, the fact is that by using BTRON-based "Cho Kanji," we can for the first time display where there is a correspondence with what. This is not quite possible with other computer systems. Because there are unified, the characters cannot be displayed.
-- Now that you mention it, it seems there is talk that the mapping tables put out by Unicode.org are really ill prepared. There is in a book called The World of Character Codes [title directly translated from Japanese] that has recently been published, and, for example, in regard to 4F47 and 4F2B of Unicode 2.0, why is 56-89 of the Chinese GB/T 12345 standard in correspondence with Unicode 4F47 (Fig. 2), etc., etc.?
Oh, is that so?
-- Cho Kanji has been reported about in quite a variety of sources. It seems like the basics of it have been understood.
Initially, there were a lot of low-level questions, such as, even if you offer as many as 130,000 characters, what will you do with them; or will ordinary people really use this kind of thing? However, recently those types of questions have disappeared. Anyone, in the end, what anyone will normally use is probably inside 10,000 characters, but one person will say he/she wants plus 100 characters, one person will say he/she wants a separate set of another 50 characters, one person will say in addition 200 characters. When you are doing this type of thing, if you do not have on the order of about 130,000 characters, things really become hopeless. If that's the case, then it's a matter of making haste slowly, so ultimately, if you do not do something like what BTRON is now doing, it becomes a matter of being unable to do it (Fig. 3).
-- How's the reaction from outside Japan?
Although there were people who said that when you put out something like this, you are going against the recent age of globalization and internationalization, that you're doing something closed that will come into common use only inside Japan, those types of things are completely false. They, of all people, are the ones who are not internationalized. In a global age on a global scale, the very differences in culture possessed by each nation are important. Different cultures holding exchanges among themselves is interesting, [but that's impossible] when there is no language other than English, when the world becomes monoculturalized. In the Book of Genesis in the Bible there appears an anecdote in which people originally spoke a single language, but humans became prideful, and just as they had erected the Tower of Babel, God became angry and caused confusion among them by making them speak different languages so that they could not understand each other. Languages are not going to collapse into one. I wonder if they haven't gotten the gist of it wrong. Making anything into one thing is not interesting. Things are impossible with one language, and the fact that there are various languages in the world in itself shows the multifaceted nature of [human] culture.
It is our idea that the computer must be something that supports to the last the multifaceted nature of [human] culture. Nevertheless, not only in Japan but throughout the world, the understanding of what we are attempting to do with BTRON is deepening. As was written in the Project Leader's Opinion from Vol. 59 of TRONWARE, Dr. [Woong-Keun] Lee, a former professor of Seoul National University's Graduate School of Public Administration, said this type of thing is needed in work such as putting Chosôn Wangjok Sillok [Royal Chronicles of (the Kingdom of) Chosôn] in electronic media form, which he carried out. If you say that's obvious, then what's obvious should be obvious. In addition, there are several universities in the U.S. where Japanese culture is researched, and as for places of the type that have research centers, in Europe also interest in Japanese culture is rising. Thus I have heard that inquiries about "Cho Kanji" are coming in from all over the world. I wonder if it's not a case of these people, who without knowing think it is something that is only Japanese, not understanding true globalization. I wonder whether they, of all people, aren't wrong in believing that anything from the U.S. is fine, so just bring it over. Throughout the world, there is great respect for preserving culture--to the degree that is mentioned when expressing respect for efforts to leave behind a cultural legacy in the manner of Larousse of the French encyclopedia.
-- That's because of the trend in Japan that all you have to do is make money using computers or the Internet, isn't it?
Today, because of the recession that may be so, but I believe it's regrettable there is a strengthening trend in which people ride a boom in which all you have to do is simply get a venture started and then get the company listed on the stock exchange. The means and the ends are reversed in this. As a result, the wishes of serious scientists and engineers who are truly trying to create unique and outstanding products get cut off. By solely making money, Japan's reputation in the world will also end up falling. Be that as it may, thanks to various individuals, such as the Konjaku Mojikyo people and the Dai Kan-Wa Jiten people of Taishukan Publishing Co., we are finally being understood. On this also, there has been a misunderstanding from way back, but as a purely private project that has been in continuous operation without the use of taxes, I'm happy to have been able to obtain the cooperation of very many people.
-- However, there are also people who say all sorts of strange things, you know.
That's right. I wonder if they aren't saying those things on purpose. They say things like "TRON is a national project," or that "mountains of taxes have been sunk into it." TRON has not received even one yen of taxes for its research. I'd like to ask who is circulating the story that TRON research is being conducted as a national project through the direct risking of tax money? I can't believe it is being confused with things like the Fifth Generation Project or the Sigma Project. Because I do research at the University of Tokyo, and since the country is paying my salary and equipment expenses, there may be some misunderstanding, though. Well, there's no use getting angry about it. But since it is not a project mainly based on the investment of taxes, there is no reason for this and that to be said. Although there are people saying all manner of things that one would grow tired of denying, actually I get the feeling that recently the group of people supporting or praising the project is greater than the number of people saying strange things about it.
-- By the way, compared to other operating systems, BTRON embodies a very unique way of thinking. I think it's a matter of the more you use it the better. Is that kind of understanding progressing?
Compared to before, I think it has become more understood, but in the end since the way of thinking is completely different from Windows and Macintosh, I wonder if there aren't individuals who are bewildered. Nevertheless, since the World Wide Web of the Internet has become very popular recently, I wonder if it hasn't become very easy to understand, though. Well, if you would take a look at the figure (Fig. 4), as you can see, if we talk about the differences with Windows and Macintosh, first there are no icons. When talking about outward differences, this is a big one.
If you ask why they aren't there, the reason is that the operating system runs on top of the real object/virtual object [filing] model of the BTRON [operating system] model. Furthermore, this is because the human interface model is also based on the real object/virtual object model. When it comes to differences with Windows and Macintosh, this is a difference at the level of basic philosophy.
-- People who haven't studied it very much are greatly mistaken in saying things like you looked at Macintosh and created it, aren't they?
Clearly, showing a portion of the data using the concept of a window inside a limited screen is a universal concept in computer science separate from the operating system model that began around the latter half of the 1960s with the Augment System of Douglas Engelbart at SRI. Then, at Xerox PARC, overlapping windows like papers arranged in a row were born with the Alto research computer (1973), which is the forefather of today's personal computer. The integrated operating environment, what we call a GUI-based environment today, was established in principle with the Xerox Star (1981), which aimed at practicalizing the Alto. The Lisa (1983) and Macintosh (1984) were developed at Apple after the company came into contact with these, and then Windows was conceived at Microsoft. Don't you remember that at the time Macintosh appeared, it received severe criticism as an imitation of Xerox PARC [technologies] that was a useless toy? However, throughout the 50 year history of the computer, there is nothing that is a completely original assemblage that didn't imitate anything. Even Alan Kay's Smalltalk, which exists as the founder of object-oriented programming that is common sense today, is something that was born following the existence of precursors, such as Ivan Sutherland's Sketchpad system (1962) and the simulation language Simula 67 (1967). I don't know how things are elsewhere, but in the world of computers, no matter how original something is, it is something that further expands on the efforts of predecessors. People who say that BTRON is an imitation because it uses the concept of a window inside a screen are not qualified to design computers. Having said that, when you hammer out a concept such as [the TRON concept of] "computers everywhere," almost everyone without understanding exactly what's new thinks you're engaging in big talk, and then 10 years later, because the U.S. becomes engaged in it, they say that was original. Even though it was valued afterwards, it's distressing.
-- I understand. I understand.
Even if, as a matter of course, we are using the window concept, the fact is that BTRON is markedly different from others in that the operation model is in the form of the real object/virtual object model. In the past, I had a lot of trouble getting people to understand, but recently thanks to the spread of the World Wide Web it has become a lot easier to explain. In World Wide Web terms, a virtual object is a link that points to a real object--a collection of data. The virtual object directly points to the real object in a way that makes it possible to access it in like manner, seamlessly and without discontinuity. It doesn't matter whether it's in the memory, on a disk, or at a network location. Moreover, what's important is that it can have mixed contents of text and graphics. By means of virtual objects, a state is created in which links to other data are buried inside data in the manner of World Wide Web pages. Leaving aside whether what exists now is like this, what I originally wanted to attempt with the real object/virtual object model is this type of seamless permeability that exceeds the network. When you open a virtual object that points to a network location, the real object of the network location opens. Taking in data through the network; if they're in the memory, having them immediately appear; they might even be in the disk--these are the things I wanted to do. I wanted to make best use of the merits of being able to manage data anywhere, without depending on single source management and physical locations where they had been stored. However, there was a difficulty as to how to do the network. There were various things such as NetWare and OSI, but in the end, it has been narrowed down to the Internet.
-- Haven't you thought of an independent TRON network?
We've done all sorts of research. However, because Internet technologies, as might be expected, are well made technologies, having been developed for defense purposes by the U.S., which spent time and money on them. Since more than anything else they are open and a lot of research and development has been added on that base, we didn't think about try to oppose them with something similar to them, as you might expect. At any rate, I'm repeating myself, but because the TRON Project is a private, non-profit project that is in no way a national project--an NGO or NPO, in recent parlance--what we can do is limited. We have no intention of doing anything such as all encompassing research on anything and everything. That was never possible from the the start. Above all, since Internet technologies are open, we also believed that their ideological affinity with TRON would be good.
-- Character code is the reverse, isn't it. You wrote somewhere that since the first draft of ISO10646 was initially progressing in an open direction that allowed for diverse codes, you were waiting and seeing with the intention of including those results. But Unicode suddenly appeared changing it in character, and thus the result was that you have ended up having to actively develop in opposition to it.
That's definitely so. I'm repeating myself, but in the TRON Project no consideration is given to excluding anything whatsoever outside of TRON. Putting everything in the world in TRON is unreasonable from the start, and for a sense of diversity, it would be better if things didn't turn out that way. Thus the point as to how to have relationships with other systems is important. What we call operating systems and applications are after all the problem of the "private within the public-private." However, what we call the network, data formats--including character codes--and furthermore the contents themselves are the "public within the public-private." These will probably gradually grow in importance in the future.
-- In that way, in the sense that it is controlled by other trends, this is a field in which difficult problems variously appear even in areas outside of technology.
Right, right. Things such as the character code problem were exactly like that. If there are things that are open and good, we would like to go on actively incorporating them. There are a mountain of things that we would like to do, that we must do, but there isn't sufficient research and development power, and we're suffering hardship because of that. Therefore, even for the character code, if someone had created a good one for us, I was prepared to thankfully incorporate it.
-- In the end, you could not help hammering it out yourself.
That's right. Therefore, conversely, that the network is in the form of the World Wide Web, that open network technology with a high affinity with TRON is spreading is just what I have wished for.
-- Then you're saying that things are generally fine as they are.
Even now, implementation also is slowly approaching the ideal. Although I wouldn't say its complete, at the end of this way of thinking lies the future. In the case of conventional operating systems, such as Microsoft's Windows, you have to state clearly where on what disk with directory designations. Because BTRON is something that attempts to designate this unspecifically, the way of thinking ends up becoming different. In specifying files, there is a big difference between having to write them while being conscious of physical locations and writing without being forced to be conscious of physical locations. This difference is mainly due to the fact that it is effective for the BTRON light application idea that "each [application] realizes the necessary functions through a combination of function blocks" locally, but I believe it also has meaning in network terms. Things such as NFS [Network Filing System] are attempting to bring a tree shaped hierarchy inside the network, but ultimately it's more convenient to directly indicate what we call resources on the network in the manner of URLs with a flat address. Well, inside that domain, whether the filing system is a UNIX style downward tree shape, or whether we bring in here a complete network shape in the manner of BTRON, directly it doesn't matter. However, when we consider maintaining consistency with network applications, I think that BTRON is more tractable.
-- For your future course, do you intend to approach the ideal course you talked about with an Internet-based proposal?
Yes, what we're thinking about at present after all is trying to get Cho Kanji distributed on the Internet, and trying to make TAD [TRON Application Databus] flow through the Internet without modifications. We would like to contribute in the direction of making Internet data formats rich. We are thinking that later on we would like to make a TRON-like proposal in the direction of attribute description and semantic description, which is important in EC [Electronic Commerce] and museum applications.
-- Windows and Macintosh have been mentioned. Compared to both of them, what do you think of the implementation and level of finish of the fine points of BTRON?
If I mention the fine details of that, we'll get into a difficult discussion. However, there are still some rough spots, so it would good to do some fine tuning. However, I cannot say this sweepingly. There are differences in the respective models, and there are rough spots in Windows, also. When Windows users use Macintosh, there are also places where they say you can't do this type of thing in a simple manner. If we get to talking about Macintosh, the counter opinion will appear that since Macintosh is faithful to its dogma, convenient functions aren't crammed one after another in the manner of extending an inn at a hot spring. It's difficult to say in a word this is rough when comparing operating systems with different models. However, even though a comparison is difficult, I believe that there are places in BTRON where it would better to do more fine tuning. But, since development manpower is limited, there will probably be a sequence to the fine tuning, also.
-- Is there anything else?
This also may be in with fine tuning, but as one gradually creates real objects, the desktop ends up getting buried, and it's a lot of trouble to put that in order or do housekeeping. If there is not some kind of strong cleaner or broom, the desktop seems disordered. It would be a function that would put the contents of the cabinets in order and line them up neatly, and that would search for real object names and their contents. Something that would make it possible to make registrations in a simple manner in MicroCard with the date attached. Since the number of users has increased, I believe that a software tool to solve this will appear. It seems like a part of it has appeared as free software. (See the "Text Search Free Software" column below.) In this area, I think it would be better if there was a little more strengthening. For people who create a lot of virtual objects and real objects, I believe it's a tool they would really like to have.
-- I believe that individual users account for most of those using BTRON at present. However, it seems like a call has also been made to use BTRON in business in the future. I have heard that there has been an increase in people who are using BTRON in business also.
Since it has become possible for BTRON to use networks, it seems like an opportunity for it to be used in business has also appeared. I think that's so for people who use lots of kanji. Another direction is using today's servers as [application] servers. Recently, there has been an increase in server applications of the kind that are accessed using a World Wide Web browser. I believe there is the possibility of using "Cho Kanji" in business further by utilizing its multicharacter functions. Without altering things in the main functions very much, the user would employ it as a client terminal based on the combination of a browser that can use multi-kanji and some software on the server. I have heard that there are a lot of people interested in this.
-- You mean a Cho Kanji-compatible browser, don't you.
That's right. The possibility of something like that appearing in the future is extremely high.
-- It seems like the TRON Project is putting a lot of effort into BTRON in particular lately.
The TRON Project is something that began nearly 20 years. Thanks to your help, ITRON is already in use in various places throughout the world and is becoming a de facto standard. As for recent trends, ITRON is being used for engine control in Toyota automobiles, and it coming to be used in all sorts of places, beginning with facsimile machines and handheld telephones. And it seems like American manufacturers are moving toward promoting the ITRON standard as an industrial standard. As for BTRON, we are in a position in which we would like to put effort into it and promote it. It's an operating system that is well matched to the human interface, to the boundary between humans and computers, and thus on our part, we are paying attention to ways to spread it. Also, there is a thing we call TRON-GUI, which is a human-machine interface from which the file structure and so on have been removed from BTRON and directly attached to ITRON. This is the extraction of an interface for small-scale embedded systems, which I believe is also being introduced in this issue of TRONWARE. And it is also scheduled to be introduced at TRONSHOW 2000. BTRON is best suited as the parent system when you use this type of system as a child device. Since an advantage appears in this area, I feel that BTRON also has progressed somewhat, so I would like you to expectations [for BTRON] in the future.
"Text Search," a freeware application written by Pi and ported to B-right/V by Gean-canough, is well known, and thus it is contained in the Cho Kanji (B-right/V R2) free software collection. This is quite a convenient tool that searches for real object names and character strings inside documents created with the Basic Text Editor and the Basic Figure Editor.
To install it, all you have to do is unfreeze it, since it is compressed with the Library Manager, and then drag it to the location of the system environment settings version. A stencil called "Search Box" will appear in the Tablet Box, so bring this to the desktop. When you open this with "Virtual Object List" from the Execution menu, a window exactly the same as a cabinet opens up. Here, you throw in documents, cabinets, etc., to be searched.
Then, when you open with "Text Search" from the Execution menu, a search panel appears. When you enter a character string you would like to search for, the virtual objects that were found are displayed. Using the search button, it is possible to search for either a character string that exists in a real object name or a character string that exists inside a document.
At the time this was written, Cho Kanji had yet to be shipped, so this software is not compatible with the 130,000 characters of Cho Kanji. However, it is a convenient tool that we expect will become Cho Kanji compatible, also.
The above interview with TRON Project Leader Dr. Ken Sakamura appeared on pages 12-21 in Vol. 60 of TRONWARE. It was translated and loaded onto this Web page with the permission of Personal Media Corporation.
Copyright © 1999 Personal Media Corporation
Copyright © 2000 Sakamura Laboratory, University Museum, University of Tokyo