SlideShare a Scribd company logo

SDL 2000 Tutorial

jonecx, profile picture
jonecx

The document discusses differences between SDL-92 and SDL-2000. SDL-2000 has been revised with removal of outdated concepts, alignment of existing concepts, and introduction of new concepts. It is now completely object-oriented and has a new formal semantics. Agents are the basic specification concept in SDL-2000 and model active system components as extended finite state machines. Agents can be hierarchically decomposed into block and process agents with different scheduling semantics.

1 of 50
Download to read offline
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
6'/7XWRULDO Prof. J. Fischer Dr. E. Holz Humboldt-Universität zu Berlin SAM 2000 Workshop Grenoble SAM 2000, Grenoble 6'/ • is a major revision of SDL – removal of outdated concepts – alignment of existing concepts – introduction of new concepts • is completely based on object-orientation • has a new formal semantics • is accompanied by a new standard Z.109 SDL-UML-Profil 2 © Humboldt-Universität zu Berlin 1
SAM 2000, Grenoble 'LIIHUHQFHVWR6'/ • document structure of standard has been reorganized • SDL-2000 is case-sensitive – two spellings for keywords all uppercase or all lowercase – removed keywords all, axioms, constant, endgenerator,endnewtype, endrefinement, endservice, error, fpar, generator, imported, literal, map, newtype, noequal, ordering, refinement, returns, reveal, reverse, service, signalroute, view, viewed – new keywords in SDL-2000 abstract, aggregation, association, break, choice, composition, continue, endexceptionhandler, endmethod, endobject, endvalue, exception, exceptionhandler, handle, method, object, onexception, ordered, private, protected, public, raise, value 3 © Humboldt-Universität zu Berlin SAM 2000, Grenoble • not available constructs in SDL-2000 – signal routes - replaced by non-delaying channels – view expression - deleted (import concept, “global” variables) – generators - generalized by parameterized types – block substructures - replaced by nested agents – channel substructures - deleted – signal refinement - deleted – axiomatic definition of data - deleted – macro diagrams - deleted – services - deleted (agents and state aggregations) 4 © Humboldt-Universität zu Berlin 2
SAM 2000, Grenoble 6'/6SHFLILFDWLRQ 6WUXFWXUH Basic1 Basic2 Dependencies between Special specification components MySystem use Basic2; System specification use Special; system MySystem A B package Basic1 Package specification(s) (...) block A (...) Referenced definition(s) 5 © Humboldt-Universität zu Berlin SAM 2000, Grenoble $JHQWV • subsumes the concepts system, block, process • basic specification concept • model active components of a system • an agent instance is an extended finite communicating state machine that has – its own identity – its own signal input queue – its own life cycle – a reactive behaviour specification 6 © Humboldt-Universität zu Berlin 3
SAM 2000, Grenoble $JHQW'HFODUDWLRQ • three main parts – attributes parameters, variables, procedures – behaviour implicit or explicit state machine – internal structure contained agents and communication paths each of the parts may be optional 7 © Humboldt-Universität zu Berlin SAM 2000, Grenoble Declaration of local variables block B dcl i Natural, Reference to agent´s c Character; state machine B Contained agents B1 B2 Communication paths 8 © Humboldt-Universität zu Berlin 4
SAM 2000, Grenoble block B dcl i Natural, c Character; B block B2 B2_1 B1 B2_2 B2 B2_3 9 © Humboldt-Universität zu Berlin SAM 2000, Grenoble 'HRPSRVLWLRQRI$JHQWV • structural decomposition into internal agents implies also decomposition of behaviour • container of an agent determines scheduling semantics of its contents – concurrent agents: block – alternating agents: process 10 © Humboldt-Universität zu Berlin 5
SAM 2000, Grenoble process A block B dcl i Natural, dcl i Natural, c Character; c Character; A B A1 B1 A2 B2 11 © Humboldt-Universität zu Berlin SAM 2000, Grenoble %ORFN$JHQW • all contained agents execute concurrently with each other and with the agents state machine – multiple threads of control – concurrent execution of multiple transitions – transitions execute with run-to-completion • contained agents may be – blocks or processes 12 © Humboldt-Universität zu Berlin 6
SAM 2000, Grenoble block B dcl i Natural, c Character; B B1 B2 13 © Humboldt-Universität zu Berlin SAM 2000, Grenoble 3URFHVV$JHQW • all contained agents execute alternating with each other and with the agents state machine – at most one transition is executed at any point in time – selection is non-determined – transitions execute in run-to-completion • contained agents – may be of kind process only 14 © Humboldt-Universität zu Berlin 7
SAM 2000, Grenoble process B dcl i Natural, c Character; B B1 B2 15 © Humboldt-Universität zu Berlin SAM 2000, Grenoble 6VWHP$JHQW • a system is one special (block) agent – must be the outermost agent – defines the border to the environment – can define communication primitives exchanged with the environment 16 © Humboldt-Universität zu Berlin 8
SAM 2000, Grenoble 9DULDEOHVLQ$JHQWV • variables – store data local to an agent – owned by agents state machine • private variables, visible only to – agents state machine • local variables, visible to – agents state machine – contained agents • public (exported) variables – visibility controlled by remote declaration 17 © Humboldt-Universität zu Berlin SAM 2000, Grenoble private variables local variables read/write access under control of state machine B dcl p Pid ; block B dcl i Natural, c Character; B block B2 B2 B2_1 B1 B2_3 B2 B2_2 18 © Humboldt-Universität zu Berlin 9
SAM 2000, Grenoble dcl p Pid ; process B dcl i Natural, c Character; B process B2 B2 B2_1 B1 B2_3 B2 B2_2 19 © Humboldt-Universität zu Berlin SAM 2000, Grenoble 5HPRWH9DULDEOHV • read-access to variables owned by an other agent – short-hand notation for signal interchanges – provision by exported-declaration – visible by remote-dedclaration – no imported declaration – access with import-operation – update (by owner) with export-operation 20 © Humboldt-Universität zu Berlin 10
SAM 2000, Grenoble block Container remote n Natural; block A block B dcl my_n Natural; dcl exported n Natural Natural, c Character; A B n‘ A1 B1 I1 A2 B2 my_n:= import (n to PidToAgentB); n:= n+10; export(n); 21 © Humboldt-Universität zu Berlin SAM 2000, Grenoble *HQHUDORPPXQLFDWLRQ • communication is based on signal exchange • a signal carries – kind of signal (signal name) – user data – implicit sender identification (Pid value) signal aSignal (Natural, Character); 22 © Humboldt-Universität zu Berlin 11
SAM 2000, Grenoble • communication requires a complete path from sender to receiver consisting of – gates – channels – connections • path may be defined – explicitly – or implicitly derived 23 © Humboldt-Universität zu Berlin SAM 2000, Grenoble • channel – uni- or bi-directional communication path between two endpoints • gate, agent, connection, state machine – safe and reliable (no loss, no re-ordering,...) – delaying or non-delaying transmission – name and signallists are optional – supersedes also signalroute concept 24 © Humboldt-Universität zu Berlin 12
SAM 2000, Grenoble channel aChannel nodelay from block1 to block2 with sig1, sig2; from block2 to block1 with sig1, sig3; endchannel; channel nodelay from block1 to block2; from block2 to block1; endchannel; aChannel block1 block2 [sig1,sig3] [sig1,sig2] 25 © Humboldt-Universität zu Berlin SAM 2000, Grenoble • gate – potential named endpoint for a channel at an agent, agent type or a state machine – uni- or bi-directional – possibly constrained by set of signals or by interface • connection – joining/splitting of channels at implicit gate 26 © Humboldt-Universität zu Berlin 13
SAM 2000, Grenoble block B; block B gate g1 in with sig2; out with sig1; dcl i Natural, gate g2 in with sig3,sig4; c Character; state B referenced; g1 block B2 referenced; c1 B [sig1] [sig2] channel c1 from env via g1 to this; from this to env via g1; c4 endchannel; c2 channel c2 from this to B2; from B2 to this; g2 endchannel; B2 channel c3 from env via g2 [sig3,sig4] g3 to B2 via g3; c3 endchannel; channel c4 from env via g2 to this; endchannel; endblock B; 27 © Humboldt-Universität zu Berlin SAM 2000, Grenoble • implicit signal lists on gates and channels • implicit gates – introduced for connections – introduced for unconnected channel endpoints – signal lists derived from channel signal lists • implicit channels – introduced for unconnected gates – gates have to have matching constraints 28 © Humboldt-Universität zu Berlin 14
SAM 2000, Grenoble system aSystem block B1 block b2 dcl i Natural, dcl i Natural, c Character; c Character; c c B1 B2 Local Local unconnected gate Implicit channel Implicit gate Implicit gate 29 © Humboldt-Universität zu Berlin SAM 2000, Grenoble $GYDQFHGRPPXQLFDWLRQ • two-way communication – implicitly mapped onto signal exchange – remote variables • read access to variables of other agents • no containment relation required – remote procedures • execution of a procedure by a different agent – remote procedure and and remote variables can be mentioned in signallists 30 © Humboldt-Universität zu Berlin 15
SAM 2000, Grenoble 6LPSOH6WDWH0DFKLQHV • behaviour of an agent – is specified by a state machine • two main constituents: – states • particular condition in which an agent may consume a signal – transitions • sequence of activities triggered by the consumption of a signal 31 © Humboldt-Universität zu Berlin SAM 2000, Grenoble 6WDWH7UDQVLWLRQ • a state machine is always either – in a state waiting for a signal – or performing a transition • a transition results in – entering a new state or – stopping the agent 32 © Humboldt-Universität zu Berlin 16
SAM 2000, Grenoble start of behaviour state start; nextstate State1; State1 state State1; input Sig1; nextstate State2; Sig2 input Sig2; Sig1 nextstate State1; endstate; input of signal State2 State1 transition 33 © Humboldt-Universität zu Berlin SAM 2000, Grenoble 7UDQVLWLRQ7ULJJHUV • the first signal of input queue is removed by an input that identifies the signal • if there is no input for the first signal, it is discarded • during an input data carried by a signal may be copied onto local variables • reference to originating agent can be obtained by implicit expression sender 34 © Humboldt-Universität zu Berlin 17
SAM 2000, Grenoble x=5 y=1 dcl x Natural, y Natural; sig1 State1 sig3(5,1) sig1 sig2 Sig2 Sig3(x,y) sig1 State1 State1 discard consume 35 © Humboldt-Universität zu Berlin SAM 2000, Grenoble • consumption of signals State1 can be deferred until a new state is entered Sig2 Sig1 – signals saved in this state – valid until a new state is entered State2 – avoids implicit discard state State1; input Sig1; nextstate State2; save Sig1; endstate; 36 © Humboldt-Universität zu Berlin 18
SAM 2000, Grenoble • input of selected signals can be preferred or constrained – priority input • transition will be selected even if the signal is not the first in the queue – enabling condition • transition will be selected only in case the attached condition is true (otherwise save) 37 © Humboldt-Universität zu Berlin SAM 2000, Grenoble prioritized input State1 state State1; priority input Sig2; Sig1 Sig2 nextstate State1; input Sig1; nextstate State2 endstate; State2 State1 state State2; input sig1; provided x5 Sig1 nextstate State2; endstate; x5 State2 constrained input 38 © Humboldt-Universität zu Berlin 19
SAM 2000, Grenoble • transitions may also be triggered without an (explicit) signal – continuous signal • transition will be selected if attached condition is true and no other transition can be selected i.e. queue is empty or all other signals are saved – non-deterministic transition • transition will be selected non-deterministically and independent from any other transition 39 © Humboldt-Universität zu Berlin SAM 2000, Grenoble non-deterministic input State1 state State1; input none; Sig1 none nextstate State1; input Sig1; nextstate State2; endstate; State2 State3 state State2; provided x5 x5 nextstate State2; endstate; State2 continuous signal 40 © Humboldt-Universität zu Berlin 20
SAM 2000, Grenoble 7UDQVLWLRQ$FWLRQV • output – generation and addressing of signals (identification of receiver or communication path) • task – sequence of simple or compound statements – informal text • decision – branching a transition into a series of alternative paths 41 © Humboldt-Universität zu Berlin SAM 2000, Grenoble state State1; State1 input Sig1(x); task { y:=x+5; if(y10) Sig2(x) y:=y-10;} Sig1(x) nextstate State2; input Sig2(x); true false x0 decision (x0); y:=x+5; true: nextstate State1; if(y10) false: nextstate State2; y:=y-10 enddecision endstate; State1 State2 Decision Sig3(x, y) State2 Task © Humboldt-Universität zu Berlin Output 42 21
SAM 2000, Grenoble $OJRULWKPLF1RWDWLRQ • statement lists provide a means for a concise textual notation of algorithms • applicable in – tasks – procedure definition – operation definition • programming-language like syntax 43 © Humboldt-Universität zu Berlin SAM 2000, Grenoble • compound statements • if-statements task { • decision-statements for( dcl P Pid :=self, not(P=null); • loop-statements dcl i Natural:=0,,i+1) • break-statements { P:=create Proc; • exception-statements if(newP=null) break; else output Sig(P,i) to parent; • output, export, return, } create,set/reset, raise, call, assignment 44 © Humboldt-Universität zu Berlin 22
SAM 2000, Grenoble UHDWHDQG6WRS • performance of a create-request action results in the existence of a new agent instance in the indicated agent set – creators implicit offspring expression refers to new createe – createe´s implicit parent expression refers to creator • initial internal structure will be created too 45 © Humboldt-Universität zu Berlin SAM 2000, Grenoble block container (1,1) block B (1,5) dcl i Natural, Natural, block A(1,1); c Character; parent offspring B A B1 B2 State1 Sig2 create stop B © Humboldt-Universität zu Berlin State2 46 23
SAM 2000, Grenoble • create request can also be based on a type definition • new instance will either be created in – existing instance set based on that type and defined in the same agent as the creator – or implicit instance set based on that type and located in the same agent as the creator 47 © Humboldt-Universität zu Berlin SAM 2000, Grenoble system MySystem BT1 BT2 block container block A(1,1); B :BT1 A implicitname :BT2 State1 State1 Sig1 Sig2 BT1 BT2 © Humboldt-Universität zu Berlin State2 State2 48 24
SAM 2000, Grenoble • the execution of a stop results in entering a special implicit stopping-state – no further execution of transitions – if agent contains no further agent instances, it will cease to exist – otherwise it will the access to the agents variables 49 © Humboldt-Universität zu Berlin SAM 2000, Grenoble 2EMHFW2ULHQWDWLRQLQ6'/ • structural typing concepts allow to define the properties of a set of specification elements • kinds of structural types – agent type – state type – signal (type) – procedures (type) – data types and interfaces 50 © Humboldt-Universität zu Berlin 25
SAM 2000, Grenoble • type concept corresponds to class concept in other OO languages and notations – inheritance – virtuality – abstraction – instance definition creation • all instance definitions in SDL are either explicitly or implicitly based on a type 51 © Humboldt-Universität zu Berlin SAM 2000, Grenoble • inheritance allows the definition of a type basing on another (super-) type of the same kind – addition of new structural elements – addition of new behavioural elements – redefinition of virtual elements • single inheritance supported by all types • multiple inheritance supported by interface 52 © Humboldt-Universität zu Berlin 26
SAM 2000, Grenoble block type B dcl i Natural; B block type B2 I2_1 I2:B2 I2_2 block type Bnew inherits B adding dcl c Natural; B I1:B3 I2 53 © Humboldt-Universität zu Berlin SAM 2000, Grenoble 7SH5HIHUHQFHVDQG 5HODWLRQV • UML-like class symbols can be used to refer to type definitions and diagrams • partial specification: – type name – type attributes – type behaviour properties • multiple references are allowed – must all be consistent with type definition 54 © Humboldt-Universität zu Berlin 27
SAM 2000, Grenoble • relations between types can be depicted by – associations (binary relations) • no predefined semantics implied – specialization • must be consistent wit a specialization in the type definition (inherits-clause) 55 © Humboldt-Universität zu Berlin SAM 2000, Grenoble «block» B - i Natural; B2 signal s1; signal s2; BNew - c Natural; B3 signal s3; signal s4; 56 © Humboldt-Universität zu Berlin 28
SAM 2000, Grenoble • type definitions for an element can be given in – the scope unit where the entity can be given – any surrounding scope unit or – any type definition for such a scope unit – a package • instances of such a type can be defined – where the type is visible and – the element is allowed 57 © Humboldt-Universität zu Berlin SAM 2000, Grenoble RQWH[W'HSHQGHQFLHV • types may refer to instances and types in their defining context – definition of instances may be limited to the same scope unit (e.g. in case of instance references) • types may refer to instances and types in their instantiation context – formal context parameters in defining context – actual context parameters in instantiation context 58 © Humboldt-Universität zu Berlin 29
SAM 2000, Grenoble block container block A; B (1,5) A State1 Sig2 B block container B (1,5) block A :Atype B State2 State1 block type Atype Sig2 block P P ? A State2 59 © Humboldt-Universität zu Berlin SAM 2000, Grenoble RQWH[W3DUDPHWHU.LQGV • block (type), process (type) • composite state type • procedure, remote procedure • signal • data type, interface, exception • variable, remote variable, timer, synonym • gate 60 © Humboldt-Universität zu Berlin 30
SAM 2000, Grenoble $EVWUDFWDQG9LUWXDO7SHV • types marked with the keyword abstract do not directly have instances – pure classification – used as super-type in an inheritance hierarchy • virtual types local to another type may be redefined in a specialisation of that type – must be contained in a type definition – redefinition can be constrained • system type can not be abstract or virtual 61 © Humboldt-Universität zu Berlin SAM 2000, Grenoble block type B • I2 is instance set of virtual B2 B virtual type B2 I2:B2 block type Bnew inherits B • I2 is instance set of redefined B2 redefined type B2 block type Bnew2 • I2 is instance set of inherits Bnew finalised type B2 • no further redefinition allowed finalized B2 62 © Humboldt-Universität zu Berlin 31
SAM 2000, Grenoble • redefinition/finalisation must be – subtype of original virtual type or – subtype of virtuality constraint ( virtual block type B atleast Base ) • references to a virtual or redefined type refer to the most recent redefinition • finalised types can not be redefined further 63 © Humboldt-Universität zu Berlin SAM 2000, Grenoble $GYDQFHG6WDWH0DFKLQHV • exceptions are used to denote and handle unexpected or exceptional behaviour – exception: the type of cause – exception handler:behaviour to occur after an exception (handle-clauses) – onexception: attaches exception handler to a behaviour unit – raise: forces a transition to throw an exception 64 © Humboldt-Universität zu Berlin 32
SAM 2000, Grenoble exception definition exception comerror; exception handler on exception association State1 eh eh Sig1(x) Sig2 DivisionbyZero comerror else parent error sender to sender y:=5/x Comerror State2 State2 State1 handle transition raise statement 65 © Humboldt-Universität zu Berlin SAM 2000, Grenoble • exception handlers can be attached to all kinds of behaviour by an onexception: • complete state machine, state, • input transition, transition/algorithm action • terminator, connect, • single procedure, single operation • single remote procedure • exception handler, handle transition • in case of an exception the most local active exception handler will be selected 66 © Humboldt-Universität zu Berlin 33
SAM 2000, Grenoble 3URFHGXUHV • procedures are a means to group and name recurrent behaviour • notation corresponds to agent state machine – local states, inputs and transitions – local variables, parameters • procedures are a type • exceptions raise but not handled in a procedure are mentioned explicitly 67 © Humboldt-Universität zu Berlin SAM 2000, Grenoble 5HPRWH3URFHGXUHV • an agent can make its procedures available for other agents – remote procedures – realized by two-way communication between caller and server • after a call to a remote procedure the caller is blocked until he receives the procedure return from the server 68 © Humboldt-Universität zu Berlin 34
SAM 2000, Grenoble • remote procedure call may deadlock – can be prevented by an associated timer, which raises an exception • server accepts calls for remote procedures in any state – execution may be deferred by save – execution may be rejected by input p raise deny • exceptions raised by the remote procedure are raised at client and server side 69 © Humboldt-Universität zu Berlin SAM 2000, Grenoble block client Initial eh block server P Make Call Initial set(10,t) P to server P raises timer t ServerStart reject State2 Initial ´initialization´ eh State2 reject t initial State2 70 © Humboldt-Universität zu Berlin 35
SAM 2000, Grenoble RPSRVLWH6WDWHV • composite states are a means to hierarchically structure state machines – nesting of states – agent can be in more than one state at a time – Harel´s state charts • composite state is itself a sub-state machine • state machine of an agent is in fact a top- level composite state 71 © Humboldt-Universität zu Berlin SAM 2000, Grenoble State1 St1 Sig3 Sig4 State1 Sig1 Sig2 State2 P1 P1 State2 State2 St1 via P1 Sig3 ex ex B State1 ex 72 © Humboldt-Universität zu Berlin 36
SAM 2000, Grenoble • composite states share agent´s input queue • internal transitions with the same trigger as external transitions have higher priority • excactly one transition is executed – possibly concatenated with triggerless transitions • special procedures may be used to define initialisation and finalisation – called implicitly upon entering/leaving a composite state 73 © Humboldt-Universität zu Berlin SAM 2000, Grenoble initialization/finalization State1 entry exit St1 St1 Sig3 Sig4 St1_1 B State1 Sig1 Sig2 Sig3 - -* State2 history nextstate 74 © Humboldt-Universität zu Berlin 37
SAM 2000, Grenoble 6WDWH$JJUHJDWLRQ • state aggregations partition the state space of an agents state machine • each partition handles a different set of the input stimuli • excactly one partition is executing a transition at any point in time – multiple enabled transitions are executed in an interleaved manner 75 © Humboldt-Universität zu Berlin SAM 2000, Grenoble State1 state aggregation st State1 St1 Sig1 Sig2 State2 State2 st_i St1 st Sig4 Sig3 st1 Sig1 st_i B st2 76 © Humboldt-Universität zu Berlin 38
SAM 2000, Grenoble • composite states and state aggregations can be classified – composite state type definition – typebased composite states/state aggregations • concept notation similar to agent types • instances are static – livedie with containing agent – multiple instances in the same scope must have different names 77 © Humboldt-Universität zu Berlin SAM 2000, Grenoble 9LUWXDO%HKDYLRXU(OHPHQWV • allow the redefinition or replacement of behaviour elements in a type specialisation • redefinition and finalisation similar to structural elements • available for – procedures – transitions – exception handle transitions 78 © Humboldt-Universität zu Berlin 39
SAM 2000, Grenoble ,QWHUIDFH • pure typing concept used for typed communication between agents • interface definition groups and names a set of – remote variable – remote procedure – signal definitions • gates and channels paths can be typed by interfaces 79 © Humboldt-Universität zu Berlin SAM 2000, Grenoble • contained remote procedures variables and signals are defined and visble where interface is visible • interface can use also existing definitions for such elements • multiple inheritance is available for interfaces • interfaces are used as matching constraints for implicit channels 80 © Humboldt-Universität zu Berlin 40
SAM 2000, Grenoble interface if1; interface if3 signal sig1; inherits if1,if2; procedure P; dcl I Natural; endinterface; [if1] [if2] block type b signal sig2,sig3; interface if2; g use sig2, sig3; if3 endinterface; 81 © Humboldt-Universität zu Berlin SAM 2000, Grenoble $JHQW,PSOLFLW,QWHUIDFH • each agent and agent type introduces an implicit interface – same name as agent (type) • contains all – signals accepted by the agents state machine – remote variables/procedures provided by agents state machine • inherits all interfaces on gates connected to agents state machine 82 © Humboldt-Universität zu Berlin 41
SAM 2000, Grenoble block type B g1 c1 B interface B [sig1] [sig2] [sig4] use sig2,sig4; c4 c2 endinterface g2 [sig3,sig4] g3 B2 c3 block type C interface C g1 inherits B inherits B, if1; [sig3] use sig3; B if1 endinterface 83 © Humboldt-Universität zu Berlin SAM 2000, Grenoble • each (explicit or implicit) interface implies a specialization of the Pid-type • can be used to refer to agent instances in – variables, expressions – output, imports or remote procedure calls – assignment attempt • dynamic type check tests – provision of remote variables/procedures – acceptance of signals – may raise an exception (InvalidReference) 84 © Humboldt-Universität zu Berlin 42
SAM 2000, Grenoble block B assignment attempt dcl myb2 B2, mybt BT; g1 c1 B [sig1] [sig2] c4 B2 c2 mybt:=offspring g2 g B2:BT [sig3,sig4] c3 Sig3(x, y) to mybt State2 © Humboldt-Universität zu Berlin compatibility check 85 SAM 2000, Grenoble 'DWD7SHV • two main kinds of data types – value types – object types (references) conversion possible • data type definition specifies – data elements and structure – operators and methods for data manipulation • package Predefined contains a set of general data type definitions 86 © Humboldt-Universität zu Berlin 43
SAM 2000, Grenoble • value types correspond to the newtype concept of SDL-92 – axioms and generators have been removed – behaviour of operators defined by algorithms or transition actions (functional description) – constructors: – literals – structs – choices 87 © Humboldt-Universität zu Berlin SAM 2000, Grenoble • object types define references to values – references are local to agents • definition similar to value types – conversion possible • polymorphic assignments 88 © Humboldt-Universität zu Berlin 44
SAM 2000, Grenoble • further properties of data types: • inheritance • context parameters • methods allow operation-calls in programming- language like dot-notation • local data types, constants, exceptions • visibility of data elements can be controlled 89 © Humboldt-Universität zu Berlin SAM 2000, Grenoble REMHFW WSH /LVW WSH (OHP! VWUXFW HOHP (OHP SULYDWH QH[W /LVW RSHUDWRUV 0DNH(OHP
!/LVW PHWKRGV DGG(OHP
! /LVW WKLV HOHP HOHP HOHP QH[W QH[W QH[W 1XOO HQGREMHFW WSH 90 © Humboldt-Universität zu Berlin 45
SAM 2000, Grenoble REMHFW WSH /LVW WSH (OHP! WKLV HOHP HOHP HOHP QH[W QH[W QH[W 1XOO RSHUDWRU 0DNHH (OHP
^ UHWXUQ H1XOO
` PHWKRG DGGH (OHP

More Related Content

PDF
Bcs facs
Bernie Cohen
 
PDF
Structural Order and Disorder Dictate Sequence And Functional Evolution of th...
Asociación Argentina de Bioinformática y Biología Computacional
 
PDF
Owl syntax
jonecx
 
PDF
tdt4260
jonecx
 
PDF
Latex tutorial
jonecx
 
PDF
Tdt4242
jonecx
 
PDF
BPMN
jonecx
 
PDF
Duplicate detection
jonecx
 
Bcs facs
Bernie Cohen
 
Structural Order and Disorder Dictate Sequence And Functional Evolution of th...
Asociación Argentina de Bioinformática y Biología Computacional
 
Owl syntax
jonecx
 
tdt4260
jonecx
 
Latex tutorial
jonecx
 
Tdt4242
jonecx
 
BPMN
jonecx
 
Duplicate detection
jonecx
 

Featured (20)

PDF
2024 Trend Updates: What Really Works In SEO & Content Marketing
Search Engine Journal
 
PDF
Storytelling For The Web: Integrate Storytelling in your Design Process
Chiara Aliotta
 
PDF
Artificial Intelligence, Data and Competition – SCHREPEL – June 2024 OECD dis...
OECD Directorate for Financial and Enterprise Affairs
 
PDF
How to Leverage AI to Boost Employee Wellness - Lydia Di Francesco - SocialHR...
SocialHRCamp
 
PDF
2024 State of Marketing Report – by Hubspot
Marius Sescu
 
PDF
Everything You Need To Know About ChatGPT
Expeed Software
 
PDF
Product Design Trends in 2024 | Teenage Engineerings
Pixeldarts
 
PDF
How Race, Age and Gender Shape Attitudes Towards Mental Health
ThinkNow
 
PDF
AI Trends in Creative Operations 2024 by Artwork Flow.pdf
marketingartwork
 
PDF
Skeleton Culture Code
Skeleton Technologies
 
PDF
PEPSICO Presentation to CAGNY Conference Feb 2024
Neil Kimberley
 
PDF
Content Methodology: A Best Practices Report (Webinar)
contently
 
PPTX
How to Prepare For a Successful Job Search for 2024
Albert Qian
 
PDF
Social Media Marketing Trends 2024 // The Global Indie Insights
Kurio // The Social Media Age(ncy)
 
PDF
Trends In Paid Search: Navigating The Digital Landscape In 2024
Search Engine Journal
 
PDF
5 Public speaking tips from TED - Visualized summary
SpeakerHub
 
PDF
ChatGPT and the Future of Work - Clark Boyd
Clark Boyd
 
PDF
Getting into the tech field. what next
Tessa Mero
 
PDF
Google's Just Not That Into You: Understanding Core Updates & Search Intent
Lily Ray
 
PDF
How to have difficult conversations
Rajiv Jayarajah, MAppComm, ACC
 
2024 Trend Updates: What Really Works In SEO & Content Marketing
Search Engine Journal
 
Storytelling For The Web: Integrate Storytelling in your Design Process
Chiara Aliotta
 
Artificial Intelligence, Data and Competition – SCHREPEL – June 2024 OECD dis...
OECD Directorate for Financial and Enterprise Affairs
 
How to Leverage AI to Boost Employee Wellness - Lydia Di Francesco - SocialHR...
SocialHRCamp
 
2024 State of Marketing Report – by Hubspot
Marius Sescu
 
Everything You Need To Know About ChatGPT
Expeed Software
 
Product Design Trends in 2024 | Teenage Engineerings
Pixeldarts
 
How Race, Age and Gender Shape Attitudes Towards Mental Health
ThinkNow
 
AI Trends in Creative Operations 2024 by Artwork Flow.pdf
marketingartwork
 
Skeleton Culture Code
Skeleton Technologies
 
PEPSICO Presentation to CAGNY Conference Feb 2024
Neil Kimberley
 
Content Methodology: A Best Practices Report (Webinar)
contently
 
How to Prepare For a Successful Job Search for 2024
Albert Qian
 
Social Media Marketing Trends 2024 // The Global Indie Insights
Kurio // The Social Media Age(ncy)
 
Trends In Paid Search: Navigating The Digital Landscape In 2024
Search Engine Journal
 
5 Public speaking tips from TED - Visualized summary
SpeakerHub
 
ChatGPT and the Future of Work - Clark Boyd
Clark Boyd
 
Getting into the tech field. what next
Tessa Mero
 
Google's Just Not That Into You: Understanding Core Updates & Search Intent
Lily Ray
 
How to have difficult conversations
Rajiv Jayarajah, MAppComm, ACC
 
Ad

SDL 2000 Tutorial

  • 1. 6'/7XWRULDO Prof. J. Fischer Dr. E. Holz Humboldt-Universität zu Berlin SAM 2000 Workshop Grenoble SAM 2000, Grenoble 6'/ • is a major revision of SDL – removal of outdated concepts – alignment of existing concepts – introduction of new concepts • is completely based on object-orientation • has a new formal semantics • is accompanied by a new standard Z.109 SDL-UML-Profil 2 © Humboldt-Universität zu Berlin 1
  • 2. SAM 2000, Grenoble 'LIIHUHQFHVWR6'/ • document structure of standard has been reorganized • SDL-2000 is case-sensitive – two spellings for keywords all uppercase or all lowercase – removed keywords all, axioms, constant, endgenerator,endnewtype, endrefinement, endservice, error, fpar, generator, imported, literal, map, newtype, noequal, ordering, refinement, returns, reveal, reverse, service, signalroute, view, viewed – new keywords in SDL-2000 abstract, aggregation, association, break, choice, composition, continue, endexceptionhandler, endmethod, endobject, endvalue, exception, exceptionhandler, handle, method, object, onexception, ordered, private, protected, public, raise, value 3 © Humboldt-Universität zu Berlin SAM 2000, Grenoble • not available constructs in SDL-2000 – signal routes - replaced by non-delaying channels – view expression - deleted (import concept, “global” variables) – generators - generalized by parameterized types – block substructures - replaced by nested agents – channel substructures - deleted – signal refinement - deleted – axiomatic definition of data - deleted – macro diagrams - deleted – services - deleted (agents and state aggregations) 4 © Humboldt-Universität zu Berlin 2
  • 3. SAM 2000, Grenoble 6'/6SHFLILFDWLRQ 6WUXFWXUH Basic1 Basic2 Dependencies between Special specification components MySystem use Basic2; System specification use Special; system MySystem A B package Basic1 Package specification(s) (...) block A (...) Referenced definition(s) 5 © Humboldt-Universität zu Berlin SAM 2000, Grenoble $JHQWV • subsumes the concepts system, block, process • basic specification concept • model active components of a system • an agent instance is an extended finite communicating state machine that has – its own identity – its own signal input queue – its own life cycle – a reactive behaviour specification 6 © Humboldt-Universität zu Berlin 3
  • 4. SAM 2000, Grenoble $JHQW'HFODUDWLRQ • three main parts – attributes parameters, variables, procedures – behaviour implicit or explicit state machine – internal structure contained agents and communication paths each of the parts may be optional 7 © Humboldt-Universität zu Berlin SAM 2000, Grenoble Declaration of local variables block B dcl i Natural, Reference to agent´s c Character; state machine B Contained agents B1 B2 Communication paths 8 © Humboldt-Universität zu Berlin 4
  • 5. SAM 2000, Grenoble block B dcl i Natural, c Character; B block B2 B2_1 B1 B2_2 B2 B2_3 9 © Humboldt-Universität zu Berlin SAM 2000, Grenoble 'HRPSRVLWLRQRI$JHQWV • structural decomposition into internal agents implies also decomposition of behaviour • container of an agent determines scheduling semantics of its contents – concurrent agents: block – alternating agents: process 10 © Humboldt-Universität zu Berlin 5
  • 6. SAM 2000, Grenoble process A block B dcl i Natural, dcl i Natural, c Character; c Character; A B A1 B1 A2 B2 11 © Humboldt-Universität zu Berlin SAM 2000, Grenoble %ORFN$JHQW • all contained agents execute concurrently with each other and with the agents state machine – multiple threads of control – concurrent execution of multiple transitions – transitions execute with run-to-completion • contained agents may be – blocks or processes 12 © Humboldt-Universität zu Berlin 6
  • 7. SAM 2000, Grenoble block B dcl i Natural, c Character; B B1 B2 13 © Humboldt-Universität zu Berlin SAM 2000, Grenoble 3URFHVV$JHQW • all contained agents execute alternating with each other and with the agents state machine – at most one transition is executed at any point in time – selection is non-determined – transitions execute in run-to-completion • contained agents – may be of kind process only 14 © Humboldt-Universität zu Berlin 7
  • 8. SAM 2000, Grenoble process B dcl i Natural, c Character; B B1 B2 15 © Humboldt-Universität zu Berlin SAM 2000, Grenoble 6VWHP$JHQW • a system is one special (block) agent – must be the outermost agent – defines the border to the environment – can define communication primitives exchanged with the environment 16 © Humboldt-Universität zu Berlin 8
  • 9. SAM 2000, Grenoble 9DULDEOHVLQ$JHQWV • variables – store data local to an agent – owned by agents state machine • private variables, visible only to – agents state machine • local variables, visible to – agents state machine – contained agents • public (exported) variables – visibility controlled by remote declaration 17 © Humboldt-Universität zu Berlin SAM 2000, Grenoble private variables local variables read/write access under control of state machine B dcl p Pid ; block B dcl i Natural, c Character; B block B2 B2 B2_1 B1 B2_3 B2 B2_2 18 © Humboldt-Universität zu Berlin 9
  • 10. SAM 2000, Grenoble dcl p Pid ; process B dcl i Natural, c Character; B process B2 B2 B2_1 B1 B2_3 B2 B2_2 19 © Humboldt-Universität zu Berlin SAM 2000, Grenoble 5HPRWH9DULDEOHV • read-access to variables owned by an other agent – short-hand notation for signal interchanges – provision by exported-declaration – visible by remote-dedclaration – no imported declaration – access with import-operation – update (by owner) with export-operation 20 © Humboldt-Universität zu Berlin 10
  • 11. SAM 2000, Grenoble block Container remote n Natural; block A block B dcl my_n Natural; dcl exported n Natural Natural, c Character; A B n‘ A1 B1 I1 A2 B2 my_n:= import (n to PidToAgentB); n:= n+10; export(n); 21 © Humboldt-Universität zu Berlin SAM 2000, Grenoble *HQHUDORPPXQLFDWLRQ • communication is based on signal exchange • a signal carries – kind of signal (signal name) – user data – implicit sender identification (Pid value) signal aSignal (Natural, Character); 22 © Humboldt-Universität zu Berlin 11
  • 12. SAM 2000, Grenoble • communication requires a complete path from sender to receiver consisting of – gates – channels – connections • path may be defined – explicitly – or implicitly derived 23 © Humboldt-Universität zu Berlin SAM 2000, Grenoble • channel – uni- or bi-directional communication path between two endpoints • gate, agent, connection, state machine – safe and reliable (no loss, no re-ordering,...) – delaying or non-delaying transmission – name and signallists are optional – supersedes also signalroute concept 24 © Humboldt-Universität zu Berlin 12
  • 13. SAM 2000, Grenoble channel aChannel nodelay from block1 to block2 with sig1, sig2; from block2 to block1 with sig1, sig3; endchannel; channel nodelay from block1 to block2; from block2 to block1; endchannel; aChannel block1 block2 [sig1,sig3] [sig1,sig2] 25 © Humboldt-Universität zu Berlin SAM 2000, Grenoble • gate – potential named endpoint for a channel at an agent, agent type or a state machine – uni- or bi-directional – possibly constrained by set of signals or by interface • connection – joining/splitting of channels at implicit gate 26 © Humboldt-Universität zu Berlin 13
  • 14. SAM 2000, Grenoble block B; block B gate g1 in with sig2; out with sig1; dcl i Natural, gate g2 in with sig3,sig4; c Character; state B referenced; g1 block B2 referenced; c1 B [sig1] [sig2] channel c1 from env via g1 to this; from this to env via g1; c4 endchannel; c2 channel c2 from this to B2; from B2 to this; g2 endchannel; B2 channel c3 from env via g2 [sig3,sig4] g3 to B2 via g3; c3 endchannel; channel c4 from env via g2 to this; endchannel; endblock B; 27 © Humboldt-Universität zu Berlin SAM 2000, Grenoble • implicit signal lists on gates and channels • implicit gates – introduced for connections – introduced for unconnected channel endpoints – signal lists derived from channel signal lists • implicit channels – introduced for unconnected gates – gates have to have matching constraints 28 © Humboldt-Universität zu Berlin 14
  • 15. SAM 2000, Grenoble system aSystem block B1 block b2 dcl i Natural, dcl i Natural, c Character; c Character; c c B1 B2 Local Local unconnected gate Implicit channel Implicit gate Implicit gate 29 © Humboldt-Universität zu Berlin SAM 2000, Grenoble $GYDQFHGRPPXQLFDWLRQ • two-way communication – implicitly mapped onto signal exchange – remote variables • read access to variables of other agents • no containment relation required – remote procedures • execution of a procedure by a different agent – remote procedure and and remote variables can be mentioned in signallists 30 © Humboldt-Universität zu Berlin 15
  • 16. SAM 2000, Grenoble 6LPSOH6WDWH0DFKLQHV • behaviour of an agent – is specified by a state machine • two main constituents: – states • particular condition in which an agent may consume a signal – transitions • sequence of activities triggered by the consumption of a signal 31 © Humboldt-Universität zu Berlin SAM 2000, Grenoble 6WDWH7UDQVLWLRQ • a state machine is always either – in a state waiting for a signal – or performing a transition • a transition results in – entering a new state or – stopping the agent 32 © Humboldt-Universität zu Berlin 16
  • 17. SAM 2000, Grenoble start of behaviour state start; nextstate State1; State1 state State1; input Sig1; nextstate State2; Sig2 input Sig2; Sig1 nextstate State1; endstate; input of signal State2 State1 transition 33 © Humboldt-Universität zu Berlin SAM 2000, Grenoble 7UDQVLWLRQ7ULJJHUV • the first signal of input queue is removed by an input that identifies the signal • if there is no input for the first signal, it is discarded • during an input data carried by a signal may be copied onto local variables • reference to originating agent can be obtained by implicit expression sender 34 © Humboldt-Universität zu Berlin 17
  • 18. SAM 2000, Grenoble x=5 y=1 dcl x Natural, y Natural; sig1 State1 sig3(5,1) sig1 sig2 Sig2 Sig3(x,y) sig1 State1 State1 discard consume 35 © Humboldt-Universität zu Berlin SAM 2000, Grenoble • consumption of signals State1 can be deferred until a new state is entered Sig2 Sig1 – signals saved in this state – valid until a new state is entered State2 – avoids implicit discard state State1; input Sig1; nextstate State2; save Sig1; endstate; 36 © Humboldt-Universität zu Berlin 18
  • 19. SAM 2000, Grenoble • input of selected signals can be preferred or constrained – priority input • transition will be selected even if the signal is not the first in the queue – enabling condition • transition will be selected only in case the attached condition is true (otherwise save) 37 © Humboldt-Universität zu Berlin SAM 2000, Grenoble prioritized input State1 state State1; priority input Sig2; Sig1 Sig2 nextstate State1; input Sig1; nextstate State2 endstate; State2 State1 state State2; input sig1; provided x5 Sig1 nextstate State2; endstate; x5 State2 constrained input 38 © Humboldt-Universität zu Berlin 19
  • 20. SAM 2000, Grenoble • transitions may also be triggered without an (explicit) signal – continuous signal • transition will be selected if attached condition is true and no other transition can be selected i.e. queue is empty or all other signals are saved – non-deterministic transition • transition will be selected non-deterministically and independent from any other transition 39 © Humboldt-Universität zu Berlin SAM 2000, Grenoble non-deterministic input State1 state State1; input none; Sig1 none nextstate State1; input Sig1; nextstate State2; endstate; State2 State3 state State2; provided x5 x5 nextstate State2; endstate; State2 continuous signal 40 © Humboldt-Universität zu Berlin 20
  • 21. SAM 2000, Grenoble 7UDQVLWLRQ$FWLRQV • output – generation and addressing of signals (identification of receiver or communication path) • task – sequence of simple or compound statements – informal text • decision – branching a transition into a series of alternative paths 41 © Humboldt-Universität zu Berlin SAM 2000, Grenoble state State1; State1 input Sig1(x); task { y:=x+5; if(y10) Sig2(x) y:=y-10;} Sig1(x) nextstate State2; input Sig2(x); true false x0 decision (x0); y:=x+5; true: nextstate State1; if(y10) false: nextstate State2; y:=y-10 enddecision endstate; State1 State2 Decision Sig3(x, y) State2 Task © Humboldt-Universität zu Berlin Output 42 21
  • 22. SAM 2000, Grenoble $OJRULWKPLF1RWDWLRQ • statement lists provide a means for a concise textual notation of algorithms • applicable in – tasks – procedure definition – operation definition • programming-language like syntax 43 © Humboldt-Universität zu Berlin SAM 2000, Grenoble • compound statements • if-statements task { • decision-statements for( dcl P Pid :=self, not(P=null); • loop-statements dcl i Natural:=0,,i+1) • break-statements { P:=create Proc; • exception-statements if(newP=null) break; else output Sig(P,i) to parent; • output, export, return, } create,set/reset, raise, call, assignment 44 © Humboldt-Universität zu Berlin 22
  • 23. SAM 2000, Grenoble UHDWHDQG6WRS • performance of a create-request action results in the existence of a new agent instance in the indicated agent set – creators implicit offspring expression refers to new createe – createe´s implicit parent expression refers to creator • initial internal structure will be created too 45 © Humboldt-Universität zu Berlin SAM 2000, Grenoble block container (1,1) block B (1,5) dcl i Natural, Natural, block A(1,1); c Character; parent offspring B A B1 B2 State1 Sig2 create stop B © Humboldt-Universität zu Berlin State2 46 23
  • 24. SAM 2000, Grenoble • create request can also be based on a type definition • new instance will either be created in – existing instance set based on that type and defined in the same agent as the creator – or implicit instance set based on that type and located in the same agent as the creator 47 © Humboldt-Universität zu Berlin SAM 2000, Grenoble system MySystem BT1 BT2 block container block A(1,1); B :BT1 A implicitname :BT2 State1 State1 Sig1 Sig2 BT1 BT2 © Humboldt-Universität zu Berlin State2 State2 48 24
  • 25. SAM 2000, Grenoble • the execution of a stop results in entering a special implicit stopping-state – no further execution of transitions – if agent contains no further agent instances, it will cease to exist – otherwise it will the access to the agents variables 49 © Humboldt-Universität zu Berlin SAM 2000, Grenoble 2EMHFW2ULHQWDWLRQLQ6'/ • structural typing concepts allow to define the properties of a set of specification elements • kinds of structural types – agent type – state type – signal (type) – procedures (type) – data types and interfaces 50 © Humboldt-Universität zu Berlin 25
  • 26. SAM 2000, Grenoble • type concept corresponds to class concept in other OO languages and notations – inheritance – virtuality – abstraction – instance definition creation • all instance definitions in SDL are either explicitly or implicitly based on a type 51 © Humboldt-Universität zu Berlin SAM 2000, Grenoble • inheritance allows the definition of a type basing on another (super-) type of the same kind – addition of new structural elements – addition of new behavioural elements – redefinition of virtual elements • single inheritance supported by all types • multiple inheritance supported by interface 52 © Humboldt-Universität zu Berlin 26
  • 27. SAM 2000, Grenoble block type B dcl i Natural; B block type B2 I2_1 I2:B2 I2_2 block type Bnew inherits B adding dcl c Natural; B I1:B3 I2 53 © Humboldt-Universität zu Berlin SAM 2000, Grenoble 7SH5HIHUHQFHVDQG 5HODWLRQV • UML-like class symbols can be used to refer to type definitions and diagrams • partial specification: – type name – type attributes – type behaviour properties • multiple references are allowed – must all be consistent with type definition 54 © Humboldt-Universität zu Berlin 27
  • 28. SAM 2000, Grenoble • relations between types can be depicted by – associations (binary relations) • no predefined semantics implied – specialization • must be consistent wit a specialization in the type definition (inherits-clause) 55 © Humboldt-Universität zu Berlin SAM 2000, Grenoble «block» B - i Natural; B2 signal s1; signal s2; BNew - c Natural; B3 signal s3; signal s4; 56 © Humboldt-Universität zu Berlin 28
  • 29. SAM 2000, Grenoble • type definitions for an element can be given in – the scope unit where the entity can be given – any surrounding scope unit or – any type definition for such a scope unit – a package • instances of such a type can be defined – where the type is visible and – the element is allowed 57 © Humboldt-Universität zu Berlin SAM 2000, Grenoble RQWH[W'HSHQGHQFLHV • types may refer to instances and types in their defining context – definition of instances may be limited to the same scope unit (e.g. in case of instance references) • types may refer to instances and types in their instantiation context – formal context parameters in defining context – actual context parameters in instantiation context 58 © Humboldt-Universität zu Berlin 29
  • 30. SAM 2000, Grenoble block container block A; B (1,5) A State1 Sig2 B block container B (1,5) block A :Atype B State2 State1 block type Atype Sig2 block P P ? A State2 59 © Humboldt-Universität zu Berlin SAM 2000, Grenoble RQWH[W3DUDPHWHU.LQGV • block (type), process (type) • composite state type • procedure, remote procedure • signal • data type, interface, exception • variable, remote variable, timer, synonym • gate 60 © Humboldt-Universität zu Berlin 30
  • 31. SAM 2000, Grenoble $EVWUDFWDQG9LUWXDO7SHV • types marked with the keyword abstract do not directly have instances – pure classification – used as super-type in an inheritance hierarchy • virtual types local to another type may be redefined in a specialisation of that type – must be contained in a type definition – redefinition can be constrained • system type can not be abstract or virtual 61 © Humboldt-Universität zu Berlin SAM 2000, Grenoble block type B • I2 is instance set of virtual B2 B virtual type B2 I2:B2 block type Bnew inherits B • I2 is instance set of redefined B2 redefined type B2 block type Bnew2 • I2 is instance set of inherits Bnew finalised type B2 • no further redefinition allowed finalized B2 62 © Humboldt-Universität zu Berlin 31
  • 32. SAM 2000, Grenoble • redefinition/finalisation must be – subtype of original virtual type or – subtype of virtuality constraint ( virtual block type B atleast Base ) • references to a virtual or redefined type refer to the most recent redefinition • finalised types can not be redefined further 63 © Humboldt-Universität zu Berlin SAM 2000, Grenoble $GYDQFHG6WDWH0DFKLQHV • exceptions are used to denote and handle unexpected or exceptional behaviour – exception: the type of cause – exception handler:behaviour to occur after an exception (handle-clauses) – onexception: attaches exception handler to a behaviour unit – raise: forces a transition to throw an exception 64 © Humboldt-Universität zu Berlin 32
  • 33. SAM 2000, Grenoble exception definition exception comerror; exception handler on exception association State1 eh eh Sig1(x) Sig2 DivisionbyZero comerror else parent error sender to sender y:=5/x Comerror State2 State2 State1 handle transition raise statement 65 © Humboldt-Universität zu Berlin SAM 2000, Grenoble • exception handlers can be attached to all kinds of behaviour by an onexception: • complete state machine, state, • input transition, transition/algorithm action • terminator, connect, • single procedure, single operation • single remote procedure • exception handler, handle transition • in case of an exception the most local active exception handler will be selected 66 © Humboldt-Universität zu Berlin 33
  • 34. SAM 2000, Grenoble 3URFHGXUHV • procedures are a means to group and name recurrent behaviour • notation corresponds to agent state machine – local states, inputs and transitions – local variables, parameters • procedures are a type • exceptions raise but not handled in a procedure are mentioned explicitly 67 © Humboldt-Universität zu Berlin SAM 2000, Grenoble 5HPRWH3URFHGXUHV • an agent can make its procedures available for other agents – remote procedures – realized by two-way communication between caller and server • after a call to a remote procedure the caller is blocked until he receives the procedure return from the server 68 © Humboldt-Universität zu Berlin 34
  • 35. SAM 2000, Grenoble • remote procedure call may deadlock – can be prevented by an associated timer, which raises an exception • server accepts calls for remote procedures in any state – execution may be deferred by save – execution may be rejected by input p raise deny • exceptions raised by the remote procedure are raised at client and server side 69 © Humboldt-Universität zu Berlin SAM 2000, Grenoble block client Initial eh block server P Make Call Initial set(10,t) P to server P raises timer t ServerStart reject State2 Initial ´initialization´ eh State2 reject t initial State2 70 © Humboldt-Universität zu Berlin 35
  • 36. SAM 2000, Grenoble RPSRVLWH6WDWHV • composite states are a means to hierarchically structure state machines – nesting of states – agent can be in more than one state at a time – Harel´s state charts • composite state is itself a sub-state machine • state machine of an agent is in fact a top- level composite state 71 © Humboldt-Universität zu Berlin SAM 2000, Grenoble State1 St1 Sig3 Sig4 State1 Sig1 Sig2 State2 P1 P1 State2 State2 St1 via P1 Sig3 ex ex B State1 ex 72 © Humboldt-Universität zu Berlin 36
  • 37. SAM 2000, Grenoble • composite states share agent´s input queue • internal transitions with the same trigger as external transitions have higher priority • excactly one transition is executed – possibly concatenated with triggerless transitions • special procedures may be used to define initialisation and finalisation – called implicitly upon entering/leaving a composite state 73 © Humboldt-Universität zu Berlin SAM 2000, Grenoble initialization/finalization State1 entry exit St1 St1 Sig3 Sig4 St1_1 B State1 Sig1 Sig2 Sig3 - -* State2 history nextstate 74 © Humboldt-Universität zu Berlin 37
  • 38. SAM 2000, Grenoble 6WDWH$JJUHJDWLRQ • state aggregations partition the state space of an agents state machine • each partition handles a different set of the input stimuli • excactly one partition is executing a transition at any point in time – multiple enabled transitions are executed in an interleaved manner 75 © Humboldt-Universität zu Berlin SAM 2000, Grenoble State1 state aggregation st State1 St1 Sig1 Sig2 State2 State2 st_i St1 st Sig4 Sig3 st1 Sig1 st_i B st2 76 © Humboldt-Universität zu Berlin 38
  • 39. SAM 2000, Grenoble • composite states and state aggregations can be classified – composite state type definition – typebased composite states/state aggregations • concept notation similar to agent types • instances are static – livedie with containing agent – multiple instances in the same scope must have different names 77 © Humboldt-Universität zu Berlin SAM 2000, Grenoble 9LUWXDO%HKDYLRXU(OHPHQWV • allow the redefinition or replacement of behaviour elements in a type specialisation • redefinition and finalisation similar to structural elements • available for – procedures – transitions – exception handle transitions 78 © Humboldt-Universität zu Berlin 39
  • 40. SAM 2000, Grenoble ,QWHUIDFH • pure typing concept used for typed communication between agents • interface definition groups and names a set of – remote variable – remote procedure – signal definitions • gates and channels paths can be typed by interfaces 79 © Humboldt-Universität zu Berlin SAM 2000, Grenoble • contained remote procedures variables and signals are defined and visble where interface is visible • interface can use also existing definitions for such elements • multiple inheritance is available for interfaces • interfaces are used as matching constraints for implicit channels 80 © Humboldt-Universität zu Berlin 40
  • 41. SAM 2000, Grenoble interface if1; interface if3 signal sig1; inherits if1,if2; procedure P; dcl I Natural; endinterface; [if1] [if2] block type b signal sig2,sig3; interface if2; g use sig2, sig3; if3 endinterface; 81 © Humboldt-Universität zu Berlin SAM 2000, Grenoble $JHQW,PSOLFLW,QWHUIDFH • each agent and agent type introduces an implicit interface – same name as agent (type) • contains all – signals accepted by the agents state machine – remote variables/procedures provided by agents state machine • inherits all interfaces on gates connected to agents state machine 82 © Humboldt-Universität zu Berlin 41
  • 42. SAM 2000, Grenoble block type B g1 c1 B interface B [sig1] [sig2] [sig4] use sig2,sig4; c4 c2 endinterface g2 [sig3,sig4] g3 B2 c3 block type C interface C g1 inherits B inherits B, if1; [sig3] use sig3; B if1 endinterface 83 © Humboldt-Universität zu Berlin SAM 2000, Grenoble • each (explicit or implicit) interface implies a specialization of the Pid-type • can be used to refer to agent instances in – variables, expressions – output, imports or remote procedure calls – assignment attempt • dynamic type check tests – provision of remote variables/procedures – acceptance of signals – may raise an exception (InvalidReference) 84 © Humboldt-Universität zu Berlin 42
  • 43. SAM 2000, Grenoble block B assignment attempt dcl myb2 B2, mybt BT; g1 c1 B [sig1] [sig2] c4 B2 c2 mybt:=offspring g2 g B2:BT [sig3,sig4] c3 Sig3(x, y) to mybt State2 © Humboldt-Universität zu Berlin compatibility check 85 SAM 2000, Grenoble 'DWD7SHV • two main kinds of data types – value types – object types (references) conversion possible • data type definition specifies – data elements and structure – operators and methods for data manipulation • package Predefined contains a set of general data type definitions 86 © Humboldt-Universität zu Berlin 43
  • 44. SAM 2000, Grenoble • value types correspond to the newtype concept of SDL-92 – axioms and generators have been removed – behaviour of operators defined by algorithms or transition actions (functional description) – constructors: – literals – structs – choices 87 © Humboldt-Universität zu Berlin SAM 2000, Grenoble • object types define references to values – references are local to agents • definition similar to value types – conversion possible • polymorphic assignments 88 © Humboldt-Universität zu Berlin 44
  • 45. SAM 2000, Grenoble • further properties of data types: • inheritance • context parameters • methods allow operation-calls in programming- language like dot-notation • local data types, constants, exceptions • visibility of data elements can be controlled 89 © Humboldt-Universität zu Berlin SAM 2000, Grenoble REMHFW WSH /LVW WSH (OHP! VWUXFW HOHP (OHP SULYDWH QH[W /LVW RSHUDWRUV 0DNH(OHP
  • 46. !/LVW PHWKRGV DGG(OHP
  • 47. ! /LVW WKLV HOHP HOHP HOHP QH[W QH[W QH[W 1XOO HQGREMHFW WSH 90 © Humboldt-Universität zu Berlin 45
  • 48. SAM 2000, Grenoble REMHFW WSH /LVW WSH (OHP! WKLV HOHP HOHP HOHP QH[W QH[W QH[W 1XOO RSHUDWRU 0DNHH (OHP
  • 49. ^ UHWXUQ H1XOO
  • 50. ` PHWKRG DGGH (OHP
  • 51. ^ GFO ODVW (OHP IRU ODVW WKLVODVWQH[W 1XOOODVWQH[W
  • 52. ODVWQH[W H
  • 53. ` HQGREMHFW WSH object type NatList inherits ListNatural; dcl myList NatList; myList:=Make(5); myList.add(2); 91 © Humboldt-Universität zu Berlin SAM 2000, Grenoble Package Predefined • Boolean, Integer, Natural, Real, Character, Duration, Time • Charstring, Bit, Bitstring, simple types • Octet, Octetstring • Time, Duration, Pid • String, Powerset, Bag, data templates • Array, Vector • OutOfRange, InvalidReference, exceptions • NoMatchingAnswer, UndefinedVariable, • UndefinedField, InvalidIndex, • DivisionByZero, Empty 92 © Humboldt-Universität zu Berlin 46
  • 54. SAM 2000, Grenoble 'DWDDVVLJQPHQWV value types object types • strong type check • references local to agent • polymorphic assignments • assignment restrictions for specialised types for specialised types • virtual methods • assignment attempts • object creation • value extraction 93 © Humboldt-Universität zu Berlin SAM 2000, Grenoble YDOXH WSH KDUVWULQJ GFO YDOBYDU KDUVWULQJ SUHGHILQHG HQGYDOXH WSH UHIBYDU REMHFW KDUVWULQJ UHIBYDU 0DNH +HOOR ZRUOG
  • 55. FUHDWHV D UHIHUHQFH WR DQ REMHFW YDOBYDU UHIBYDU YDOXH H[WUDFWLRQ FUHDWHV D UHIHUHUHQFH DQG FRSLHV UHIBYDU YDOBYDU WKH YDOXH LQWR WKH REMHFW FORQH
  • 56. YDOBYDU IRR YDOXH DVVLJQPHQW UHIBYDU 1XOO H[FHSWLRQ ,QYDOLG5HIHUHQFH YDOBYDU UHIBYDU 94 © Humboldt-Universität zu Berlin 47
  • 57. SAM 2000, Grenoble object type MyStruct; dcl m MyStruct, struct n Natural; m1 MyStruct1, y Boolean optional; m2 MyStruct2; endobject type; object type MyStruct1 m1:= (. 5, true,’c’); inherits MyStruct adding m2:= (. 5,,’ABC’); c Character; m:= m1; assignment endobject type; m:=m2; m2:=Null; object type MyStruct2 assignment inherits MyStruct adding m1:=m; attempt s Charstring m2:=m; endobject type; m2:=m1; static error m1 m m2 (null) (null) 5 5 true * ‘c’ ‘ABC’ 95 © Humboldt-Universität zu Berlin SAM 2000, Grenoble • each data type (obejct or value) implicitly inherits from Any: – operators equal and clone – methods copy and is_equal • each object type has additionally the operations – Make and Null • each struct type has for each file the methods – fieldModify, fieldExtract, – fieldPresent (optional fields only) 96 © Humboldt-Universität zu Berlin 48
  • 58. SAM 2000, Grenoble 'HYHORSPHQWRI5HDO7LPH 6VWHPVZLWK6'/ communication and external procedures,synonyms, operators or methods executable SDL spec non-SDL user-defined application or components pre-defined virtual SDL machine exceptions target platform 97 © Humboldt-Universität zu Berlin SAM 2000, Grenoble )XUWKHU,QIRUPDWLRQ • tutorial slides and author contact www.informatik.hu-berlin.de/Institut/struktur/systemanalyse/ {fischer|holz}@informatik.hu-berlin.de • ITU standards and recommendations – www.itu.ch or www.itu.int/itudoc/itu-t/approved/z/index.html • SDL Forum Society – www.sdl-forum.org • conferences and workshops – bi-annual SDL Forum - next Copenhagen 2001 98 © Humboldt-Universität zu Berlin 49
  • 59. SAM 2000, Grenoble block type Telephone virtual Terminal Telephone: [PBX] [Phone] Terminal Keys:Keypad T Handset block type FaxPhone process type FaxCodec inherits Telephone redefined Terminal Sender Telephone: Terminal FaxCodec [data] [data] S:Scanner SendReceive: Receiver FaxCodec P:Printer 99 © Humboldt-Universität zu Berlin 50