Formal Foundations for the Evolution of Hypermedia Systems Mª José Rodríguez-Fórtiz (Universidad de Granada) Lina García-Cabrera (Universidad de Jaén) José Parets-Llorca (Universidad de Granada) Good evening, My name is Lina García-Cabrera and am going to present the work “Toward a Formalisation of Evolutionary Hypermedia Systems based on System Theory ” by José Parets-Llorca from the University of Granada and myself from the University of Jaén (Spain)

Why Evolutionary Hypermedia? Dynamic and evolving nature of Hp Systems. Represents a conceptual domain Ways of representing, structuring and browsing it. Life-cycle- of Hp is not sufficiently considered. Implicit, diluted and unaffordable structuring process inside the documents. In the same way as other authors we consider that the “process of tying two items together is the important thing”. Models must provide an abstraction of functionality from structure instead of abstracting the hypermedia connectivity from information content. Therefore, the current information systems based on links between information chunks must be converted into real knowledge systems based on structured information items But this structuring process requires the identification of its components and their conceptual associations. So, we need cognitive models and hypermedia must be based on this cognitive models

A Semantic-Dynamic Model based on Systems Memorisation system: information semantics, conceptual domain. Navigation system: extends semantics, order relationships to cover the conceptual domain. And Why a cognitive model? In others words, why should the semantic structure of hypermedia systems be made explicit? Because this benefits the author and reader during the development and use activities: construction, maintenance and navigation. Authors design their hypermedia systems step by step in an incremental way Authors can maintain the hypermedia system if they have a representation of the decision process carried out during their construction And finally, the reader can navigate easily and obtain a better understanding if the tool provides this semantic structure meintenance Evolution Changes and updates integrated in the development process

Formalisms of a Higher Abstraction Level Meta-level Evolution operations (Ace) (meta-operations with preconditions or meta-restrictions) changes changes Memorisation System MS = (CS, RT) concepts, items, associations, restrictions Navigation System NS = (CSn , RTn, PN) navigation rules or restrictions Therefore, we believe that the hypermedia systems must be constructed by a Semantic-Dynamic Model (SDM). With the aim of clarifying the understanding of the model I have divided his explanation in three parts. First of all, I will describe briefly the architecture of the model, second I will define and describe the three most important components of the model: Information Items, Conceptual Structure and Restrictions. Finally, I will come back to the architecture in order to locate the different components inside the model and show their functionality. Arquitecture

Formalisms of a Higher Abstraction Level Meta-level Evolution operations (Ace) (meta-operations with preconditions or meta-restrictions) changes changes Memorisation System MS = (CS, RT) concepts, items, associations, restrictions Navigation System NS = (CSn , RTn, PN) navigation rules or restrictions Therefore, we believe that the hypermedia systems must be constructed by a Semantic-Dynamic Model (SDM). With the aim of clarifying the understanding of the model I have divided his explanation in three parts. First of all, I will describe briefly the architecture of the model, second I will define and describe the three most important components of the model: Information Items, Conceptual Structure and Restrictions. Finally, I will come back to the architecture in order to locate the different components inside the model and show their functionality. Arquitecture

Memorisation System Specification: Conceptual Structure CS (Graph) concept, C item, II association concepts, Ac From our point of view, a Hypermedia System can be conceived by three interrelated and interacting systems: the Knowledge System, the Navigation System and the Learning System. The Knowledge System is in charge of the storage, structuring and maintenance of information items. It stores the knowledge acquired about the information system. This knowledge will guide the design and structuring processes of the information system. It will determine the possibilities of transformation and change in this structure throughout its evolution procesis consivd association concept-item, Ai

Memorisation System Specification: Restrictions RTs & RTa (Temporal logic) RTs : “is_a association is not recursive” <c, is_a, c1>  not  <c1, is_a, c>  c, c1  C RTa: “every planet will rotate around the sun” <c, rotate, Sun>   <c, is_a, Planets>  c C Finally, The Learning System optimises the knowledge acquisition process from the hypermedia system, adapting navigation to the information needs and to the knowledge gained by the reader. óptimaisis

Formalisms of a Higher Abstraction Level Meta-level Evolution operations (Ace) (meta-operations with preconditions or meta-restrictions) changes changes Memorisation System MS = (CS, RT) concepts, items, associations, restrictions Navigation System NS = (CSn , RTn, PN) navigation rules or restrictions Therefore, we believe that the hypermedia systems must be constructed by a Semantic-Dynamic Model (SDM). With the aim of clarifying the understanding of the model I have divided his explanation in three parts. First of all, I will describe briefly the architecture of the model, second I will define and describe the three most important components of the model: Information Items, Conceptual Structure and Restrictions. Finally, I will come back to the architecture in order to locate the different components inside the model and show their functionality. Arquitecture

Memorisation System Evolution: Modification of CS (Set-theory) (Set-theory or Temporal logic) Ace: add_concept: Saturn Meta-RT: Saturn  C Ace: add_concept_assoc: <Saturn, is_a, Planets > Meta-RT: (set-theory) < Planets, is_a, Saturn >  Ac (temporal logic) <c, is_a, c1>  not  <c1, is_a, c> with c = Saturn and c1= Planets Internal propagation of the changes The Navigation System helps the reader in his interaction with the information system. Using the knowledge base and the reader activity over time dynamically, this system determines: –firstly- the accessible information and – secondly- its interaction possibilities. dainamicaly Evolving MS: CS  CS' CS' = (C', II', Ac', Ai') C'= C  {Saturn}; II' = II ; Ac'= Ac  {<Planets, is_a, Saturn>}; Ai' = Ai

Memorisation System Evolution: Modification of RTa (Predicate temporal logic) Ace: addRest (<c,is_a,Planets>  <c,rotate,Sun>) Meta-RT: “cycles in concept associations are not allowed” addRest(ac2, ac1) not  isRest(ac1,ac2) ac1,ac2 Ac This clause can be instantiated: addRest (<c, rotate, Sun>, <c, is_a, Planets>)  not  isRest(<c, is_a, Planets>, <c,rotate, Sun>) Now I will define and describe the three most important components of the model: Information Items, Conceptual Structure and Restrictions.

Memorisation System Specification: Restrictions RTs & RTa (Temporal logic) RTs : “is_a association is not recursive” <c, is_a, c1>  not  <c1, is_a, c>  c, c1  C RTa: “every planet will rotate around the sun” <c, rotate, Sun>   <c, is_a, Planets>  c C Finally, The Learning System optimises the knowledge acquisition process from the hypermedia system, adapting navigation to the information needs and to the knowledge gained by the reader. óptimaisis

Formalisms of a Higher Abstraction Level Meta-level Evolution operations (Ace) (meta-operations with preconditions or meta-restrictions) changes changes Memorisation System MS = (CS, RT) concepts, items, associations, restrictions Navigation System NS = (CSn , RTn, PN) navigation rules or restrictions Therefore, we believe that the hypermedia systems must be constructed by a Semantic-Dynamic Model (SDM). With the aim of clarifying the understanding of the model I have divided his explanation in three parts. First of all, I will describe briefly the architecture of the model, second I will define and describe the three most important components of the model: Information Items, Conceptual Structure and Restrictions. Finally, I will come back to the architecture in order to locate the different components inside the model and show their functionality. Arquitecture

Navigation System: Specification: SubGraph CSn So, as we have seen, the Conceptual Structure allows us to label the information items in order to identify them.

Navigation System: Specification: Restrictions RTn (Temporal logic) RTn: c.Portugal.map   c.Countries.list and a.is_a c.Portugal.map   c.Countries.cities and a.is_a c.Portugal.history   c.Countries.cities and a.is_a We distinguish two important types of associations The first type of associations are the relationships between concepts and are domain dependent. So they must be defined by the author The second type of associations are the dependencies between concepts that are independent of the Conceptual Domain. For instance, Process-States and Process-States-Transitions are partOf Process Definition. And Process Active States and Process Suspend States are Process States. Others more informal dependencies are complementary and priority. For instance, Some Management-Interrupt Produces Process-Swithching Defaind

Navigation System: Specification: Creating PN from RTn (Petri Nets) the Conceptual Structure also allows the author to compose items. For instance, Running, Ready and Blocked have the relationship aKindOf with Process-Active-States. Then, a composed item, which is labelled with the generic concept “Process-Active-States” can be constructed by grouping all four items. This composed item inherits the functions or roles of the four items.

Formalisms of a Higher Abstraction Level Meta-level Evolution operations (Ace) (meta-operations with preconditions or meta-restrictions) changes changes Memorisation System MS = (CS, RT) concepts, items, associations, restrictions Navigation System NS = (CSn , RTn, PN) navigation rules or restrictions Therefore, we believe that the hypermedia systems must be constructed by a Semantic-Dynamic Model (SDM). With the aim of clarifying the understanding of the model I have divided his explanation in three parts. First of all, I will describe briefly the architecture of the model, second I will define and describe the three most important components of the model: Information Items, Conceptual Structure and Restrictions. Finally, I will come back to the architecture in order to locate the different components inside the model and show their functionality. Arquitecture

Navigation System: Evolution: Modification of RTn (Predicate temporal logic) Ace: delRest ( c.Portugal.map   c.Countries.list and a.is_of ) Meta-RT: “Every item must be reachable” delRest(c.i, nav_rest)   existRest(c.i, nav_rest1) or  ( existRest(c1.i1, nav_rest2) and ref(nav_rest2, c.i) )  c  C ,  i  II, These restrictions constrain the associations that can be established between the information items. Are always applied by the Navigation System. This system has to take into account at all times: first the items where the document reader is located at any moment; second, concept environment of the item, and third, the information framework or the set of preconditions that is true for an information item.

Navigation System: Evolution: Generating a new PN We can distinguish two types of preconditions: The preconditions derived from semantic structure mean that the navigation will be restricted inside the Conceptual Structure. We can also establish preconditions that are derived from the navigation itself. We can establish a set of preconditions that allow a more or less guided navigation. /gaidid/ In addition, we can take into account the navigation carried out by the reader, and finally, we can distinguish between different readers and show for one reader only a part of the conceptual domain, only some versions of the items or only some of the functions The possibility of adding preconditions implies adaptations and changes in the hypermedia system. diraivd

Evolution and flexibility Information System CSm1 . CSmn CSn11 CSn1n CSnn1 CSnnn PN1n1 PN1nn PN111 PN11n PNnn1 PNnnn PNn11 PNn1n Different ways of representing, structuring and browsing the same Information System

Specification formalisation In particular, Knowledge System contains the semantic structure, therefore, labelled graphs are the more suitable mechanism for representing it. Restrictions are represented by means of temporal logic in order to indicate what associations are valid in the Conceptual Structure. In the Navigation System, the main objective is to restrict the possible paths or navigations. Temporal logic allows to specify the order relationships and Petri nets are operational formalisms which can be executed in order to show these paths and analyse their properties.

Evolution formalisation 1) Consistency 2) Propagation inside 3) Propagation outside All systems, include a set of evolving actions, that allow to make and propagate changes in the hypermedia: Actions that redefine some aspects of the system: the conceptual structure, the restrictions. Actions that control the propagation of these change inside of the system itself. Actions that control the propagation of these change outside the system. So, we use set theory to check changes in the graph and predicate temporal logic to manage the meta-restrictions in order to redefine or evolving restrictions. Thank you for your attention I’m sorry if I’ve run over time

Evolving Actions 5) CS evolves (Set Theory) 6) RTa evolve (PTL) 7) If RTa change --> CS changes 8) RTn evolve (PTL) 1) RTs or RTa <--> CS 2) CSn, subset of the CS If CS changes --> CSn changes 3) RTn constraint CSn 4) PN created from RTn. If RTn change --> PN changes This is the appearance of the prototype. As we can see in the pictures, the reader can navigate through the hypermedia or using the concept-map. These are the properties and content of one of the information item. About the same concept, the information system can offer other items with different functions.

Conclusions MS and NS allow a specification and management of the semantics of information and its navigation in a separated way, using different formalisms. The model guarantees the conservation of the integrity of the system by meta-restrictions. Evolution actions inside and outside each system. The selected formalisms allow an easy specification and change of the structure of each system. These evolution formalisms can also be applied to other evolving systems (reactive or temporal ones). We can conclude that.... We have proposed a semantic-dynamic model that provides: contextoalisation

Suggestions for discussion topics Evolution during the development process. Specification and evolution: relationships. Evolution (changes in the structure): Meta-level. Guarantee of consistency: invariants, pre or postconditions. Change impact, propagation of the change. Which formalism is the better? Now, we are working towards three objectives: