Lyee Internet Information Interview with Stuart Toole at University of Central England, England Domenico Pisanelli and Aldo Gangemi at ITBM-CNR, Italy Stuart Domenico Aldo Toole Pisanelli Gangemi Date: Thursday, June 28, 2001 Place: The Institute of Computer Based Software Methodology and Technology Moderator: Hamid Fujita, professor at the Iwate Prefectrual University [Stuart Toole] Born in 1950 1970 Licentiate of the Inst. of Rubber & Plastics Industries (John Dolton College of HE) 1972 Associate of the Inst. of Rubber & Plastics Industries (John Dolton College of HE) 1977 Post Graduate DSW (University of Leicester) 1
1999-Present Senior Lecturer in Computing University of Central England 1994-Present Director ST Solutions Ltd. 1984-Present Editor Information Technology in the Human Services 1986-Present Member of the HUSITA board Human Services Information Technology Association <Research Interest> Clinical & medical field, computers in human services, etc. <Selected Publication> "Expert Systems in Social Work" (1986) "Computers and Social Work Key Issues"(1987) "Expert Systems and Social Work"(1987) [Domenico Pisanelli] 1959 Born in Rome, Italy 1986 Laurea degree in Electronic Engineering (Computer Science in Medicine) at the University of Rome 1987-Present Researcher at ITBM-CNR (National Research Council, Institute of Biomedical Technologies) <Research Interest> Medical Informatics (expert systems in medicine, computerized guidelines, and telemedicine) [Aldo Gangemi] 1962 Born in Rome, Italy 1989 Laurea degree in Philosophy (Analysis of Scientific Languages) at the University of Rome 1995-Present Researcher at ITBM-CNR (National Research Council, Institute of Biomedical Technologies) <Research Interest> Formal ontologies and methodologies for information integration 2
Prof. Fujita Today, we have three colleagues: two persons from CNR in Rome and one from University of Central England. The purpose of their research is related to medical application, medical ontology. Their visit this time is for collaboration for research project we have for the 21 st century software development. They came just last week. Today, we have the interview as a part of their preparation to contribute to this research project. You have been discussing new methodology, which is Lyee. I think I already talked with Domenico about my visit to Rome. I talked about my intention to establish international collaboration for new software development for new century. As you have been using conventional software development, I think you have some feeling what software development in a conventional frame is and you have studied for last few days about Lyee methodology. So, I think you have a vision about that world and this world. This is one question. You can have projection about this comparison. Mr. Toole My expertise is in developing self-care computer programs for people who have psychological problems. I was interested in Lyee because we have used non-traditional approach for about twelve years now in designing our systems. We had some research two months ago which demonstrated that compared with SSADM and SS(soft systems), our current methodology was 50% more efficient on start of the project to get into specification. Our problem is** that we have not yet found a way of speeding up this programming and in particular the testing and re programming. This takes a large part of the total system development time. I think that Lyee will fit in nicely with our ideas. Here is something new that could potentially greatly improve our productivity. Mr. Pisanelli First, I feel quite confused by the theory and I think it is impossible to understand fully the theoretical explanation in a week. As far as implementation of the tool is concerned, we are quite impressed by the fact that traditional lifecycle of software is very different. With Lyee approach, it is non-traditional lifecycle any more. So we are used to different lifecycle and have many feedbacks between implementers and users. With Lyee, lifecycle is shortened and there is no feedback. Feedback is in every step and is inside. So, it is quite a revolutionary approach and very interesting. But in a 3
week, you just can get a feeling. Mr. Gangemi A week or three days is not so much to grasp Lyee but I felt it is very impressive and revolutionary for software development. I feel that good points are philosophical perspectives and it wants to shorten the part from the intention of users to programs. In this sense, it is promising to start from the intention expressed in natural language and visual codes. This is a good point. As for the detailed procedure, I feel the need to distinguish two different issues in this approach. First issue is what the input for requirement extraction is. The other issue is actual input for software design and implementation. As far as requirement extracted from the analysis is concerned, I would like to know more about actual process of extraction from natural language. I heard about using only nouns but which nouns? On the other hand, I wonder if there is no modeling. Actually, the release of requirement can be given in several ways. I have seen that requirements are given in words, diagrams, structures, etc. In any way, there is some modeling process. Maybe we can expect that there could be more modeling approach even within Lyee. I feel that in traditional requirement analysis, tools are built in ontology. If you have model-oriented approach, it enhances clarity. It has advantage for Lyee, not drawback. As for software design and implementation from ontological point, I can only say that Lyee approach seems to be very adequate ontological flowchart. As far as I understand a kind of universal structure to optimizing software development, Lyee is very promising. Prof. Fujita I would like you to talk about your research background and your impression making a possible connection with Lyee and your research related to medical terminology and applications. Mr. Toole One very interesting finding that has become more important over last few years has been the realization of how important it is to reflect the philosophy and human values held by the professions and the patients actially in the design and coding final system for them to be most effective. A seemingly small mis match can result in patients not using the system or in clinicians having no confidence in the system. The values of the human beings involved in the system design and testing :patients, doctors, and software 4
engineers are all important. Even programming languages can imposes certain values on the system because of how it is written. The Lyee approach offers a completely different perspectives. It iwill be interesting to see how it will affect our ability to program in one way or the other. I am sure it will affect things though I do not know how and it is quite an exciting thing. If you take a spreadsheet and bring an accountant from the year 1850, for example, and show him the spreadsheet, he will immediately see the value of this in two minutes. Ie it reflect the values of the accountancy profession. If you give a spreadsheet to doctors and ask them to model a medical system, they will not be comfortable with it. They have other values and those values cannot be expressed in a spreadsheet. Doctors want something with medical human values, not accountancy proffession values. I used this example to illustrate a point that program language in itself does impose restrictions on human values that can be expressed in a final system. Prof. Fujita From Lyee point of view, engineers are not conventional engineers. I think you saw only young staff in the factory. They do not have any background in computer or software engineering. They graduated from different departments and universities. So, when they come, they are educated what words are, how to make scenario function, etc. and they have source codes. They never care what formal design, specification, spreadsheet, and language are. They have different paradigm. Now, if this paradigm is given from doctors, engineers, etc., then, these guys can make software. So here is the difference. They need to classify knowledge for different users and different requirement engineers. Classification means to classify the knowledge and ontology. In order to bridge their world, we should make user-friendly interfaces, define a lot of concepts for bridging user interface into difficult ontology inside the system. I think that Lyee allows to take doctors requirements as they are and implement them. Mr. Toole That is almost correct. We will still have to spend a lot of time making sure that the screens which are produced reflect the requirements of all the users such as doctors, patients, and the software engineers. When those are produced well, then, given the Lyee approach, it is hoped that we will be able to more quickly 5
produce code in a way we want and also speed up testing process. We spend a disproportionate amount of time (compared to may other domains) testing selfcare programs because of the ethical problems relating to the risks of testing poorly designed and programmed software on real patients. Prof. Fujita I would like to ask the same question in a different prospect. We discussed ontology applied for very complicated requirements. How do you make ontology to requirements? This will be outside of Lyee but very important to Lyee. This is very interesting to Lyee. Requirements are conflicting. I would like you to comment on how you deal with requirements using ontology. Mr. Pisanelli Yesterday was very interesting because I just realized that not all requirements are equal. Till now, Lyee has been using requirements as input. If you just go thorough implementing, you realize software. The requirements are not all the possible words. We discovered this by the example. There were two different words. One was not good for Lyee input. It should be rephrased. I think it was a word describing action. They cannot be used as nouns. We could analyze the kind of words that can be admitted by Lyee and what words cannot be. Ontological approach might be interesting. Mr. Gangemi Maybe there is something that ontology can do to requirements but we have to study Lyee a little bit more. I am quite sure that when analysts try to read requirements and extract relevant words for Lyee approach, then, they have to find the definition. This morning, I asked about what they do if they do not find definition in the text and they said they would call users. In that case, they have to add words. If the text is very informal, they have to transform and normalize the text in a Lyee-compliant form. In some sense, they are doing modeling. So, this means to me that conceptual modeling can be in use for this work, especially for constructing what users wanted to explain in a language but actually did not emerge in the surface structure of the language. Prof. Fujita This project has started since June. We will have conferences. Next year will be in Toulouse and the year after next will be in Kyoto. I would like to hear your vision of the collaboration. 6
Mr. Toole First of all, I would like to say that I have already discussed this with the dean and faculty before coming here. They would be very interested in such collaboration. I am a committee member of the HUSITA(HUman Service Information Technology Applications)organisation which has its bi annual conference in September, which is a good timing. I would like to develop some exchange and learning and then I could share some ideas with the experts of the field. We have 98 papers and 50 of these are about system developments. So, it would be good if I could owe that with other people from human services and information technology associations to discuss these issues with the expert mind. And then, after that, we could start producing pilot project where we could take a new idea, for example eating disorders and work with Imperial College team and produce a pilot project putting as much Lyee as we can. And then, also using Lyee for programming to see if that will work. Mr. Pisanelli I think Lyee is an interesting approach and this can be a possible research project we might carry on, which is ontological analysis of Lyee requirements. Lyee itself contains ontology of flowchart, control mechanism, and structure. We might use our formalization analysis to analyze inside structure of Lyee. This could be a framework and one application of this framework is the analysis of ontological idea of requirements. So, the result of this could be that what kinds of input user can give to Lyee and this could shorten even more the lifecycle of software because the extraction of requirements would be shortened. Mr. Gangemi I think that is one of the possibilities of the research project. We make requirements for the system and guideline management tool, which is the general framework for describing medical procedures. This could be a way of using Lyee. Prof. Fujita From our perspective, we would like to bring a new academic foundation to software science. Software science through my career, especially software engineering, has a lot of conferences and we have a lot of panel discussions. But all of these have never established stable concrete foundation for software. Software is going through a very 7
critical stage. In the market, there are assessment companies and they say that the loss of the money is trillions of dollars because of software development. It is very terrible and critical trouble brought with. This scientific community is not made on purpose but it is made in this way. Software technology is going very fast and we should do something. Since the industry is in a hurry, academic people tried to establish some foundation. Speed is very fast and we are going very slowly, and technology cannot bridge that speed. I found some light in fact when I was acquainted with Lyee and this light can bring something, either spark or new very huge light. In any case, it can bring great results. This is my impression in general. I think that all colleagues are coming here for the same purpose. Everybody shares the same opinion and the same dream but objectives or the way of achieving objectives are different, but they are going into the same direction. Unauthorized reproduction of the contents hereof are strictly prohibited. All copyrights pertaining to this document are the property of Catena Corporation. Copyright (c)2001 CATENA CORPORATION 8