a journal of a researcher

Wednesday, November 17, 2004

My new web site

I registered a new web site. http://www.flydragontech.com/ is new web site for Yuhong Yan to host my personal web site and serve as the testing field of my projects. I do this is because it is too complex to use NRC’s resources. You need to compliant to the regulations, and pass through the technical staffs. I pay for this web site hosting service (actually it is yahoo small business hosting service) is because I need to bring my research results faster to the real world. Do what I want to do.

IST 2004

IST 2004 runs from 15-17 November, in the Hague, Netherlands. This year's edition of the most important European Information Society Technologies (IST) event includes a conference, an exhibition of research results and networking facilities for the participants, and should appeal to industry, SMI and SME, businesses, public sector decision-makers, the research community and academics, and anyone generally interested in the Information Society, including young people and students.

I have no time to write full text. Here are the points of my slides to introduce IST to my colleagues.

IST: The big picture
• Citizen centric: learning, healthcare, government, quality of life
• Information and Knowledge society
• Service oriented industry
• Globalization: less difference in economics, global competition, mobility
• Themes
– E-learning
– E-government
– E-health
– E-inclusion
– E-collaboration

Information Technology Trends
• Knowledge management
– Semantics, ontology
– Nature language processing
– Information retrieval
• Ambient intelligence
• Open source and Service oriented software industry
• Grid computing
• Telecommunication (beyond G3)
• Security

Networking: Canada-EU
• Networking session on Nov. 15
• Reception hosted by the Canadian Embassy in Netherlands on Nov. 16 evening
• Canadian delegation booth
• Results:
– Many, many interests, more than my expectation
– East European countries: Hungary, Czech, Slovakia,
– EU10 countries: Germany, UK, Netherlands
– Canadian: Athabasca University, Richard Hornsby in UNB
– Forward you information
• Digital Olympics
– News and reports
– Resource management
– Broadcasting, traffic management, security
• high performance computing
• open source
• semiconductor manufacturing.
• Serve the global market
– Software engineering
– Product engineering
– Accounting
– Call center
– Technical supporting
– Heath care services
• Collaboration
– Joint R&D projects
– E-government
– Education: alumni ecosystem for 100 European schools

Saturday, November 13, 2004

Where to go? Web Service vs. Grid Services (3) - My visions

1) Grid Services so far are not matured enough to integrate with Web Service specification. But it is a necessary step for them to go. So you can see a lot of efforts on-going.

2) Grid Services and Web Services will merge on WSRF/GT4. So do not spend time on GT3. Most of them will go away. Try WSRF directly.

3) The benefit to use grid is to improve the computational performance, using clustering computers, workload balance, job submission etc. These are high performance issues. The e-business servers, like websphere, weblogic event do not have these functions. I read that websphere will have these features soon, but it will be very expensive. So we have to survive with the servers without these features. You can say we are in a prehistorical era of computer science by seeing this. Well, it is true.

4) Besides that, the features like states, notification, logging, lifecycle management are not the motivations to use Grid Services. It is because we can implement them easily using exisiting tools and will not suffer from prematured grid services.

5) High performance computation is more for traditionally time sharing OS like Unix and Linux. So you can see that most of the tools are for Unix and Linux. This will be a barrier for e-business applications.

6) Actually it is not sure that using WSRF will solve the high performance issues. So far I see it is only a set of protocol, no real workload management tools or other tools.

7) It depends where you want to conduct your research. For me, high performance computing is another domain. I will not study workload balance or clustering algorithms. So I would see a service will be wrapped into a web services, and study the Web Services issues. Actually I see the QoS issues are very important. That is a joint point of the two domains. I will also focus on the process management issues and portal issues, because these are the joint points of e-business. I can share the knowledge in my other projects.

I do not say we have to avoid Grid Services. Actually I will try WSRF. I just say my research will not go beyond using WSRF on the Grid Services.

Where to go? Web Service vs. Grid Services (2) - Merge

Actually people in the Grid Services are active to make the Grid integrated with Web Service specifications seamlessly. Maybe WSRF will be better than what I claimed above. Here are my visions:

1)the Grid Services will wrap the service as web services, so pay more attention to wrap your services.
2)The benefit to use grid is to improve the computational performance, using clustering computers, workload balance, job submission etc.
3)Besides that, the features like states, notification, logging, lifecycle management are not the motivations to use Grid Services.
4)High performance computation is more for platform like Unix and Linux, because these OS are traditionally time sharing OS. So you can see that most of the tools are for Unix and Linux.
5)Try GT4 and WSRF from now
6)Follow the new specifications of Web Services
7)The gap between the two will be less and less.

Where to go? Web Service vs. Grid Services (1) - Critics on Grid Services

Grid Services is a combination of Grid + Web Services. But it is in a pre-mature status. The specification so far does not present anything attractive to the Web Services community. Here are my critics:

1) The specification is not stable. WSRF will replace OGSI in Jan, 2005. At that time GWSDL will go back to WSDL1.1, which does not support the portType defined for Grid, and service data (SDE) will go too.

2) it still relies on external solutions to resource management (it has no scheduler). I saw the listed tools, such as GridPort and OGCE; Grid Packaging Toolkit (GPT); MPICH-G2 (Grid Enabled MPI); Network Weather Service (NWS) (Quality-of-Service monitoring and statistics) ; Condor (CPU Cycle Scavenging) and Condor-G (Job Submission) . If without these tools, Globus has nothing left, or has only web services left.

3) the features like state, notification, can be easily implemented by using web services + EJB or other tools. I do not think these are the points to use Grid Service.

4) Globus tutorial says that Globus does not integrate well with existing tools.

5) Most of Globus modules are tested on Linux or Unix only.

Sunday, November 07, 2004

e-learning conferences and events

13th European Conference and Specialist Trad Fair for Educational and Information Technology: learntec 15-18 Feb, 2005, Karlsruhe, Germany

European E-learning Award: eureleA

10th Int. Conf. on Technology Supported Learning & Training Dec, 1-3, 2004, Berlin, Germany

ED-MEDIAJune 2-July 27, 2005, Montreal, Canada (deadline Dec. 12, 2004)

iceer2005 March, 2005. Taiwan. (Submitted)

ACM Conference on Electronic Commerce (EC’05) June 5-8, 2005, Vancouver, Canada (deadline Dec. 7, 2004)

An interesting discussion with a German publisher

I met Dr. Jurgen Schmidt from Ernst Klett Verlag GmbH in prolearn meeting. I asked him if he thinks e-learning will change the publication. He sees this trend, but he does not expect it to happen soon. Two obstacles, for publisher, publishing in multi-media greatly increases the cost of publication, because more workers are needed to work on media part, such as sound, image, picture and animation. If this era does come, the publisher will change greatly. Second, so far the teachers are not convinced about the effects of new technology. The young generations are still taught on paper. So he expects that e-learning will not completely replace traditional learning methods in the next generation. He greatly concerns the quality of the learning material in new media. It seems to him that the quality of the learning material in new media is not easy to achieve.

From Cyberlab’s business model to education system

Cyberlab is a portal of online experiment system. Its business model is to provide administration service to existing online experiment systems, and charge the service providers. This model obviously does not work, which is confirmed by Eikaas Tor Ivar, the creator of the system. He said the universities who provide the services simply do not want to pay the services, because they believe they can do the same. What I saw is that the European education system eliminates the possibility. Simple answer is that the university education is free in Europe. If the university provides free services, where is the motivation to improve its services? To get more students, and even remote students? Why should the universities be bothered to do so? The consequences are: the good technology is not used, no business is created, no job is created. Everyone waits the government to pay to improve education.

Because we are in North America, we simply reverse the business model: the students will pay for the online experiment services. By using online experiment system, the students can save tuition; the universities can provide the education to more students. And the universities are willing to share their resources if they are got paid. It seems it is a workable system in North America. The consequences: new business is created, jobs are created, and students are got trained better. Therefore the society is better.

The European profs simply claim that the education should be free. But universities are not yet another welfare system, nor are they the pool for the students to wait for 21. The effectiveness of the education system is a main concern. I saw much more undisciplined students in European universities than in North American universities. The students in North American universities do not come to waste their money and time. They demand well training so that they can get a job to pay back the tuition. The profs have the obligations to provide quality services and are obliged to reform the courses to meet new demands. Thus it is not the profs demand the students, but the students demand the profs when considering teaching. I do not see in North America, the students can’t go to university because of money. On the contrary, the North American universities have much more scholarships to give.

I should claim that even in North America, universities are non-profit organizations. Tax money is invested in university.

e-learning in Europe

In Hanover, I took part in two Prolearn workshops: technology enhanced learning and online experiments. Both were very interesting and I had very interesting discussions with many people.

EducaNext impresses me. It is a University Brokerage Platform to manage the learning resources and foster collaboration. You can register your course material on this platform. I plan to register my course on it. It will provide more functions on community building and collaborative learning. The Vienna University of Economics and Business Administration spends half of a million euro each year on this platform for the content and management.

The learning management system (LMS) is a commonly used concept. WebCT is a typical LMS which has the abilities to manage users, manage content, collaborative tools like discussion forums. I’ve done some survey on the existing open source LMS, such as Moodle, ATutor. Most of them use php and MySQL. In this conference I knew another one, Stud.IP , which seems superior to the existing LMS. The interesting functions that the Stud.IP provide: literature management, survey and test, search, wiki web, export functions, scheduler for institute, timetable for students, online community support (personal web pages, who-is-online, chat room, discussion forum).

How to foster the collaborative learning and tools to support collaborative learning both for tutors and students are important topics. I think we need profound research and solid theories on this direction. I saw the work of 5D project – learning beyond 3 D and time, i.e. learning anytime, anywhere. In 5D, a maze and learning task cards are designed to guide the students on a learning project. But I want to see more intelligent tutor, or customized learning project built on solid theories.

Online experiment system is a mixture of virtual laboratories (simulation) and tele-laboratories (remote laboratories). One interesting project shows how a real laboratory can connect to the simulated virtual devices (DERVIE project). People also pay attention to enable team work in a distributed environment. We presented to use grid technology and web services to compose online experiment services, which will integrate learning resources in different locations. We think it is the trend, because that provides more opportunities to learn remotely, using distributed resources. Cyberlab is doing a portal of existing online lab. We present an opposite business model to Cyberlab’s. More detail is in another post.

A la Maison de Paris

I am in Paris again. The city is on its regular rhythm. Pied de Cochon is still full of people day and night. Its seafood counter is facing the street. You can see the chefs open the shells and prepare the big fruit de meer platters for the well dressed people in the hall. Bisto Roman is besides Pied de Cochon, a successful restaurant chain. Then, several 7/24 bars. People are walking, eating, and drinking. Out of Fredericton, I feel the contrast of Paris. Montougei’s little shops are so different than the spacious super store in Fredericton. You can’t image how the little shops can survive in the expensive area and hire so many people. Well, you can find the best quality food here and a lot of varieties -- Jambon cur from Italy and Span, fruits and vegetables from Africa and South America.

Galeries Lafayette: decoration below glass dome (1910-1912, architect Ferdinand Chanut) (from http://www.cambridge2000.com/gallery/html/P40214377.html)

Galeries Lafayette is preparing for Christmas season. The shopping windows are dressed like fairy tale world. The animated animals and puppets attract the children and parents. Each time I come to Galeries Lafayette, I see more Chinese, the Chinese from main land China to visit Paris. They like to spend money in Galeries Lafayette and Printemps, the two names they know in China. The first floor is their best love. They like purses from the best brands, like Lancell, Gucci, Luis Vinton, and Christian Dior. They buy these brands so much that all these stores hire Chinese speaking sales. They also like Parisian perfumes and jewelries. Clothes are second to the Accessories. From this year, Lafayette has regularly greetings in Chinese.

Galeries Lafayette is a place for Chinese to bring Paris to their home. In Paris, there is another place where the Chinese bring China to Paris. The newly emerged Chinese district is near Metro Art et Metiers, where the Chinese from Wenzhou sell their products of purses, accessories and small electronic applicants.

Thursday, November 04, 2004

Manage Open Source Projects

Some literatures that not out of your imagination:

three tips to survive open source projects

choose open source

Commercialize Open Source Software

In the prolearn project meeting, the Europeans demonstrate their effort in e-learning domain. Many choose to make their research results open source. Open source is of growing interest to the education and training sectors for a number of reasons: Low cost solutions for effective learning provision; loose “community” of educational institutes; benefits from collaboration. I am impressed by the number of the high quality software provided by this community.

Though for research work, it can end up at the open source software, I want to see the opportunities to commercialize open source. Globus, an open source implementation of OGSA (Open Grid Service Architecture) is under commercialization by Platform, a Toronto based company. I think there are two ways to commercialize open source software (OSS): providing linkers to existing products or providing commercial support of the open source.

Don Rosenberg of Stromian Technologies (http://www.stromian.com) writes a monthly column for businesses. Many interesting points:

. Government funded projects and academic projects are the source for OSS.
. The future of software industry is the service, may the software is free.

Wednesday, November 03, 2004

Fostering Collaboration between Canada and Europe

I am on my trip in Europe. The first thing is to take part in the Prolearn project meeting (e-learning). The second thing is to attend IST2004 for my institute.

Europe has world-class researchers and research project. The EU Framework program efficiently brings the researchers from different countries together. It is a good tool to build a united Europe.

Many of my research topics still results from my working in Europe. I gained a lot of my research ability by working with the best researchers there. So from a researcher point of view, collaborating with Europe is valuable. I hope my European counter parts think the same. I think they do, because I helped them to prepare FP6 calls.

What I do not know is from the country level. It seems Canada is a gateway to US; or some policy, such as privacy law, is similar between Canada and Europe than between US and Europe. These are Larry Kobra’s points of view. But I am wondering why there is no such a push between Canada and US, if US is so important and close to Canada. If I want to do my mission better, I have to know more about Canadian society and also the European society.