Related
I was wondering if this is Standard, or a bug in my code. I'm trying to compare a pair of my homegrown function objects. I rejected the comparison if the type of function object is not the same, so I know that the two lambdas are the same type. So why can't they be compared?
Every C++0x lambda object has a distinct type, even if the signature is the same.
auto l1=[](){}; // one do-nothing lambda
auto l2=[](){}; // and another
l1=l2; // ERROR: l1 and l2 have distinct types
If two C++0x lambdas have the same type, they must therefore have come from the same line of code. Of course, if they capture variables then they won't necessarily be identical, as they may have come from different invocations.
However, a C++0x lambda does not have any comparison operators, so you cannot compare instances to see if they are indeed the same, or just the same type. This makes sense when you think about it: if the captured variables do not have comparison operators then you cannot compare lambdas of that type, since each copy may have different values for the captured variables.
Is the equality operator overloaded for your lambda object? If not I'm assuming you'll need to implement it.
I guess that most factory-like methods start with create. But why are they called "create"? Why not "make", "produce", "build", "generate" or something else? Is it only a matter of taste? A convention? Or is there a special meaning in "create"?
createURI(...)
makeURI(...)
produceURI(...)
buildURI(...)
generateURI(...)
Which one would you choose in general and why?
Some random thoughts:
'Create' fits the feature better than most other words. The next best word I can think of off the top of my head is 'Construct'. In the past, 'Alloc' (allocate) might have been used in similar situations, reflecting the greater emphasis on blocks of data than objects in languages like C.
'Create' is a short, simple word that has a clear intuitive meaning. In most cases people probably just pick it as the first, most obvious word that comes to mind when they wish to create something. It's a common naming convention, and "object creation" is a common way of describing the process of... creating objects.
'Construct' is close, but it is usually used to describe a specific stage in the process of creating an object (allocate/new, construct, initialise...)
'Build' and 'Make' are common terms for processes relating to compiling code, so have different connotations to programmers, implying a process that comprises many steps and possibly a lot of disk activity. However, the idea of a Factory "building" something is a sensible idea - especially in cases where a complex data-structure is built, or many separate pieces of information are combined in some way.
'Generate' to me implies a calculation which is used to produce a value from an input, such as generating a hash code or a random number.
'Produce', 'Generate', 'Construct' are longer to type/read than 'Create'. Historically programmers have favoured short names to reduce typing/reading.
Joshua Bloch in "Effective Java" suggests the following naming conventions
valueOf — Returns an instance that has, loosely speaking, the same value
as its parameters. Such static factories are effectively
type-conversion methods.
of — A concise alternative to valueOf, popularized by EnumSet (Item 32).
getInstance — Returns an instance that is described by the parameters
but cannot be said to have the same value. In the case of a singleton,
getInstance takes no parameters and returns the sole instance.
newInstance — Like getInstance, except that newInstance guarantees that
each instance returned is distinct from all others.
getType — Like getInstance, but used when the factory method is in a
different class. Type indicates the type of object returned by the
factory method.
newType — Like newInstance, but used when the factory method is in a
different class. Type indicates the type of object returned by the
factory method.
Wanted to add a couple of points I don't see in other answers.
Although traditionally 'Factory' means 'creates objects', I like to think of it more broadly as 'returns me an object that behaves as I expect'. I shouldn't always have to know whether it's a brand new object, in fact I might not care. So in suitable cases you might avoid a 'Create...' name, even if that's how you're implementing it right now.
Guava is a good repository of factory naming ideas. It is popularising a nice DSL style. examples:
Lists.newArrayListWithCapacity(100);
ImmutableList.of("Hello", "World");
"Create" and "make" are short, reasonably evocative, and not tied to other patterns in naming that I can think of. I've also seen both quite frequently and suspect they may be "de facto standards". I'd choose one and use it consistently at least within a project. (Looking at my own current project, I seem to use "make". I hope I'm consistent...)
Avoid "build" because it fits better with the Builder pattern and avoid "produce" because it evokes Producer/Consumer.
To really continue the metaphor of the "Factory" name for the pattern, I'd be tempted by "manufacture", but that's too long a word.
I think it stems from “to create an object”. However, in English, the word “create” is associated with the notion “to cause to come into being, as something unique that would not naturally evolve or that is not made by ordinary processes,” and “to evolve from one's own thought or imagination, as a work of art or an invention.” So it seems as “create” is not the proper word to use. “Make,” on the other hand, means “to bring into existence by shaping or changing material, combining parts, etc.” For example, you don’t create a dress, you make a dress (object). So, in my opinion, “make” by meaning “to produce; cause to exist or happen; bring about” is a far better word for factory methods.
Partly convention, partly semantics.
Factory methods (signalled by the traditional create) should invoke appropriate constructors. If I saw buildURI, I would assume that it involved some computation, or assembly from parts (and I would not think there was a factory involved). The first thing that I thought when I saw generateURI is making something random, like a new personalized download link. They are not all the same, different words evoke different meanings; but most of them are not conventionalised.
I'd call it UriFactory.Create()
Where,
UriFactory is the name of the class type which is provides method(s) that create Uri instances.
and Create() method is overloaded for as many as variations you have in your specs.
public static class UriFactory
{
//Default Creator
public static UriType Create()
{
}
//An overload for Create()
public static UriType Create(someArgs)
{
}
}
I like new. To me
var foo = newFoo();
reads better than
var foo = createFoo();
Translated to english we have foo is a new foo or foo is create foo. While I'm not a grammer expert I'm pretty sure the latter is grammatically incorrect.
I'd point out that I've seen all of the verbs but produce in use in some library or other, so I wouldn't call create being an universal convention.
Now, create does sound better to me, evokes the precise meaning of the action.
So yes, it is a matter of (literary) taste.
Personally I like instantiate and instantiateWith, but that's just because of my Unity and Objective C experiences. Naming conventions inside the Unity engine seem to revolve around the word instantiate to create an instance via a factory method, and Objective C seems to like with to indicate what the parameter/s are. This only really works well if the method is in the class that is going to be instantiated though (and in languages that allow constructor overloading, this isn't so much of a 'thing').
Just plain old Objective C's initWith is also a good'un!
I'm in the middle of reading Code Complete, and towards the end of the book, in the chapter about refactoring, the author lists a bunch of things you should do to improve the quality of your code while refactoring.
One of his points was to always return as specific types of data as possible, especially when returning collections, iterators etc. So, as I've understood it, instead of returning, say, Collection<String>, you should return HashSet<String>, if you use that data type inside the method.
This confuses me, because it sounds like he's encouraging people to break the rule of information hiding. Now, I understand this when talking about accessors, that's a clear cut case. But, when calculating and mangling data, and the level of abstraction of the method implies no direct data structure, I find it best to return as abstract a datatype as possible, as long as the data doesn't fall apart (I wouldn't return Object instead of Iterable<String>, for example).
So, my question is: is there a deeper philosophy behind Code Complete's advice of always returning as specific a data type as possible, and allow downcasting, instead of maintaining a need-to-know-basis, that I've just not understood?
I think it is simply wrong for the most cases. It has to be:
be as lenient as possible, be as specific as needed
In my opinion, you should always return List rather than LinkedList or ArrayList, because the difference is more an implementation detail and not a semantic one. The guys from the Google collections api for Java taking this one step further: they return (and expect) iterators where that's enough. But, they also recommend to return ImmutableList, -Set, -Map etc. where possible to show the caller he doesn't have to make a defensive copy.
Beside that, I think the performance of the different list implementations isn't the bottleneck for most applications.
Most of the time one should return an interface or perhaps an abstract type that represents the return value being returned. If you are returning a list of X, then use List. This ultimately provides maximum flexibility if the need arises to return the list type.
Maybe later you realise that you want to return a linked list or a readonly list etc. If you put a concrete type your stuck and its a pain to change. Using the interface solves this problem.
#Gishu
If your api requires that clients cast straight away most of the time your design is suckered. Why bother returning X if clients need to cast to Y.
Can't find any evidence to substantiate my claim but the idea/guideline seems to be:
Be as lenient as possible when accepting input. Choose a generalized type over a specialized type. This means clients can use your method with different specialized types. So an IEnumerable or an IList as an input parameter would mean that the method can run off an ArrayList or a ListItemCollection. It maximizes the chance that your method is useful.
Be as strict as possible when returning values. Prefer a specialized type if possible. This means clients do not have to second-guess or jump through hoops to process the return value. Also specialized types have greater functionality. If you choose to return an IList or an IEnumerable, the number of things the caller can do with your return value drastically reduces - e.g. If you return an IList over an ArrayList, to get the number of elements returned - use the Count property, the client must downcast. But then such downcasting defeats the purpose - works today.. won't tomorrow (if you change the Type of returned object). So for all purposes, the client can't get a count of elements easily - leading him to write mundane boilerplate code (in multiple places or as a helper method)
The summary here is it depends on the context (exceptions to most rules). E.g. if the most probable use of your return value is that clients would use the returned list to search for some element, it makes sense to return a List Implementation (type) that supports some kind of search method. Make it as easy as possible for the client to consume the return value.
I could see how, in some cases, having a more specific data type returned could be useful. For example knowing that the return value is a LinkedList rather than just List would allow you to do a delete from the list knowing that it will be efficient.
I think, while designing interfaces, you should design a method to return the as abstract data type as possible. Returning specific type would make the purpose of the method more clear about what they return.
Also, I would understand it in this way:
Return as abstract a data type as possible = return as specific a data type as possible
i.e. when your method is supposed to return any collection data type return collection rather than object.
tell me if i m wrong.
A specific return type is much more valuable because it:
reduces possible performance issues with discovering functionality with casting or reflection
increases code readability
does NOT in fact, expose more than is necessary.
The return type of a function is specifically chosen to cater to ALL of its callers. It is the calling function that should USE the return variable as abstractly as possible, since the calling function knows how the data will be used.
Is it only necessary to traverse the structure? is it necessary to sort the structure? transform it? clone it? These are questions only the caller can answer, and thus can use an abstracted type. The called function MUST provide for all of these cases.
If,in fact, the most specific use case you have right now is Iterable< string >, then that's fine. But more often than not - your callers will eventually need to have more details, so start with a specific return type - it doesn't cost anything.
Update: Please read this question in the context of design principles, elegance, expression of intent, and especially the "signals" sent to other programmers by design choices.
I have two "views" of a set of objects. One is a dictionary/map indexing the objects by a string value. The other is a dictionary/map indexing the objects by an ordinal (ordering integer). There is no "master" collection of the objects by themselves that can serve as the authoritative source for the number of objects, but the two dictionaries should always both contain references to all the objects.
When a new item is added to the set a reference is added to both dictionaries, and then some processing needs to be done which is affected by the new total number of objects.
What should I use as the authoritative source to reference for the current size of the set of objects? It seems that all my options are flawed in one dimension or another. I can just consistently reference one of the dictionaries, but that would codify an implication of that dictionary's superiority over the other. I could add a 3rd collection, a simple list of the objects to serve as the authoritative list, but that increases redundancy. Storing a running count seems simplest, but also increases redundancy and is more brittle than referencing a collection's self-tracked count on the fly.
Is there another option that will allow me to avoid choosing the lesser evil, or will I have to accept a compromise on elegance?
I would create a class that has (at least) two collections.
A version of the collection that is
sorted by string
A version of the
collection that is sorted by ordinal
(Optional) A master collection
The class would handle the nitty gritty management:
The syncing of the contents for the collections
Standard collection actions (e.g. Allow users get the size, Add or retrieve items)
Let users get by string or ordinal
That way you can use the same collection wherever you need either behavior, but still abstract away the "indexing" behavior you are going for.
The separate class gives you a single interface with which to explain your intent regarding how this class is to be used.
I'd suggest encapsulation: create a class that hides the "management" details (such as the current count) and use it to expose immutable "views" of the two collections.
Clients will ask the "manglement" object for an appropriate reference to one of the collections.
Clients adding a "term" (for lack of a better word) to the collections will do so through the "manglement" object.
This way your assumptions and implementation choices are "hidden" from clients of the service and you can document that the choice of collection for size/count was arbitrary. Future maintainers can change how the count is managed without breaking clients.
BTW, yes, I meant "manglement" - my favorite malapropism for management (in any context!)
If both dictionaries contain references to every object, the count should be the same for both of them, correct? If so, just pick one and be consistent.
I don't think it is a big deal at all. Just reference the sets in the same order each time
you need to get access to them.
If you really are concerned about it you could encapsulate the collections with a wrapper that exposes the public interfaces - like
Add(item)
Count()
This way it will always be consistent and atomic - or at least you could implement it that way.
But, I don't think it is a big deal.
The title may not really explain what I'm really trying to get at, couldn't really think of a way to describe what I mean.
I was wondering if it is good practice to check the arguments that a function accepts for nulls or empty before using them. I have this function which just wraps some hash creation like so.
Public Shared Function GenerateHash(ByVal FilePath As IO.FileInfo) As String
If (FilePath Is Nothing) Then
Throw New ArgumentNullException("FilePath")
End If
Dim _sha As New Security.Cryptography.MD5CryptoServiceProvider
Dim _Hash = Convert.ToBase64String(_sha.ComputeHash(New IO.FileStream(FilePath.FullName, IO.FileMode.Open, IO.FileAccess.Read)))
Return _Hash
End Function
As you can see I just takes a IO.Fileinfo as an argument, at the start of the function I am checking to make sure that it is not nothing.
I'm wondering is this good practice or should I just let it get to the actual hasher and then throw the exception because it is null.?
Thanks.
In general, I'd suggest it's good practice to validate all of the arguments to public functions/methods before using them, and fail early rather than after executing half of the function. In this case, you're right to throw the exception.
Depending on what your method is doing, failing early could be important. If your method was altering instance data on your class, you don't want it to alter half of the data, then encounter the null and throw an exception, as your object's data might them be in an intermediate and possibly invalid state.
If you're using an OO language then I'd suggest it's essential to validate the arguments to public methods, but less important with private and protected methods. My rationale here is that you don't know what the inputs to a public method will be - any other code could create an instance of your class and call it's public methods, and pass in unexpected/invalid data. Private methods, however, are called from inside the class, and the class should already have validated any data passing around internally.
One of my favourite techniques in C++ was to DEBUG_ASSERT on NULL pointers. This was drilled into me by senior programmers (along with const correctness) and is one of the things I was most strict on during code reviews. We never dereferenced a pointer without first asserting it wasn't null.
A debug assert is only active for debug targets (it gets stripped in release) so you don't have the extra overhead in production to test for thousands of if's. Generally it would either throw an exception or trigger a hardware breakpoint. We even had systems that would throw up a debug console with the file/line info and an option to ignore the assert (once or indefinitely for the session). That was such a great debug and QA tool (we'd get screenshots with the assert on the testers screen and information on whether the program continued if ignored).
I suggest asserting all invariants in your code including unexpected nulls. If performance of the if's becomes a concern find a way to conditionally compile and keep them active in debug targets. Like source control, this is a technique that has saved my ass more often than it has caused me grief (the most important litmus test of any development technique).
Yes, it's good practice to validate all arguments at the beginning of a method and throw appropriate exceptions like ArgumentException, ArgumentNullException, or ArgumentOutOfRangeException.
If the method is private such that only you the programmer could pass invalid arguments, then you may choose to assert each argument is valid (Debug.Assert) instead of throw.
If NULL is an inacceptable input, throw an exception. By yourself, like you did in your sample, so that the message is helpful.
Another method of handling NULL inputs is just to respont with a NULL in turn. Depends on the type of function -- in the example above I would keep the exception.
If its for an externally facing API then I would say you want to check every parameter as the input cannot be trusted.
However, if it is only going to be used internally then the input should be able to be trusted and you can save yourself a bunch of code that's not adding value to the software.
You should check all arguments against the set of assumptions that you make in that function about their values.
As in your example, if a null argument to your function doesn't make any sense and you're assuming that anyone using your function will know this then being passed a null argument shows some sort of error and some sort of action taken (eg. throwing an exception). And if you use asserts (as James Fassett got in and said before me ;-) ) they cost you nothing in a release version. (they cost you almost nothing in a debug version either)
The same thing applies to any other assumption.
And it's going to be easier to trace the error if you generate it than if you leave it to some standard library routine to throw the exception. You will be able to provide much more useful contextual information.
It's outside the bounds of this question, but you do need to expose the assumptions that your function makes - for example, through the comment header to your function.
According to The Pragmatic Programmer by Andrew Hunt and David Thomas, it is the responsibility of the caller to make sure it gives valid input. So, you must now choose whether you consider a null input to be valid. Unless it makes specific sense to consider null to be a valid input (e.g. it is probably a good idea to consider null to be a legal input if you're testing for equality), I would consider it invalid. That way your program, when it hits incorrect input, will fail sooner. If your program is going to encounter an error condition, you want it to happen as soon as possible. In the event your function does inadvertently get passed a null, you should consider it to be a bug, and react accordingly (i.e. instead of throwing an exception, you should consider making use of an assertion that kills the program, until you are releasing the program).
Classic design by contract: If input is right, output will be right. If input is wrong, there is a bug. (if input is right but output is wrong, there is a bug. That's a gimme.)
I'll add a couple of elaborations (in bold) to the excellent design by contract advice offerred by Brian earlier...
The priniples of "design by contract" require that you define what is acceptable for the caller to pass in (the valid domain of input values) and then, for any valid input, what the method/provider will do.
For an internal method, you can define NULLs as outside the domain of valid input parameters. In this case, you would immediately assert that the input parameter value is NOT NULL. The key insight in this contract specification is that any call passing in a NULL value IS A CALLER'S BUG and the error thrown by the assert statement is the proper behavior.
Now, while very well defined and parsimonius, if you're exposing the method to external/public callers, you should ask yourself, is that the contract I/we really want?
Probably not. In a public interface, you'd probably accept the NULL (as technically in the domain of inputs that the method accepts), but then decline to process gracefully w/ a return message. (More work to meet the naturally more complex customer-facing requirement.)
In either case, what you're after is a protocol that handles all of the cases from both the perspective of the caller and the provider, not lots of scattershot tests that can make it difficult to assess the completeness or lack of completeness of the contractual condition coverage.
Most of the time, letting it just throw the exception is pretty reasonable as long as you are sure the exception won't be ignored.
If you can add something to it, however, it doesn't hurt to wrap the exception with one that is more accurate and rethrow it. Decoding "NullPointerException" is going to take a bit longer than "IllegalArgumentException("FilePath MUST be supplied")" (Or whatever).
Lately I've been working on a platform where you have to run an obfuscator before you test. Every stack trace looks like monkeys typing random crap, so I got in the habit of checking my arguments all the time.
I'd love to see a "nullable" or "nonull" modifier on variables and arguments so the compiler can check for you.
If you're writing a public API, do your caller the favor of helping them find their bugs quickly, and check for valid inputs.
If you're writing an API where the caller might untrusted (or the caller of the caller), checked for valid inputs, because it's good security.
If your APIs are only reachable by trusted callers, like "internal" in C#, then don't feel like you have to write all that extra code. It won't be useful to anyone.