1. Language features for concurrency
   1. Semaphores
   2. Monitors
   3. Message Passing
2. Programming Language Selection
   1. Strategic issues in choosing a language
   2. Criteria for Language Design
      1. Writeability
         1. Simplicity
         2. Uniformity:
         3. Expressiveness
         4. Orthogonality & generality
         5. Clear, unambiguous syntactic and semantic description
      2. Readability - Modifiability and Maintenance.
         1. Documentation:
         2. Lexical and syntactic conventions:
         3. Not synonymous with wordiness:
         4. Importance depends on number of programmers.
      3. Reliability
         1. Static checks
         2. Separate (but not independent) compilation
         3. Minor modifications should not require major changes.
         4. Clear semantics supporting verification.
         5. Simplicity of compiler implementation.
      4. Fast translation
      5. Efficient object code
      6. Machine independence:
   3. Hoare on Designing Programming Languages
      1. Features:
      2. Language:

Language features for concurrency

1. Semaphores (for mutual exclusion)

2. Monitors (for mutual exclusion)

3. Message passing (using "tasks")

Focus here on producer/consumer or bounded buffer problem.
Two processes cooperating, one by adding items to a buffer, the other removing items. Ensure not remove when nothing there and not overflow buffer as well.

Text also focuses on parallel matrix multiplication (read on own).

Text also discusses some ways of handling with simple extensions of existing languages:

• Parallel for loops

• library routines for spawning and managing processes.

Coroutines also worth noting (part of Modula-2). Iterators were a special case.

Idea is co-equal routines which pass control back and forth.
E.g., our Modula-2 has library supporting routines:

    NewCoroutine(P: PROC; workspaceSize:CARDINAL; 
                                                        VAR q: COROUTINE);

        Transfer(VAR from, to: COROUTINE)

Can have multiple coroutines executing same procedure or can all be distinct.

Usually run on single processor.

Can think of as supporting multi-tasking. Good for writing operating systems.

See Modula-2 code in text for bounded buffer problem with coroutines.

Semaphores

Three operations:

    InitSem(S: Semaphore; value: integer);
    Wait(S: Semaphore);
    Signal(S: Semaphore);

InitSem starts up semaphore with an initial (non-negative) value.

Wait(S): If S > 0 then S := S - 1 else suspend self

Signal(S): if processes are waiting, then wake up a process, else S := S + 1;

Think of Wait(S) as claiming a resource so that no one else can get it, while Signal(S) releases the resource.

In order to solve mutual exclusion problem, must ensure that Wait and Signal execute atomically (i.e., cannot be interrupted and no one else can execute at same time).

If start w/S = 1 then protect a critical region by:

    Wait(S);    -- grab token
    {Critical region}
    Signal(S);  -- release token

Solution to bounded buffer:

Suppose also have procedures:

    CreateProcess(p:PROC; workspacesize: CARDINAL);
        Creates nameless process
    StartProcesses;  -- starts all processes which have been created.
    Terminate;  -- stop execution of process

Main program:
CreateProcess(Producer,WorkSize);   
                                        -- create at least one producer
CreateProcess(Consumer,WorkSize);   
                                        -- create at least one consumer
BufferStart := 1;  BufferEnd := 0
InitSem(NonEmpty, 0)    
            -- semaphore w/initial value of  0 to indicate empty
InitSem(NonFull, MaxBuffSize)   
            -- semaphore w/initial value of size of buffer
InitSem(MutEx,1)        -- semaphore used for mutual exclusion
StartProcesses
end;

Procedure Producer;
begin
    loop
        read(ch)
        Wait(NonFull);
        Wait(MutEx);
        BufferEnd := BufferEnd MOD MaxBuffSize + 1;
        Buffer[BufferEnd] := ch;
        Signal(MutEx);
        Signal(NonEmpty);
    end loop;
end;

Procedure Consumer;
begin
    loop
        Wait(NonEmpty);
        Wait(MutEx);
        ch := Buffer[BufferStart];
        BufferStart := BufferStart MOD MaxBuffSize + 1;
        Signal(MutEx);
        Signal(NonFull);
        Write(ch)
    end loop
end;

Why is there a MutEx semaphore?

Technically it is not necessary here since Producer only changes BufferEnd, while Consumer only changes BufferStart, but if they both changed a count of the number of items in the buffer would be important to keep them from executing at the same time!

What would go wrong if you changed the order of the two Wait's at the beginning of either Producer or Consumer?

Biggest problem with semaphores is that they are too low level and unstructured. Accidentally reversing order or Wait'ing twice could be disastrous.

Monitors

The idea is to provide an ADT with "condition variables", each of which has an associated queue of processes and suspend (or delay) and continue ops for en and dequeuing processes. Suspend always stops current, continue starts up new if any waiting.

Concurrent Pascal uses monitors.

• Continue by process in monitor causes it to be immediately suspended.

• Each queue can only hold one suspended process at a time. (owise error)

type buffer = monitor;

var store: array[1..MaxBuffSize] of char;
    BufferStart, BufferEnd, BufferSize: integer
    nonfull, nonempty: queue;

procedure entry insert(ch: char);
begin
    if BufferSize = MaxBuffSize then delay(nonfull);
    BufferEnd :=BufferEnd mod MaxBuffSize + 1;
    store[BufferEnd] := ch;
 BufferSize := BufferSize + 1;
   continue(nonempty)
end;

procedure entry delete(var ch: char);
begin
   if BufferSize = 0 then delay(nonempty);
 ch := store[BufferStart];
   BufferStart := BufferStart mod MaxBuffSize + 1;
    BufferSize := BufferSize -1;
    continue(nonfull);
end;

begin (* initialization *)
  BufferEnd := 0;
 BufferStart := 1;
   BufferSize := 0
end;

type producer = process (b: buffer);
var ch: char;
begin
    while true do begin
     read(ch);
        b.insert(ch)
    end;
end

type consumer = process(b: buffer);
var ch: char;
begin
    while true do begin
        b.delete(ch);
        write(ch)
    end
end;

var p: producer;
     q: consumer;
     b:buffer;
begin
    init b, p(b), q(b)
end.

Notice improved structure!

See text for simple way of emulating Semaphores w/ Monitors.

Message Passing

With distributed systems, cannot rely on shared variables for synchronization and communication.

With message passing, rely on send and receive operations:

    Procedure Send(To: Process; M: Message);
    Procedure Receive(From: Process; M: Message);

Synchronization questions:

• Does sender have to wait for receiver to accept message or keep in buffer?

• Does receiver have to wait for message or can just check to see if any?

Ada's tasks combine features of monitors and message passing.

Process in Ada is called a task (defined much like a package).

Task exports entry names to world (like a monitor) which can be "called".

• Entries can have parameters as well.

• Each entry has associated queue (managed as FIFO)

Select allows choice of which accept is selected.

Syntax:

    select
      [when <cond> =>] <select alternative>
        {or [when <cond> =>] <select alternative>}
       [else <statements>]
    end select

System determines which "when" conditions are true and select one which has an "accept" process waiting. If several than choose arbitrary one.

Caller is blocked only during body of "accept" clause.
Can have other statements outside the "accept".

Ex from Ada:

task Buffer is      -- specification
    entry insert(ch: in CHARACTER);
 entry delete(ch: out CHARACTER);
end;

task body Buffer is
 MaxBuffSize: constant INTEGER := 50;
    Store: array(1..MaxBuffSize) of CHARACTER;
    BufferStart: INTEGER := 1;
    BufferEnd: INTEGER := 0;
    BufferSize: INTEGER := 0;
begin
    loop
        select
            when BufferSize < MaxBufferSize =>
                accept insert(ch: in CHARACTER) do
                    BufferEnd := BufferEnd mod MaxBuffSize + 1;
                 Store(BufferEnd) := ch;
                end;
                BufferSize := BufferSize + 1;
            or when BufferSize > 0 =>
                accept delete(ch: out CHARACTER) do
                    ch := Store(BufferStart);
                end;
                BufferStart := BufferStart mod MaxBufferSize + 1;
                BufferSize := BufferSize -1;
        end select;
    end loop
end Buffer;

task type Producer;
task body Producer is
    ch: CHARACTER;
begin
    loop
        TEXT_IO.GET(ch);
        Buffer.insert(ch);
    end loop;
end Producer;

task type Consumer;
task body Consumeris
    ch: CHARACTER;
begin
    loop
        Buffer.delete(ch);
        TEXT_IO.PUT(ch);
    end loop;
end Consumer;

Note Producer/Consumer run forever - no entries.

Exit from task when at end and no open subprocesses or waiting with "terminate" alternative open and parent has executed to completion. If this true of all children of process then all terminate simultaneously.

Comparing semaphores and monitors and tasking:

Semaphores very low level.

Monitors are passive regions encapsulating resources to be shared (mutual exclusion). Cooperation enforced by delay and continue statements.

Everything active in Ada tasks (resources and processes)

Monitors and processes can easily be structured as Ada tasks and vice-versa.

Ada tasks better represents behavior of distributed system.

Programming Language Selection and Design

a. Intended application.

b. Intended programmers.

c. Intended computers to be used.

But many general principles!

Strategic issues in choosing a language

Support for:

1. Abstraction: What versus How

2. Programming in the Large: Can 100 people work together on a program?

3. Software reuse: Can you use old or produce new modules which can be used in other programs?

Criteria for Language Design

Writeability

Simplicity

- bad ex: PL/I & Ada - don't throw in kitchen sink!

If too complex, danger of unexpected interactions, subsetting & (if redundant) development of dialects.

Simple conceptual model of semantics.

Can go too far! E.g., Basic, LISP syntax.

"Everything should be as simple as possible, but not simpler" Einstein

Uniformity:

Bad example: subscript vs fcn def in FORTRAN.

Good example: Instance vble vs parameterless function in Eiffel

Expressiveness

Success depends on appl.

Orthogonality & generality

Ex. of probs w/ Pascal: Fcn can only return simple type, constants only defined for built-in basic types, enumerated types cannot be read or written, etc.

Algol 68 carries to extremes, everything has value.

Smalltalk, Eiffel, other pure object-oriented languages do it well.

Clear, unambiguous syntactic and semantic description

Original presentation usually too vague. Refined.

ALGOL 68 problem w/ redefining terms.

Readability - Modifiability and Maintenance.

Documentation:

commenting conventions: line vs stream oriented.

Lexical and syntactic conventions:

blocking of compound statements - if, fi versus begin, end

two-dimensional syntax, syntax directed editors.

Not synonymous with wordiness:

Importance depends on number of programmers.

Reliability

Related to writeability, readability, and modifiability.

Static checks
Separate (but not independent) compilation
Minor modifications should not require major changes.

Clear semantics supporting verification.

Simplicity of compiler implementation.

Practical Considerations

Fast translation

Efficient object code

Machine independence:

Hoare on Designing Programming Languages

Features:

Make sure solves problem, without causing new ones.

Show simply and efficiently implemented.

Carefully specify semantics

Use in lots of examples

Language:

Watch for inconsistencies, interactions.

(Knuth article on remaining problems in Algol 60: interactions major theme,
e.g. side effects and order of evaluation, unexpected exit from control structure, aliasing - 2 diff. ways of referring to vble, integers also labels)

Be especially wary of new features

Be ready to make modifications to solve minor problems.

Know intended applications, allowable size and complexity.

Implement on several machines, write manuals and texts

Be prepared to sell it to customers

Don't include untried ideas - consolidation, not innovation.

Make design group as small as possible.

Don't be afraid to revise - once distributed, too hard to change.

Simplicity is really the key - avoid complexity. Too many "solutions" make language hard to understand.

(See also Hoare, "The emperor's old clothes", CACM, Feb., 1981, pp. 75-83.)

Base programming language on:

1. a minimum number of independent concepts combined in a uniform manner.

2. a comprehensive definition mechanism to provide the breadth.

3. a small core language on which the extensions are based.

4. a syntax chosen for its readability.

Most dangerous approach to problems: add new features

Trap PL/I, Ada, and now C++ fell into.