<TITLE: Conceptual Modelling Languages
ACADEMIC DOMAIN: technology
DISCIPLINE: information technology
EVENT TYPE: doctoral defence presentation
FILE ID: UDEFP040
NOTES: continued in UDEFD040

RECORDING DURATION: 15 min 53 sec

RECORDING DATE: 22.3.2004

NUMBER OF PARTICIPANTS: unknown

NUMBER OF SPEAKERS: 1

S1: NATIVE-SPEAKER STATUS: Finnish; ACADEMIC ROLE: research student; GENDER: male; AGE: 31-50>


<S1> <START MISSING> and that the actual model that you er create the system with er they are difficult my main sources of inspiration are the following ones the other one is a report by the I-S-O technical committee called concepts and terminology for the conceptual schema and the information base and the other one is an article by professor kangassalo COMIC a system and methodology for conceptual modelling and information construction erm conceptual modelling and ins- information construction indeed the acronym er COMIC comes from there , and one can say that the methods of er my thesis is more like literature survey however it's some constructive parts like algorithms for translations and stuff like that , the main question i would s- say of the whole thing is what is the role of the conceptual schema and the different chapters of er the thesis er shed light to it from different angles but er from the sources from the I-S-O source they say the conceptual schema is at the conceptual level but it must be expressed using this conceptual schema language and it must be based on principles of logic so they emphasise er formalism and once we have a formalism er we can evaluate the computability and other formal issues of the presentation whereas COMIC has different approach over there the conceptual schema is based on intensions of concept- the conceptual level only and it is more or less independent of the pre-representation issue so any pre-representation is er i would say er rendering of the conceptual schema and that kind of a- approach makes the conceptual schema something very intuitive and conceptual but one could claim that it maybe doesn't pay enough attention to the question of semantics of the conceptual schema semantics that indeed is underlined with the I-S-O report with their so-called helsinki principle , however now on the other hand we would have the formal bases on the other hand we would have the intuitions of what what conceptual modelling is all about is there any way that we could er combine them , and the natural way er even though we remember that the re- representation issues are secondary with the COMIC approach the natural way of er presen- er presenting the conceptual schema er is the concept diagram that actually expresses concepts and intensional relations between them they are economical since there are basically only concepts and the intensional relations even though we are not always so sure what er what we mean by intensionality over there now if the concept diagrams would capture some features of other well-known technologies like the E-R the IFO anything like that that would be simply fun 'cause er we could express first of er er conceptual schema using the intuitive way and then translate it into something more formal , the other question that comes to fore if we use dec- description logics especially because they are interested in computational formalism if we can map into them is there any way that er we can then erm use logic as the basis of some manipulation of the schema , if we analyse this problem er in a more detailed level erm we come to the three following questions first of them if indeed the COMIC approach is right in in its er emphasis then there are probably some kind of languages er er we deal with the intensional er or conceptual level of modelling and erm of course we ask is there (xx) er the other question if erm these conceptual schemata are expressed what is the semantics that could be built in using them and the third question can be kind of er smuggle the semantics in by first translating these concept diagrams into other er for- formalism or presentation and here are some spoilers yes erm 'cause in the er in my manuscript i have er presented erm in joint papers with er <NAME> a language that is based actually completely on the purely intensional level and the background is in professor kauppi's er concept theory and this approach has been discussed in chapters er two four and five that are called intensional and extensional er modelling languages and four o- five functional approaches to kauppi's concept theory as for the semantics there are several different alternatives i discuss in chapter seven that is er that is semantics and conceptual modelling erm implicating the meaning of er concept diagrams over there i discuss the er possibilities of using possible-world semantics situation semantics and HIT semantics of (xx) and the translation question can we translate these COMIC diagrams that has been actually tried so many times er either direct mapping to databases or some other formalism erm there in er chapters two and six erm six er is er er concept D and IFO we er do some translations even i hav- er even though i have to say that the algorithm er is a bit rudimentary but er basically we can translate concept diagrams into something er (xx) sure has more formal erm semantics , but let's go back to the questions of intensional and extensional languages the working definitions of this theses is something extremely simple intensional modelling language would be a language where the intensional containment relationship is used between in the set of concepts whereas an extensional modelling language there the objects relations and attribute be considered then to be in the domain of application and we think they are represented more or less directly between the lang- the language and a hybrid language would feature both of them erm people who are have been interested in modelling would s- have the following questions first of all hasn't everything after the erm famous entity-relationship approach er been hybrid so that you could have intensional and extensional parts at will erm that is duly so and the other question in COMIC or concept diagrams we talk about intensional containment with some other languages we talk about attributes entities something like that isn't this more or less a terminological question if indeed we can point or we can pinpoint exactly the same thing in the er domain of application and that is as well the case however it is possible and i have kind of demonstrated it to construct a language where only concepts and intensional containment is used , and on the other hand it is possible to construct a language that has very explicitly extensional semantics if this link works i can show the language of er er concepts and intensional containment er over there the underlined curves concepts er the lines er (cutter) lines intensional containment , and what's what about the explicitly extensional semantics er something like this in chapter two we start with named sets doesn't sound very extensional and then we have named relations between those sets later we add some syntactic sugar to make language a bit more expe- er expressive but i would still think that it stays very much in on the extensional level , many people have seen things like this you can't avoid it er even the honourable opponent has been working er with things like this this is er end very end of er entity-relationship model where some things are very er well extensional for instance things having attributes like a er faculty having a (phone) and some things are very apparently erm i would say intensional for instance a member of faculty er or student being a person so this would be kind of a well-known example of a hybrid language . er we saw the first example of the er er intensional modelling language we saw that er the diagram was simple 'cause there were only those concepts and intensional containment but we can duly ask what the intensional containment would mean in these er concept diagrams and the first one that er many er from the philosophical point of view is completely (radical) they say it means nothing er it's just the basic relationship like in for instance professor kauppi's concept theory , now from the er computer science point of view we can say that hey then we can't do anything 'cause after all this is about data data and database have semantics so er where does the semantics come from . whereas the other possibility would be we simply map them something that looks very much like the E-E-R that we saw over there for instance something means exactly the same as something being an attribute of something else some of these lines mean something being in a relationship with other entity and something is to (build an) super er concept sub concept relationship as over there . erm unluckily this would kind of mean that the intensional containment arches in the original er schema like the concept er diagram they would be illustrations they wouldn't mean anything erm and to get them the actual meaning we would need an translation as in my thesis i map them to IFO and this might well be the case erm at least er you can easily fill tens of pages with that kind of thing but on the other hand in chapter seven we consider the er semantical er alternative to it for instance er mapping something into er semantics or explaining something using semantics that would duly give it some kind of a er logical interpretation , and the er possibilities or different alternative in semantics are those of possible worlds that has a kind of a minor glitch that we might er come back later , situation semantics that is an extensional possible world semantics then theories of predication and HIT semantics of (prague) where we kind of er next (thesis) is slightly technical but this won't last long anymore over there erm the main thesis of HIT model is that concepts are meanings of expressions and constructions are methods of reaching them there is an epistemic base there are methods of building combinations out of the er epistemic base there are things called constructions , constructions are simply for instance the function that would have arguments and the er way you build it up is the construction and then there are sub constructions for instance the arguments of the said function when applied would be sub constructions so the sub construction relationship actually looks very interesting doesn't it erm we might even be tempted to explain all of the intensional containment using the sub construction relationship and er that is indeed question on the next slide and chapter seven would could it be actually undest- understood as the intensional containment and er of course we say no erm jari palomki over there has demonstrated in his article that the sub construction relationship er is not the same as the intensional containment relationship in kauppi's theory on the other hand if we do not follow concept theoretical line but er something a bit er more con- er computer scientist line we could say probably yes that could very well a sub-construction relation could very well explain most of the things that we find in concept diagrams there are some additional details but i think we can cover them later , and as the er summary of this presentation and of the thesis in general the following things could be considered fo- first of all the role of the er conceptual schema what is so intensional or conceptual in it i hope that i have indicated some of it by using the categorisation of different languages then about concept theory that was er important in two chapters like the functional presentation of kauppi's theory erm we have er demonstrated that er it is good as background and then help formulating er the er background and finding some missing concepts for instance but it might not always be useful and about the concept diagrams we can see that in many cases they contain enough information to make a feasible database implementation but if we are really interested in the different variants of the intensional containment we either need the mapping or then we have to use another approach as for the concept theory that will be covered later . list of errors the main error i omitted one important contributor in the acknowledgement and er now doctor professor <NAME S2> as my opponent i'd like to ask you to provide comments that you find pertinent </S1>
