Download presentation
Presentation is loading. Please wait.
1
Modeling Community with Tiles joint work with Ivan Lanese Ugo Montanari Dipartimento di Informatica Università di Pisa Roberto Bruni Dipartimento di Informatica Università di Pisa Workshop AGILE, Pisa, 28-29 Luglio 2003 Ongoing Work!
2
Modeling CommUnity with Tiles Roadmap Goal Tiles Translation Example: Buffer Advantages Ongoing and future work Localities
3
Modeling CommUnity with Tiles Roadmap Goal Tiles Translation Example: Buffer Advantages Ongoing and future work Localities
4
Modeling CommUnity with Tiles Goal Operational semantics for CommUnity Modular translation Preserving of “separation of concerns” Abstract semantics via observations
5
Modeling CommUnity with Tiles Roadmap Goal Tiles Translation Example: Buffer Advantages Ongoing and future work Localities
6
Modeling CommUnity with Tiles Tile Logic Operational and Abstract Semantics of Open Concurrent Systems Compositional in Space and Time Specification Formats Depending on chosen connectors Congruence results Category based
7
Modeling CommUnity with Tiles Tiles
8
Modeling CommUnity with Tiles Configurations input interface output interface
9
Modeling CommUnity with Tiles Observations initial interface final interface concurrent computation
10
Modeling CommUnity with Tiles Roadmap Goal Tiles Translation Example: Buffer Advantages Ongoing and future work Localities
11
Modeling CommUnity with Tiles Tile Format for CommUnity Interfaces action, input (typed), output (typed) System Configuration (gs-monoidal) graphs components: empty input interface coordinators: binary input interface Observations monoid of unary boxes Tiles computation (empty trigger) coordination
12
Modeling CommUnity with Tiles From Modules to Tile Configurations o 1 =v 1 … o m =v m p 1 =w 1 … p k =w k I O A I O … … externalization of the state?
13
Modeling CommUnity with Tiles From Modules to (Schemes of) Basic Tiles I O A ’’ I O A act v w (G) one tile for any ( ,w) that satisfies the guard G of action act actions are translated into tiles
14
Modeling CommUnity with Tiles Data Coordinators I/O I v I v v
15
Modeling CommUnity with Tiles Action Coordinators A A bibi A A aiai (a 1,b 1 ) … (a n,b n ) (a 1,b 1 ) … (a n,b n )
16
Modeling CommUnity with Tiles From Diagrams to Tile Systems M1M1 M2M2 C M1M1 M2M2 C
17
Modeling CommUnity with Tiles Roadmap Goal Tiles Translation Example: Buffer Advantages Ongoing and future work Localities
18
Modeling CommUnity with Tiles Buffer Design design BUFFER [ t:sort, n:nat ] is in i:t out o:t prv b:list[t,n] do put : |b|<n b:=b.i fwd : |b|==n o:=head(b) || b:=tail(b).i get : |b|>0 o:=head(b) || b:=tail(b) shft : |b|==0 o:=i
19
Modeling CommUnity with Tiles A Buffer Configuration o=v b=v 1.v 2 …v k ITIT OTOT A i o BUFFER[T,N]
20
Modeling CommUnity with Tiles Put as a Tile o=v b=v 1.v 2 …v k ITIT OTOT A o=v b=v 1.v 2 …v k.w ITIT OTOT A put v w (k<N)PUT
21
Modeling CommUnity with Tiles Building a Modular Buffer design CABLE [ t:sort] is in i:t do fwd : true skip shft : true skip BUFFER[T,1] CABLE i fwd shft o fwd get shft i fwd put shft BUFFER[T,1]
22
Modeling CommUnity with Tiles Intuition put fwd 1 … fwd k put k+1 get get k shft k+1 … shft n shft shft 1 … shft n fwd fwd 1 … fwd n
23
Modeling CommUnity with Tiles Outcome Bisimilarity between BUFFER[T,N] and M buffers BUFFER[T,K] with N=M*K Bisimilarity between the system and its specification
24
Modeling CommUnity with Tiles Roadmap Goal Tiles Translation Example: Buffer Advantages Ongoing and future work Localities
25
Modeling CommUnity with Tiles Preliminary Results and Advantages Concurrent operational semantics for CommUnity Translation independent from the serialization of represented morphisms Encoding of the colimit behaviourally equivalent to the diagram encoding Standard notion of abstract equivalence (tile bisimilarity) Separation of concerns
26
Modeling CommUnity with Tiles Roadmap Goal Tiles Translation Example: Buffer Advantages Ongoing and future work Localities
27
Modeling CommUnity with Tiles Dealing with Localities Localities as data types input localities (I L ) and output localities (O L ) Actions can involve localities Observable constraints on localities dynamically generated via coordinators statical constraints on moves Open problems which flavour of connectors? what should be observed?
28
Modeling CommUnity with Tiles Future Work Application to AGILE case studies Dealing with system reconfigurations Refinement Common framework for CommUnity and UML
Similar presentations
© 2024 SlidePlayer.com. Inc.
All rights reserved.