sfjp280.png

Project of This Month

http://sourceforge.jp/projects/sourceforge/wiki/potm_50x50.gif August 2007 - GLOBALBASE

Table of Contents
  1. Project Overview
  2. Interview to the Project Administrator (Mr. Hirohisa Mori)
    1. What kind of software is GLONABALBASE?
    2. What motivated you to start the project? How did you start the project?
    3. Who are the target users of this software?
    4. How many people do you think use the software?
    5. What are the typical usages of this software?
    6. When do you feel that the project is progressing successfully?
    7. What is the event that surprized you the most while you worked on this project?
    8. What is the most difficult aspect of this project?
    9. Why did this project begin to be recognized widely?
    10. What is the future direction of the project?
    11. What requests do you receive from the users?
    12. What are the features of the software you are proud of?
    13. What points of the project do you want to work over, if possible?
    14. How do you coordinate the project?
    15. Are you a dedicated developer of this project? Or do you have another job?
    16. What is your development environment?
    17. Version History
    18. How can I contribute to this project?
    19. Please let us know your requests to SourceForge.jp.

Project Overview

GLOBALBASE is a completely autonomous distributed GIS. Our ultimate goal is for geographical information created and made available by individual users in GLOBALBASE to grow unbounded in volume, interlinked with each other, eventually forming a mirror image of the Earth itself (like a parallel world).

Most Web-based geographical information systems employ a mechanism where the base maps are maintained on their own servers, protected from outside interference. GLOBALBASE is special in that it allows geographical information of any form, including base maps, to be coordinated with information on other servers by overlapping or connecting. Not only may information providers combine and upload geographical information at their own will, but the viewer side may also pick and choose any combination of geographical information he/she wishes to be displayed in his/her browser.

The aim of this project is to develop and implement the new technologies required to build such an open and free geographical information society.

Interview to the Project Administrator (Mr. Hirohisa Mori)

What kind of software is GLOBALBASE?

GLOBALBASE is like the "Web of maps." You might have heard the term "Web-GIS." At first glance, GLOBALBASE is similar to a Web-GIS, but the underlying concept is very different. The "Web of maps" is a network (Web) where an infinite number of "maps" are linked together based on mutual positional relationships and mutually associated. Using this mechanism, "maps" provided by an infinite number of people and organizations can be connected with each other autonomously, which can be viewed as a kind of mirror Earth in cyberspace.

The relationship between GLOBALBASE and maps is exactly the same as the relationship between WWW and documents. Documents uploaded to the WWW link to each other via hypertext links and can in principle be thought of as a single text spreading all over the Earth.

Web-GISs are different in this respect. For instance, as you may know, it is not possible to overlap or connect maps in e.g. GoogleMaps or Mapion at the same level. Maps in Web-GISs are, in a way, like illustrations of a document. Conceptually, it is the document that is hyperlinked, not the illustrations (although this may not necessarily be the case on the implementation level).

Now, let's take a look at our architecture, in which maps are properly hyperlinked both on the conceptual and the implementation level. The following images show views of the world as presented in COSMOS, the dedicated browser of GLOBALBASE.

image2_thumb.png

Figure 1. COSMOS Opening Window (→Display with original size)

As shown in Figure 1, when you launch COSMOS with the default settings, the Earth is displayed just as in other spatial browsers. From here, you can move to any place you wish on the Earth, by dragging the mouse on the screen or clicking the zoom in/out buttons at the upper left corner.

The main point that makes GLOBALBASE clearly different from other spatial browsers is that it allows editing the conditions of each map layer and search conditions via the tabs below the browser panel, thereby for example displaying maps of various times in history or various types of maps. For example, Figure 2 shows a browser view where several local maps of some of humanity's great monuments are diplayed. You can see that the Earth is displayed with the local maps superimposed, just like a patchwork.

human2_thumb.png

Figure 2. From Humanity's Monument Database at the Kochi University of Technology (→Display with original size)

The next figure shows how a map of Kyoto in the Edo period (around 1750) is overlapped with a corresponding current map using contents on a server at the International Research Center for Japanese Studies (Figure 3).

oldmap2_thumb.png

Figure 3. Map from International Research Center for Japanese Studies overlapped with Modern Map (→Display with original size)

Such different contents are stored in servers of individual organizations, but they are not treated as independent entities; all of them are connected via positional information. For example, in the case of humanity's monuments, satellite pictures owned by the Osaka City University have been augmented by data supplied by the Kochi University of Technology. It is of course also possible to overlap the data from the database of humanity's monuments at the Kochi University of Technology onto the old maps provided by the International Research Center for Japanese Studies, as well as any other combination you can think of.

The potential of the "Web of maps" is not limited to these applications. Figure 4 shows another example of overlapping information. Here, you can see aerial photos and satellite photos of a part of the Kansai area uploaded by various people displayed while maintaining their correct positional relationship. Each of these photos is a huge image, comprising from several gigabytes to several hundred gigabytes of data.

p6_thumb.png

Figure 4. Large Collection of Uploaded Aerial Photos/Satellite Photos (→Display with original size)

Normal GISs provide pre-defined coordinate systems common to the entire world called reference systems. By converting all data to such a reference system, overlapping and other processing can be performed on the data. GLOBALBASE, on the other hand, does not impose such a reference system. Instead, it allows users to define any coordinate systems they wish, and represent the contents therein. It employs a mechanism for browsing networks comprising huge numbers of coordinate systems, relying only on relative positional relationships between the coordinate systems and relative scales of enlargement. For this reason, anything, even very tiny things, can be displayed by gradually increasing the resolution of the linked maps. For example, Figure 5.(1) defines a coordinate system on the desk of a certain company. You can even see luster of coat of a stuffed toy placed on the table. This clearly illustrates that GLOBALBASE allows defining very small coordinate systems. Figure 5.(2) shows the same coordinate system inserted in another coordinate system, in this case the floorplan of the company. This floorplan can then be inserted in a map of Kyoto and so forth, repeating this process as many times as necessary.

dog1_thumb.png

Figure 5.(1) A View of a Desk of a Certain Company (→Display with original size)

dog2_thumb.png

Figure 5.(2) Floorplan of the Company (→Display with original size)

In this way, the internal layout of an integrated circuit and the entire Earth can be represented using the exact same architecture. Note that the view above can be displayed in COSMOS by setting the time to around 2000 and zooming in on Karasumaoike in Kyoto.

What motivated you to start the project? How did you start the project?

I worked as an asistant at the University Museum at the University of Tokyo from 1996 to 1999. During the last year, I was in charge of an exhibition called "Birth of News -- Information World of Kawara-ban and Newspaper Nishiki-e (about tile block print and color woodblock print newspapers). Together with a professor studying the history of natural disasters, I created various maps of disaster areas brought about by the Ansei Edo Earthquake (Ansei 2, 1855). I created maps showing how the manors of daimyos and suburban residences and trademen's houses in the Edo peropd were damaged, which I then printed on transparent acrylic sheets such that they could be overlapped on Edo area maps by physically placing the sheets on top of each other. It was a truly analog system.

My creation turned out to be quite unpopular. I received comments such as, "Why on earth did you print images created in Illustrator and VectorWorks on acrylic sheets and place them on top of each other? Why didn't you think about displaying the overlapped maps directy on the computers?" Receiving these bad reviews motivated me to start development and research of GLOBALBASE.

Afterward, my job description changed to creating databases for the International Research Center for Japanese Studies, where I was asked to create databases of old maps; this served to spur me on as well. Once started, I found it unexpectedly interesting. The maps themselves and the associated data are interesting to begin with. I was surprised to know that so many old maps are still kept throughout Japan.

Then I came up with the idea that by applying computer science technologies to this database project, it might be possible to connect old maps to create a complete, albeit deformed, overview of the Japanese archipelago. Then, if that idea was extended to cover the entire world, the end result would be a deformed map of the entire Earth. Furthermore, if there were a protocol that would allow connecting any maps, of arbitrary nature, on overlay networks, many people and researchers could participate in creating such a deformed world.

I realized that this idea could be extended even to hand-drawn maps. If hand-drawn maps can be handled, absolutely anybody would be able to participate. If such a project were to be a success, who could say what kind of outrageous world map might be created!?

I believe it was in October 1999 that the International Research Center for Japanese Studies decided to launch the GLOBALBASE project. At first, I expected the project to be over in a couple of years, but now 8 years have gone by already.

Who are the target users of the software?

I want as many people as possible to view GLOBALBASE maps, regardless of age and sex, both professionals and non-professionals. I also want lots of people not only to view, but also create and upload maps, because the world created in GLOBALBASE will become truly interesting if each user uploads maps of their imminent environment, and such maps end up linked just like web documents. However, I started to realize that this poses a considerably difficult problem.

The type of people who are "interested in maps" are often completely disinterested in "network technologies" and such. If a user wants to upload maps, she/he must launch a server. In order to do so, he/she must first be able to execute basic commands in Linux or a similar OS. What is worse, if one does not have knowledge of TCP/IP, one cannot link maps properly. Currently, although it is an autonomous distribution system, GLOBALBASE is still a client-server architecture. Thus, I figured it would be necessary to construct a complete P2P architecture eventually.

I also eventually began to understand that another bottleneck is a common perception among the general public regarding the concept of "maps" itself. There seem to be many people who think that maps are too technical to draw for ordinary people. I believe that breaking this barrier is one of the main challenges we have to deal with.

How many people do you think are using this software?

I believe quite a considerable number of peoplehas downloaded the software just for "viewing maps." We receive emails with comments, or complaints that the software does not work etc., from time to time. GLOBALBASE is mentioned in blogs in South Korea, the Website is accessed from DoD (Department of Defense) or similar, etc. All in all, I feel that GLOBALBASE is proving itself to be decent entertainment.

However, the software for "creating and uploading map data," as I mentioned earlier, is far from successful. Apart from the impression of "maps being difficult," there are various reasons for this such as difficulties in handling servers, insufficient manuals and other non-technical issues that have to be solved before the usage becomes smooth and painless. Nonetheless, I believe that there are many potential users out there who wish to provide their own maps.

What are the typical usages of the software?

The primary usage is to "view maps and have fun." Satellite photos, for example, are fascinating. You will be suprised by how many interesting geographical formations can be found on the Earth. Also, there are highly unconventional maps, such as "Tomorrow's Maps", which are maps drawn by children, or sound-maps at Research Institute of Soundscape Planning, which are provided through the use of the Web gateway function of GLOBALBASE. I want people to realise that maps are not something difficult, as these examples show.

ashita-map_thumb.png

FIgure 6. Tomorrow's Maps (→Display with original size)

Please take a look at Figure 6, "Tomorrow's Maps." In normal GISs, it would be unthinkable to view imaginary maps freely drawn by children. In many of the exhibitions and similar events we attended in past, various specialists who tried GLOBALBASE seemed to have a hard time understanding that so many different maps, from survey maps to such pictorial maps, could be handled at the same level. However, it is after all easier for specialists to accept this fact without hesitation.

I hope that more and more people will launch servers and upload their original maps and the maps they have on hand from now on. For example, I would very much like students studying information systems to upload campus maps and similar to GLOBALBASE.

webgateway1_thumb.png

Figure 7. GLOBALBASE - WWW Gateway (→Display with original size)

Moreover, some people may hesitate to use the dedicated browser. Figure 7 shows an example of how our gateway technology allows translating GLOBALBASE contents into WWW format and insert it in websites, so it may be viewed in a normal browser. It is also possible to use maps provided by other users and overlap them with your own maps, and view the results in a WWW browser. We plan to provide various gateway templates with different designs as well.

When do you feel that the project is progressing successfully?

When tuning progresses well, bugs will (hopefully) be fixed... I often tell my wife that "getting rid of those bugs feels as if getting rid of a big lump of earwax." Sorry for being so trivial. A particularly good example is a bug fix event we had during the fall last year. For several years in the past, the load on the autonomous network was much heavier than theoretically considered appropriate, but we simply could not find the reason no matter how hard we tried. Finally, we found an error of just one character. We should have written:

flags |= F_HOGEHOGE;

but, instead, we wrote:

flags = F_HOGEHOGE;

thereby deleting all other important flags. When we corrected this error, the network load throughout the entire Japan was reduced immediately.

In a way the project is an ongoing struggle that requires hard work and efforts. I have witnessed impressive breakthroughs in software, such as Mosaic, Linux and so on, which seem particularly brilliant because they came into sudden prominence virtually out of nowhere. However, the true brilliance of most such pieces of software can only be revealed and maintained by polishing tirelessly every day, just like with antiques.

We can continue our efforts because we have people who understand and support the project. I feel that the project is progressing successfully whenever we encounter other such people. We do not need to meet these supportive people face to face, we just need to know they exist. For example, sometimes we find traces of people trying to access documents explaining how to set up GLOBALBASE servers in the logs, and we are very much touched.

On another note, I believe it was a success that we employed the concept of open source code from an early stage. At the beginning, we were criticized for not delivering software living up to the expectations of a commercial product etc., but by now I am sure that our project is marching with the time. I first encountered open source in 1989 when I was a university student. I started using it when I migrated NetRelease1 on ITRON. Whether NetRelease1 is open source software or not is arguable, but nonetheless, it is (part of) the source code that triggered the Berkeley and AT&T UNIX lawsuit. After this, Richard Stallman's concept of free software began to spread, slowly but surely.

I remember decoding the source code with absorbing interest. While on that subject, I have a suggestion to all computer science students. It is my impression that computer science students no longer tend to read sorce code lately. I suggest that if you are studying to be a computer scientist and want to be a developer/resercher, you should take the time to read as much source code as you possibly can. There are no novelists who do not read novels, either. Excellent source code provides you with the mental food that will allow you to become good developers and researchers yourselves. Open source code is not based on the happy-go-lucky way of thinking that you can "steal" others' source code and make your own development easier. It is meant to help future developers and researchers to study and grow by reading wonderful source code written by other reserchers. This is also clearly stated by Mr. Stallman.

Later, taking the opportunity of migration of NetRelease1, I got involved with the TRON project. Soon thereafter, when the TRON project faced a crisis as the US accused it of being a trade barrier using the Super 302 trade law, I became keenly aware of the importance of the open source concept. The TRON project claimed to adopt an open architecture and open specifications, but there was no open source code for BTRON and CTRON. Licensed companies were afraid of becoming the targets of lawsuits and bad publicity and not only stopped product sales, but also closed their source code. Several years later, after repeated negotiations by Mr. Ken Sakamura, the project achieved commercialization, but its time had already passed. If open source BTRON had existed at that time, on the other hand, I believe the source code might have survived even to this day.

Based on the experience back then, I decided to employ the open source code policy strictly for all software related to the GLOBALBASE project. Open source code projects may be low-key compared to open architecture/specification projects. On the other hand, they cannot become targets of law suits and such.

Which event surprised you the most while you worked on this project?

Recently, when seeing GLOBALBASE, I often hear people say with surprise, "there is Google Earth in Japan as well?!" They mean no harm, I guess. However,I would like to point out that the correct expression is, in fact, "there is GLOBALBASE in the US as well?!" At one point I performed a demonstration of GLOBALBASE at a meeting of economists; I believe it was before Google Earth had appeared. One of the professors attending the demonstration remarked, "This is a wonderful technology. I believe it is based on US military technologies." He seemed to have gotten the idea in his head that Japanese would never be able to develop such technologies by themselves. At that moment I understood what a nation of surrender-monkeys we Japanese are.

When I created a system specification for a certain government office for system implementation, I was told that proper nouns should be left out wherever possible, because they would emphasize certain products and inhibit competition. However, the specification contained the phrase "an OS equivalent to Windows" throughout the entire document, and I pointed out that that phrase referred to a specific product name. I received a reply saying that Windows is the name of a foreign product and thus does not inhibit competition. In other words, the only names to be deleted were names of specifications and products developed by companies and organizations within Japan. I sometimes get fed up with the bizzare sense of equality among Japanese people, as well as our lack of self-respect and overestimation of foreign countries.

What is the most difficult aspect of this project?

We use multi-threading in our architecture and are experiencing many problems related to that.

Take, for example, the POSIX file descriptor. It is a very basic structure, which you should be learning during your first year of information technology studies. This concept, however, was developed in an era when multi-threading was practically unheard of and is thus not very compatible with threads. For example, let us consider the following case and assume there are 3 threads A, B and C.

1. Thread A reads file descriptor 1.

2. Thread B closes file descriptor 1.

3. Thread C opens a new file and obtains a descriptor.

If the events occur in the order 2->3, and event 1 occurs in parallel while file descriptor 0 is in use, the descriptor assigned to thread C in event 3 is likely to be descriptor 1, according to the POSIX specification. If this is the case, however, it becomes difficult to say in advance whether file descriptor 1, which is read by thread A, refers to the file just closed by thread B or the file newly opened by thread C. Once released, the architecture should not allow a descriptor to be reused easily.

A similar phenomenon occurs in C++ during creation and release of objects using malloc - free. In multi-threaded contexts it becomes necessary to protect the allocated memory resources with locks/mutex constructions, which can be done but the method of protection is unnecessarily complicated.

We were also annoyed with many OS and library bugs related to threads. We did not experience many simple bugs, where certain functions simply did not work as specified; rather, many of the bugs we experienced manifested themselves under certain combinations of conditions, such as "when there are threads waiting for reception of the connect instruction, and another thread forks, the entire process locks up." Moreover, there were many problems of a random nature, network fluctuations, and other infrequently occurring problems that puzzled us.

Also, our debugging environment presented us with a challenge. There are no functions available to stop the debugger at a certain point and allowing you to search within the stack for a function called by a certain thread by the function name, you see. If you know any debuggers that allow this, please let me know. It is quite troublesome to display and visually inspect each stack one by one, you know.

With gdb, it is possible to display all stacks with "thread apply all bt" etc. to search for function names with the terminal search function, although doing so takes some time, but in a general environment, we often get stuck. Imperfect GUIs for these debugging tools are problematic, I figure.

Why did this project begin to be recognized widely?

I believe that there is a certain type of Shangri-la (lotus land) in the direction the history of computer technology development is headed for. It may not be an image of "lotus land" originated in China. Route 66, Buston, UCLA, student activism, hitch-hike and similar keywords may be more specific. In the US way, it should be "Hill of Tara." Unfortunately, I could not appreciate such culture in real time, but I believe various computer science concepts born from such culture, such as "free software," "open source," "distribution technologies" and "public key cryptosystem," were born necessarily. People in Japan tend to misunderstand that such technologies were created by authorities. None of them, however, could not be developed without the power of stray scientists.

Maybe you might say that computer science was born in the US. But the "Hill of Tara" can be shared throughout the world without words. Linux, WWW and other software systems are not something created by the US alone. People involved are aiming for the "Hill of Tara" through the way they formulate the organization of their development and concepts of architecture.

When we take snapshots of a given moment, it seems to me that various technologies are appearing and then disappearing pointlessly. Some say that technological innovation is significant and computer technologies shift as rapidly as a chameleon changes its colors. However, this is clearly a very limited view of computer science. You can understand this immediately when you consider basic technologies such as TCP/IP and POSIX. Such basic technologies take several decades to mature. People who feel computer science is as changeable as a chameleon are simply not on the road leading to the Hill of Tara or fairyland.

Overall, I feel that the general trend of most roadmaps seems to be that centralization is somehow avoided and the creative energy is being poured into progressive distribution of technologies. At the end the day, it may be that having the energy to code constantly is the fundamental secret of taking initiative in a project.

What is the future direction of the project?

First of all, I want to complete the major release of version B within this year. In addition to the basic map browsing functions, we have implemented a search function for names of features and places in published maps in this version. It also supports plot data, altitude (DEM) data, mapping (coordinate conversion) data, as well as large scale (matrix) data types in addition to the already supported raster and large scale vector image data. In this version, map data on the terabyte-scale can be browsed smoothly.

Moreover, in a joint study with Digital Earth Technology (Figure 8), we would like to put the 1-meter distribution system for aerial photos and maps into practical use within the next year. If this is actualized, the latest versions of Google Earth-class images can be supplied anytime. A demonstration of the joint study will be given at Innovation Japan 2008 held from September 14. Please visit our booth.

p2_thumb.png

Figure 8. Digital Earth Technology's 1 m mesh aerial photo (joint research) (→Display with original size)

Moreover, we are considering putting our map editor into practical application, as I will explain in the next section. Furthermore, we are considering to start development of version C. With this version, it will become possible to handle 3D data through the introduction of support for multi-dimensional spatial coordinates.

We also have certain problems related to standardization of the GLOBALBASE protocol/architecture that must be dealt with. Currently, the Ministry of Economy, Trade and Industry is trying to standardize various distributed geographical system protocols into JIS and ISO under the names of gXML and Place Identifier (PI) (see Database Promotion Center). I am involved in these standardization tasks, but the current versions of the standards are very rough. I feel that standardization of functionality equivalent to that of GLOBALBASE will need to wait for some time.

Lastly, I would like to mention our diffusion activities. Of course we work hard to diffuse GLOBALBASE within Japan, but our real goal is to see GLOBALBASE used not only in Japan, but in the rest of the world as well. In order to achieve this goal, we intend to translate manuals and other materials within this year. Moreover, we would like to become properly established on sourceforge.net. Even though the official hosting is moved to sf.net, all the development activities are actually still handled within sourceforge.japan.

What requests do you receive from the users?

We receive many requests from our users stating that they want editors and they want to be able to handle 3D. Both of these require significant efforts, however. Also, if we rush out and create an editor without thinking carefully about our users, GLOBALBASE as a whole may grow unnecessarily complex, like some of the GIS software already existing on the market; therefore, at the moment I would actully rather not release any official editor. I feel that there is a need to keep the browser separated from the accompanying authoring tools in a sensible manner.

We have various ideas involving 3D, on the other hand. When the term "map" is replaced by the more advanced notion of "space," GLOBALBASE naturally evolves into 3D to become a "Web of space." We humans are equipped with a very basic, natural cognition of "space," which can be said to be a human archetype. Even the ability to speak using "words" and exchange ideas via documents of various kinds, another basic human form of understanding, develops later than the cognition of space.

This, I believe, will eventually lead to WWW (i.e., the document Web) to become transformed into a spatial Web. In other words, as various technologies and social conditions continue to improve, an even more drastic shift will occur to the Web than what we have aready witnessed for the semantic Web.

GLOBALBASE rendered in 3D will allow using various different coordinate systems compared to 2D GLOBALBASE. Even now, it is possible to render a Moebius strip, a Klein bottle and other similar 2-dimensional non-Euclidean geometries. I believe that there are no other GISs today that can achieve such a feat. With 3D GLOBALBASE, it will be naturally possible to handle Minkowski and other non-Euclidean spaces. Thus, GLOBALBASE will allow simulating how the cosmos would look like if we were able to move at the speed of light.

One of the tough issues in this context is quantum-scale coordinate systems. Since being able to view something in GLOBALBASE is equivalent to it being observable, a view through an electronic microscope is just another view. The final goal of GLOBALBASE is to be able to simulate all natural phenomena. That is the goal that I will try to push GLOBALBASE toward, inch by inch (although I have no idea when that goal will be reached).

What are the features of the software you are proud of?

Above all, we are very proud that we managed to implement the basic functionality, delivery of spatial information via the internet, at this early stage, and that GLOBALBASE employs the most advanced technologies in the world today.

I used the term autonomous distribution above. This technology is equivalent to current technologies such as Web2.0, P2P and similar. When we released the first GLOBALBASE version in 1999, we implemented a function for exchanging information among maps and maintaining links via networks. This is the same as the backtracking function we know today. Similarly, the function for exchanging information about a maintained link and resolving overlapping of maps involves all sorts of routing and PSP technologies.

What points of the project do you want to re-do, if possible?

What I have become keenly aware of is the importance of daring to make a drastic work-over when the amount of brushing up required threatens to grow out of hand. Keeping on making additions in an ad-hoc manner eventually generates too many distortions, eventually preventing you from progressing altogether.

I often discard entire programs, even after writing thousands of lines. The sad fact is that we can only identify problems after actually trying out the software. I would say that code does not become really good until you have rewritten it at least three times.

How do you coordinate the project?

Basically, we coordinate the project by exchanging ideas, opinions etc. via the mailing list and emails. Since each of us more or less works on independent tasks, we tend to use emails more often than the mailing list.

We work by forking separate CVS branches, identified by an individual tag based on the main developer's name, and when a task is basically completed, we merge the branch back into the main source code tree. We thought about migrating to Subversion, but we have simply not been able to find the time yet... We still use CVS. We also often use the primitive method of meeting face to face. That way we are often able to solve a problem on the first attempt.

In coordinating a Japanese open source project, I have become very aware of the significance of language barriers. Some of our developers are Malaysian and Chinese, and they tend to have a hard time communicating directly with the Japanese developers. After all, it is hard for non-Japanese developers to join Japanese communities. On the other hand, if English were used as the common language, the sad fact is that the Japanese developers would fall back. I suppose we are very conservative, which becomes a bottleneck.

Are you a dedicated developer of this project, or do you have another job?

I am a dedicated developer of this project. I am an associate professor of a university, and I am fortunate enough to have GLOBALBASE as my research theme. The idea of GLOBALBASE was born and nurtured in my previous work places and, in my current workplace as well, I recognize that my mission is research and archiving using GLOBALBASE.

At this point I would like to make one thing clear, though. The department I currently work for is the Department of Geography, Osaka City University, which might lead you to think that I am a researcher in the field of humanities; however, I am 100% a researcher in the mathematics and computer science fields. For example, the title of my dissertation is "Visual Dataflow Language for Signal Processing" and the title of my master's thesis is "Hardware Description Language for MPU Development." I used to study hardware, rather than software, including signal processing, data flow languages and hardware descriptive languages. However, you might have heard the phrase "paper machine." Hardware does not function without applications. When I thought about this fact while I was a doctoral student, I started to think about the importance of the concept of "future applications."

It is very difficult to answer the question, "what are the future applications?" One thing is certain, though. One cannot answer the question unless one knows the world surrounding computer science inside out. In the world of computer science, there are researchers who are not able to go out into the outside world, no matter how hard they try. Such researchers will find themselves going around in circles without going anywhere, ending up with "software that provides methods for selling this software." I, on the other hand, actively tried to involve myself in projects in other fields. In this, I was strongly influenced by my teacher at that time, Mr. Ken Sakamura. However, we had different ideas on research policies. After I worked for him as an assistant for 3 and a half year, I deceided to go my own way.

What is your development environment?

My development environments are the Hankyu Kyoto Line and the Midosuji Line, the Osaka Loop Line, the Hanwa Line, sometimes various Super Bullet Trains, the Yamate Line, the Tsukuba Express, the Tobu Tojo Line, various airplanes... just kidding. Due to the nature of my job, I move around frequently. As a natural consequence, I need a framework that allows me to work on development anytime, anywhere. For example, I started to write the manuscript for this interview on my way home, just after leaving Ohyamazaki station, where Suntory is located.

I always carry a 17" MacBook Pro and an AirH"Pro with me, although it may be annoying for people around me. I check the functionality out via the AirH" and then debug, compile and commit. Since Darwin is a BSD system, my machine allows me performing most of the debugging using POSIX tools. Also, it can of course handle debugging of the Carbon user interface. When an application works mostly OK, I open up a virtual machine and debug in Windows. Debugging of some network applications can also be handled using the AirH". Recently, EMOBILE was released, which I intend to try out as well.

I have access to Linux, FreeBSD and Solaris machines at my university. I check the behavior of codes that cannot be checked standalone at the university when I do not move around. I place great importance on "portability."

I handle most of my business contacts using my MacBook Pro and seldom use mobile phones. The MacBook Pro also provides the "Sha-Mail" service. Come to think of it, Mac users seem to be the mobilers of yesteryear. When was the style of going out with a cute Mac Plus in its special carrying bag on one's shoulder popular? It has at least become last year's fashion, it seems.

The machines at the university can be accessed via networks anytime. Project members are using these machines as compiler farms and so on.

Version History:

How can I contribute to this project?

You can contribute to this project in many ways. First of all, using the software itself is a big contribution in and of itself. Starting from there, please try looking for maps in your imminent environment; you may find maps without copyright. Alternatively, you may try drawing your own maps of your surroundings based on various existing maps. The next step is thinking about uploading such maps on a GLOBALBASE server. If you have any problems, please do not hesitate to contact us. We are happy to assist you.

If you are a person who cannot be satisfied only by simple maps and own or want to provide huge satellite photos or similar, we welcome you; please contact us by all means.

The development of GLOBALBASE is a very diverse undertaking. There are of course areas that require technical skils and knowledge, but there are also non-technical tasks where anyone can make contributions in various ways. The role of tester, for example, where you try to run a finished application and report on the results, is one way of contributing. There are also people who create various sample data, which requires knowledge on creation of GLOBALBASE data. Such people do not have to be computer specialists, either. Also, we need people that can design user interfaces for COSMOS and HTTP-GATEWAY. If you are good at designing icons and similar, please join us.

While on the subject, I would like to take the opportunity to briefly introduce some of the people who have contributed to the project. Mr. Tomoki Sekiyama joined the project a long time ago when he was a student working after school, and is still working on a volunteer basis on GUI and GPS adapters, even after he found himself a steady job. Mr. Oh Watanabe and Mr. Tomohito Nakajima from Zeta Co. have been handling development of Windows-specific areas as well as HTTP-GATEWAY. Mr. Toshifumi Saito from Saito Office has been dealing with designs and advertisement and PR activities. Mr. Koichi Watanabe and Mr. Hisanori Aoki from WEBMAX took part in data creation and design. Mr. Keisuke Yotsui from Community & Research Assist Network Co., Ltd. and many other people contributed to the project in various ways. I would like to take this opportunity to express my sincere gratitude and thank them all. Some people are professionals, some are not, and some people work for companies; yet, each of these people has brought his/her own special knowledge and skills to the table, volunteering to participate in the project and providing great help.

If you are interested in joining the project, please take a look at the GLOBALBASE Guidebook. If you cannot find what you are looking for in the guidebook, you are always welcome to email me.

If you want to become involved in the development in any particular way, please contact me; let's work together.

Please send us your requests to SourceForge.jp.

sf.jp is also providing us with great assistance, for which I am very grateful. At this moment, provision of development environment, release processing and many other tasks would be entirely impossible without sf.jp.

My only request would be an environment where multiple projects could cooperate together, rather than having each project use sf.jp independently, in order to diffuse open source. That is, although open source has started to attract attention, it is difficult to approach for general people who lack professional skills or are computer beginners. There is a myriad of interesting and convenient software projects other than GLOBALBASE on sf.jp, but many of these may be sleeping without being recognized by the general public.

In particular, the barrier of intallation seems to be hard to overcome for beginners. Look at projects like Fink and Darwin Ports (both are Mac OS X OSS package suites) and NSUG(Solaris OSS package suites for Japanese Sun user groups), I wish there were a project for creating a suite of sf.jp package tools that would allow the average user to "simply get this software and install anything." Let me temporarily refer to this suite as "sf.jp.port."

Ideally, the use of sf.jp.port could be promoted by approaching OS manufacturers and requesting them to sell the software bundled with the OS, or approaching hardware manufacturers to request that they sell PCs with sf.jp.port already installed. What do you think?

(Interviewed on: July 23, 2007)