Have I understood correctly that in (most? some?) multiple dispatch languages each method gets added to the function at some point in time of program's execution.
Can I then conclude that multiple dispatch as a feature forces functions to be mutable?
Is there a multiple dispatch language, where all methods are attached to a (generic)function together (at load time?), so that it's not possible to see the function in different states at different points in time?
at some point in time of program's execution.
In Common Lisp the methods get added/replaced when the method definitions are executed - for a compiled system this is typically at load-time of the compiled code - not necessarily during the program's execution.
Remember, that Common Lisp has an object system (CLOS, the Common Lisp Object System), which is defined by its behaviour. It's slightly different from a language or a language extension.
Common Lisp allows runtime modification of the object system. For example also adding/removing/replacing methods.
Common Lisp also may combine more than one applicable method into an effective method, which then gets executed. Typical example: all applicable :before methods and the most specific applicable primary method will be combined into one effective method.
There exist extensions for CLOS in some implementations, which seal a generic function against changes.
For a longer treatment of the idea of an object system see: The Structure of a Programming Language Revolution by Richard P. Gabriel.
In Common Lisp, you can read the following from the specification:
7.6.1 Introduction to Generic Functions
When a defgeneric form is evaluated, one of three actions is taken (due to ensure-generic-function):
If a generic function of the given name already exists, the existing generic function object is modified. Methods specified by the current defgeneric form are added, and any methods in the existing generic function that were defined by a previous defgeneric form are removed. Methods added by the current defgeneric form might replace methods defined by defmethod, defclass, define-condition, or defstruct. No other methods in the generic function are affected or replaced.
If the given name names an ordinary function, a macro, or a special operator, an error is signaled.
Otherwise a generic function is created with the methods specified by the method definitions in the defgeneric form.
7.6.2 Introduction to Methods
When a method-defining form is evaluated, a method object is created and one of four actions is taken:
If a generic function of the given name already exists and if a method object already exists that agrees with the new one on parameter specializers and qualifiers, the new method object replaces the old one. For a definition of one method agreeing with another on parameter specializers and qualifiers, see Section 7.6.3 (Agreement on Parameter Specializers and Qualifiers).
If a generic function of the given name already exists and if there is no method object that agrees with the new one on parameter specializers and qualifiers, the existing generic function object is modified to contain the new method object.
If the given name names an ordinary function, a macro, or a special operator, an error is signaled.
Otherwise a generic function is created with the method specified by the method-defining form.
The definition of ensure-generic-function:
If function-name specifies a generic function that has a different value for the :lambda-list argument, and the new value is congruent with the lambda lists of all existing methods or there are no methods, the value is changed; otherwise an error is signaled.
If function-name specifies a generic function that has a different value for the :generic-function-class argument and if the new generic function class is compatible with the old, change-class is called to change the class of the generic function; otherwise an error is signaled.
If function-name specifies a generic function that has a different value for the :method-class argument, the value is changed, but any existing methods are not changed.
You also have add-method and remove-method.
As you can see, generic functions retain their identify between defmethod definitions, and even between defgeneric definitions. Generic functions are mutable in Common Lisp.
In Julia, you can read the following from the documentation:
Defining Methods
To define a function with multiple methods, one simply defines the function multiple times, with different numbers and types of arguments. The first method definition for a function creates the function object, and subsequent method definitions add new methods to the existing function object.
As you can see, functions objects are mutable in Julia.
This says nothing about all other multiple dispatch languages. You can invent a multiple dispatch language right now just for the purpose of showing you can do it with immutability, e.g. adding methods would return a new function similar to the previous function but with the added method. Or a language where functions are generated statically at compile-time, such that you can't change it at runtime in any way, not even to add or remove methods.
Paraphrasing from the excellent "Getting started with Julia" book which has a nice section on this (emphasis mine):
We already saw that functions are inherently defined as generic, that is, they can be used for different types of their arguments. The compiler will generate a separate version of the function each time it is called with arguments of a new type. A concrete version of a function for a specific combination of argument types is called a method in Julia. To define a new method for a function (also called overloading), just use the same function name but a different signature, that is, with different argument types.
A list of all the methods is stored in a virtual method table ( vtable ) on the function itself; methods do not belong to a particular type. When a function is called, Julia will do a lookup in that vtable at runtime to find which concrete method it should call based on the types of all its arguments; this is Julia's mechanism of multiple dispatch, which neither Python, nor C++ or Fortran implements. It allows open extensions where normal object-oriented code would have forced you to change a class or subclass an existing class and thus change your library. Note that only the positional arguments are taken into account for multiple dispatch, and not the keyword arguments.
For each of these different methods, specialized low-level code is generated, targeted to the processor's instruction set. In contrast to object-oriented (OO) languages, vtable is stored in the function, and not in the type (or class). In OO languages, a method is called on a single object, object.method(), which is generally called single dispatch. In Julia, one can say that a function belongs to multiple types, or that a function is specialized or overloaded for different types. Julia's ability to compile code that reads like a high-level dynamic language into machine code that performs like C almost entirely is derived from its ability to do multiple dispatch.
So, the way I understand this (I may be wrong) is that:
The generic function needs to be defined in the session before you can use it
Explicitly defined methods for concrete arguments are added to the function's multiple-dispatch lookup table at the point where they're defined.
Whenever a function is called with specific arguments for which an explicitly defined method does not exist, a concrete version for those arguments is compiled and added to the vtable. (however, this does not show up as an explicit method if you run methods() on that function name)
The first call of such a function will result in some compilation overhead; however, subsequent calls will use the existing compiled version*.
I wouldn't say this makes functions mutable though, that's an altogether different issue. You can confirm yourself they're immutable using the isimmutable() function on a function 'handle'.
*I know modules can be precompiled, but I am not entirely sure if these on-the-fly compiled versions are saved between sessions in any form -- comments welcome :)
Dynamicity can be a real asset in your application, if only for debugging. Trying to prevent a function from being later updated, redefined, etc. might be a little bit short-sighted. But if you are sure you want static dispatch, you can define your own class of generic functions thanks to the MOP, the Meta-Object Protocol, which is not part of the standard but still largely supported. That's what the Inlined-Generic-Function library provides (and this is possible because CLOS is open to extensions).
In Dylan, methods are generally added to a generic function at compile time, but they may also be added at run time (via add-method or remove-method). However, a generic function may be sealed, which prevents libraries other than the one in which the g.f. is defined from adding methods. So to answer your question, in Dylan generic functions are always mutable within the defining library but they may be rendered immutable to other libraries.
Related
I have seen and heard many times that Julia allows "multiple dispatch", but I am not really sure what that means or looks like. Can anyone provide me an example of what it looks like programmatically and what it enables?
From the Julia docs
The choice of which method to execute when a function is applied is called dispatch. Julia allows the dispatch process to choose which of a function's methods to call based on the number of arguments given, and on the types of all of the function's arguments. This is different than traditional object-oriented languages, where dispatch occurs based only on the first argument, which often has a special argument syntax, and is sometimes implied rather than explicitly written as an argument. 1 Using all of a function's arguments to choose which method should be invoked, rather than just the first, is known as multiple dispatch. Multiple dispatch is particularly useful for mathematical code, where it makes little sense to artificially deem the operations to "belong" to one argument more than any of the others: does the addition operation in x + y belong to x any more than it does to y? The implementation of a mathematical operator generally depends on the types of all of its arguments. Even beyond mathematical operations, however, multiple dispatch ends up being a powerful and convenient paradigm for structuring and organizing programs.
So in short: other languages rely on the first parameter of a method in order to determine which method should be called whereas in Julia, multiple parameters are taken into account. This enables multiple definitions of similar functions that have the same initial parameter.
A simple example of multiple dispatch in Julia can be found here.
I know that from here:
Julia function arguments follow a convention sometimes called "pass-by-sharing", which means that values are not copied when they are passed to functions. Function arguments themselves act as new variable bindings (new locations that can refer to values), but the values they refer to are identical to the passed values. Modifications to mutable values (such as Arrays) made within a function will be visible to the caller. This is the same behavior found in Scheme, most Lisps, Python, Ruby and Perl, among other dynamic languages.
Given this, it's clear to me that to pass by reference, all you need to do is have a mutable type that you pass into a function and edit.
My question then becomes, how can I clearly distinguish between pass by value and pass by reference? Does anyone have an example that shows a function being called twice; once with pass by reference, and once with pass by value?
I saw this post which alludes to some similar ideas, but it did not fully answer my question.
In Julia, functions always have pass-by-sharing argument-passing behavior:
https://docs.julialang.org/en/v1/manual/functions/
This argument-passing convention is also used in most general purpose dynamic programming languages, including various Lisps, Python, Perl and Ruby. A good and useful description can be found here:
https://en.wikipedia.org/wiki/Evaluation_strategy#Call_by_sharing
In short, pass-by-sharing works like pass-by-reference but you cannot change which value a binding in the calling scope refers to by reassigning to an argument in the function being called—if you reassign an argument, the binding in the caller is unchanged. This means that in general you cannot use functions to change bindings, such as for example to swap to variables. (Macros can, however, modify bindings in the caller.) In particular, if a variable in the caller refers to an immutable value like an integer or a floating-point number, its value cannot be changed by a function call since which object the variable refers to cannot be changed by a function call and the value itself cannot be modified as it is immutable.
If you want to have something like R or Matlab pass by value behavior, you need to explicitly create a copy of the argument before modifying it. This is precisely what R and Matlab do when an argument is passed in a modified and an external reference to the argument remains. In Julia it must be done explicitly by the programmer rather than being done automatically by the system. A downside is that the system can sometimes know that no copy is required (no external references remain) when the programmer cannot generally know this. That ability, however, is deeply tied with the reference counting garbage collections technique, which is not used by Julia due to performance considerations.
By convention, functions which mutate the contents of an argument have a ! postfix (e.g., sort v/s sort!).
What is the function of new() in Julia? Is this question even specific enough?
I am looking through the module Mocha where new(...) is used quite commonly, but I don't see any definition of new(), only uses of it, nor do I find reference to it in the Julia documentation.
I thought it might then be defined in a module that is being used by Mocha, but then I would think I could learn about new() with Mocha.new from the REPL, but that comes back with ERROR: UndefVarError: new not defined.
For the life of me I can't figure out what new(...) is doing. If it doesn't sound like something common to Julia, what can I do to try to track down where it's defined?
From http://docs.julialang.org/en/release-0.4/manual/constructors/
Inner Constructor Methods
While outer constructor methods succeed in
addressing the problem of providing additional convenience methods for
constructing objects, they fail to address the other two use cases
mentioned in the introduction of this chapter: enforcing invariants,
and allowing construction of self-referential objects. For these
problems, one needs inner constructor methods. An inner constructor
method is much like an outer constructor method, with two differences:
It is declared inside the block of a type declaration, rather than
outside of it like normal methods.
It has access to a special locally existent function called new that creates objects of the block’s type.
I'm trying to simulate something akin to Haskell's typeclasses with Common Lisp's CLOS. That is, I'd like to be able to dispatch a method on an object's "typeclasses" instead of its superclasses.
I have a metaclass defined for classes which have and implement typeclasses(which are just other classes). Those classes(those that implement typeclasses) have a slot containing the list of the typeclasses they implement.
I'd like to be able to define methods for a typeclass, and then be able to dispatch that method on objects whose class implement that typeclass. And I'd like to be able to add and remove typeclasses dynamically.
I figure I could probably do this by changing the method dispatch algorithm, though that doesn't seem too simple.
Anybody is comfortable enough with CLOS and the MOP to give me some suggestions?
Thanks.
Edit: My question might be specified as, how do I implement compute-applicable-methods-using-classes and compute-applicable-methods for a "custom" generic-function class such that if some of the specializers of a generic function method are typeclasses(classes whose metaclass is the 'typeclass' class), then the corresponding argument's class must implement the typeclass(which simply means having the typeclass stored in a slot of the argument's class) for the method to be applicable?
From what I understand from documentation, when a generic function is called, compute-discriminating-functionis first called, which will first attempt to obtain applicable methods through compute-applicable-methods-using-classes, and if unsuccessful, will try the same with compute-applicable-methods.
While my definition of compute-applicable-methods-using-classes seems to work, the generic function fails to dispatch an applicable function. So the problem must be in compute-discriminating-function or compute-effective-method.
See code.
This is not easily achievable in Common Lisp.
In Common Lisp, operations (generic functions) are separate from types (classes), i.e. they're not "owned" by types. Their dispatch is done at runtime, with the possibility of adding, modifying and removing methods at runtime as well.
Usually, errors from missing methods are signaled only at runtime. The compiler has no way to know if a generic function is being "well" used or not.
The idiomatic way in Common Lisp is to use generic functions and describe its requirements, or in other words, the closest to an interface in Common Lisp is a set of generic functions and a marker mixin class. But most usually, only a protocol is specified, and its dependencies on other protocols. See, for instance, the CLIM specification.
As for type classes, it's a key feature that keeps the language not only fully type-safe, but also makes it very extensible in that aspect. Otherwise, either the type system would be too strict, or the lack of expressiveness would lead to type-unsafe situations, at least from the compiler's point of view. Note that Haskell doesn't keep, or doesn't have to keep, object types at runtime, it takes every type inference at compile-time, much in contrast with idiomatic Common Lisp.
To have something similar to type classes in Common Lisp at runtime, you have a few choices
Should you choose to support type classes with its rules, I suggest you use the meta-object protocol:
Define a new generic function meta-class (i.e. one which inherits from standard-generic-function)
Specialize compute-applicable-methods-using-classes to return false as a second value, because classes in Common Lisp are represented solely by their name, they're not "parameterizable" or "constrainable"
Specialize compute-applicable-methods to inspect the argument's meta-classes for types or rules, dispatch accordingly and possibly memoize results
Should you choose to only have parameterizable types (e.g. templates, generics), an existing option is the Lisp Interface Library, where you pass around an object that implements a particular strategy using a protocol. However, I see this mostly as an implementation of the strategy pattern, or an explicit inversion of control, rather than actual parameterizable types.
For actual parameterizable types, you could define abstract unparameterized classes from which you'd intern concrete instances with funny names, e.g. lib1:collection<lib2:object>, where collection is the abstract class defined in the lib1 package, and the lib2:object is actually part of the name as is for a concrete class.
The benefit of this last approach is that you could use these classes and names anywhere in CLOS.
The main disadvantage is that you must still generate concrete classes, so you'd probably have your own defmethod-like macro that would expand into code that uses a find-class-like function which knows how to do this. Thus breaking a significant part of the benefit I just mentioned, or otherwise you should follow the discipline of defining every concrete class in your own library before using them as specializers.
Another disadvantage is that without further non-trivial plumbing, this is too static, not really generic, as it doesn't take into account that e.g. lib1:collection<lib2:subobject> could be a subclass of lib1:collection<lib2:object> or vice-versa. Generically, it doesn't take into account what is known in computer science as covariance and contravariance.
But you could implement it: lib:collection<in out> could represent the abstract class with one contravariant argument and one covariant argument. The hard part would be generating and maintaining the relationships between concrete classes, if at all possible.
In general, a compile-time approach would be more appropriate at the Lisp implementation level. Such Lisp would most probably not be a Common Lisp. One thing you could do is to have a Lisp-like syntax for Haskell. The full meta-circle of it would be to make it totally type-safe at the macro-expansion level, e.g. generating compile-time type errors for macros themselves instead of only for the code they generate.
EDIT: After your question's edit, I must say that compute-applicable-methods-using-classes must return nil as a second value whenever there is a type class specializer in a method. You can call-next-method otherwise.
This is different than there being a type class specializer in an applicable method. Remember that CLOS doesn't know anything about type classes, so by returning something from c-a-m-u-c with a true second value, you're saying it's OK to memoize (cache) given the class alone.
You must really specialize compute-applicable-methods for proper type class dispatching. If there is opportunity for memoization (caching), you must do so yourself here.
I believe you'll need to override compute-applicable-methods and/or compute-applicable-methods-using-classes which compute the list of methods that will be needed to implement a generic function call. You'll then likely need to override compute-effective-method which combines that list and a few other things into a function which can be called at runtime to perform the method call.
I really recommend reading The Art of the Metaobject Protocol (as was already mentioned) which goes into great detail about this. To summarize, however, assume you have a method foo defined on some classes (the classes need not be related in any way). Evaluating the lisp code (foo obj) calls the function returned by compute-effective-method which examines the arguments in order to determine which methods to call, and then calls them. The purpose of compute-effective-method is to eliminate as much of the run-time cost of this as is possible, by compiling the type tests into a case statement or other conditional. The Lisp runtime thus does not have to query for the list of all methods each time you make a method call, but only when you add, remove or change a method implementation. Usually all of that is done once at load time and then saved into your lisp image for even better performance while still allowing you to change these things without stopping the system.
In Ada, Primitive operations of a type T can only be defined in the package where T is defined. For example, if a Vehicules package defines Car and Bike tagged record, both inheriting a common Vehicle abstract tagged type, then all operations than can dispatch on the class-wide Vehicle'Class type must be defined in this Vehicles package.
Let's say that you do not want to add primitive operations: you do not have the permission to edit the source file, or you do not want to clutter the package with unrelated features.
Then, you cannot define operations in other packages that implicitely dispatches on type Vehicle'Class.
For example, you may want to serialize vehicles (define a Vehicles_XML package with a To_Xml dispatching function) or display them as UI elements (define a Vehicles_GTK package with Get_Label, Get_Icon, ... dispatching functions), etc.
The only way to perform dynamic dispatch is to write the code explicitely; for example, inside Vechicle_XML:
if V in Car'Class then
return Car_XML (Car (V));
else
if V in Bike'Class then
return Bike_XML (Bike (V));
else
raise Constraint_Error
with "Vehicle_XML is only defined for Car and Bike."
end if;
(And a Visitor pattern defined in Vehicles and used elsewhere would work, of course, but that still requires the same kind of explicit dispatching code. edit in fact, no, but there is still some boilerplate code to write)
My question is then:
is there a reason why operations dynamically dispatching on T are restricted to be defined in the defining package of T?
Is this intentional? Is there some historical reasons behind this?
Thanks
EDIT:
Thanks for the current answers: basically, it seems that it is a matter of language implementation (freezing rules/virtual tables).
I agree that compilers are developped incrementally over time and that not all features fit nicely in an existing tool.
As such, isolating dispatching operators in a unique package seems to be a decision mostly guided by existing implementations than by language design. Other languages outside of the C++/Java family provide dynamic dispatch without such requirement (e.g. OCaml, Lisp (CLOS); if that matters, those are also compiled languages, or more precisely, language for which compilers exist).
When I asked this question, I wanted to know if there were more fundamental reasons, at language specification level, behind this part of Ada specifications (otherwise, does it really mean that the specification assumes/enforces a particular implementation of dynamic disapatch?)
Ideally, I am looking for an authoritative source, like a rationale or guideline section in Reference Manuals, or any kind of archived discussion about this specific part of the language.
I can think of several reasons:
(1) Your example has Car and Bike defined in the same package, both derived from Vehicles. However, that's not the "normal" use case, in my experience; it's more common to define each derived type in its own package. (Which I think is close to how "classes" are used in other compiled languages.) And note also that it's not uncommon to define new derived types afterwards. That's one of the whole points of object-oriented programming, to facilitate reuse; and it's a good thing if, when designing a new feature, you can find some existing type that you can derive from, and reuse its features.
So suppose you have your Vehicles package that defines Vehicle, Car, and Bike. Now in some other package V2, you want to define a new dispatching operation on a Vehicle. For this to work, you have to provide the overriding operations for Car and Bike, with their bodies; and assuming you are not allowed to modify Vehicles, then the language designers have to decide where the bodies of the new operation have to be. Presumably, you'd have to write them in V2. (One consequence is that the body that you write in V2 would not have access to the private part of Vehicles, and therefore it couldn't access implementation details of Car or Bike; so you could only write the body of that operation if terms of already-defined operations.) So then the question is: does V2 need to provide operations for all types that are derived from Vehicle? What about types derived from Vehicle that don't become part of the final program (maybe they're derived to be used in someone else's project)? What about types derived from Vehicle that haven't yet been defined (see preceding paragraph)? In theory, I suppose this could be made to work by checking everything at link time. However, that would be a major paradigm change for the language. It's not something that could be easily. (It's pretty common, by the way, for programmers to think "it would be nice to add feature X to a language, and it shouldn't be too hard because X is simple to talk about", without realizing just what a vast impact such a "simple" feature would have.)
(2) A practical reason has to do with how dispatching is implemented. Typically, it's done with a vector of procedure/function pointers. (I don't know for sure what the exact implementation is in all cases, but I think this is basically the case for every Ada compiler as well as for C++ and Java compilers, and probably C#.) What this means is that when you define a tagged type (or a class, in other languages), the compiler will set up a vector of pointers, and based on how many operations are defined for the type, say N, it will reserve slots 1..N in the vector for the addresses of the subprograms. If a type is derived from that type and defines overriding subprograms, the derived type gets its own vector, where slots 1..N will be pointers to the actual overriding subprograms. Then, when calling a dispatching subprogram, a program can look up the address in some known slot index assigned to that subprogram, and it will jump to the correct address depending on the object's actual type. If a derived type defines new primitive subprograms, new slots are assigned N+1..N2, and types derived from that could define new subprograms that get slots N2+1..N3, and so on.
Adding new dispatching subprograms to Vehicle would interfere with this. Since new types have been derived from Vehicle, you can't insert a new area into the vector after N, because code has already been generated that assumes the slots starting at N+1 have been assigned to new operations derived for derived types. And since we may not know all the types that have been derived from Vehicle and we don't know what other types will be derived from Vehicle in the future and how many new operations will be defined for them, it's hard to pick some other location in the vector that could be used for the new operations. Again, this could be done if all of the slot assignment were deferred until link time, but that would be a major paradigm change, again.
To be honest, I can think of other ways to make this work, by adding new operations not in the "main" dispatch vector but in an auxiliary one; dispatching would probably require a search for the correct vector (perhaps using an ID assigned to the package that defines the new operations). Also, adding interface types to Ada 2005 has already complicated the simple vector implementation somewhat. But I do think this (i.e. it doesn't fit into the model) is one reason why the ability to add new dispatching operations like you suggest isn't present in Ada (or in any other compiled language that I know of).
Without having checked the rationale for Ada 95 (where tagged types were introduced), I am pretty sure the freezing rules for tagged types are derived from the simple requirement that all objects in T'Class should have all the dispatching operations of type T.
To fulfill that requirement, you have to freeze type and say that no more dispatching operations can be added to type T once you:
Derive a type from T, or
Are at the end of the package specification where T was declared.
If you didn't do that, you could have a type derived from type T (i.e. in T'Class), which hadn't inherited all the dispatching operations of type T. If you passed an object of that type as a T'Class parameter to a subprogram, which knew of one more dispatching operation on type T, a call to that operation would have to fail. - We wouldn't want that to happen.
Answering your extended question:
Ada comes with both a Reference Manual (the ISO standard), a Rationale and an Annotated Reference Manual. And a large part of the discussions behind these documents are public as well.
For Ada 2012 see http://www.adaic.org/ada-resources/standards/ada12/
Tagged types (dynamic dispatching) was introduced in Ada 95. The documents related to that version of the standard can be found at http://www.adaic.org/ada-resources/standards/ada-95-documents/