I read quite a lot about the visitor pattern and its supposed advantages. To me however it seems they are not that much advantages when applied in practice:
"Convenient" and "elegant" seems to mean lots and lots of boilerplate code
Therefore, the code is hard to follow. Also 'accept'/'visit' is not very descriptive
Even uglier boilerplate code if your programming language has no method overloading (i.e. Vala)
You cannot in general add new operations to an existing type hierarchy without modification of all classes, since you need new 'accept'/'visit' methods everywhere as soon as you need an operation with different parameters and/or return value (changes to classes all over the place is one thing this design pattern was supposed to avoid!?)
Adding a new type to the type hierarchy requires changes to all visitors. Also, your visitors cannot simply ignore a type - you need to create an empty visit method (boilerplate again)
It all just seems to be an awful lot of work when all you want to do is actually:
// Pseudocode
int SomeOperation(ISomeAbstractThing obj) {
switch (type of obj) {
case Foo: // do Foo-specific stuff here
case Bar: // do Bar-specific stuff here
case Baz: // do Baz-specific stuff here
default: return 0; // do some sensible default if type unknown or if we don't care
}
}
The only real advantage I see (which btw i haven't seen mentioned anywhere): The visitor pattern is probably the fastest method to implement the above code snippet in terms of cpu time (if you don't have a language with double dispatch or efficient type comparison in the fashion of the pseudocode above).
Questions:
So, what advantages of the visitor pattern have I missed?
What alternative concepts/data structures could be used to make the above fictional code sample run equally fast?
For as far as I have seen so far there are two uses / benefits for the visitor design pattern:
Double dispatch
Separate data structures from the operations on them
Double dispatch
Let's say you have a Vehicle class and a VehicleWasher class. The VehicleWasher has a Wash(Vehicle) method:
VehicleWasher
Wash(Vehicle)
Vehicle
Additionally we also have specific vehicles like a car and in the future we'll also have other specific vehicles. For this we have a Car class but also a specific CarWasher class that has an operation specific to washing cars (pseudo code):
CarWasher : VehicleWasher
Wash(Car)
Car : Vehicle
Then consider the following client code to wash a specific vehicle (notice that x and washer are declared using their base type because the instances might be dynamically created based on user input or external configuration values; in this example they are simply created with a new operator though):
Vehicle x = new Car();
VehicleWasher washer = new CarWasher();
washer.Wash(x);
Many languages use single dispatch to call the appropriate function. Single dispatch means that during runtime only a single value is taken into account when determining which method to call. Therefore only the actual type of washer we'll be considered. The actual type of x isn't taken into account. The last line of code will therefore invoke CarWasher.Wash(Vehicle) and NOT CarWasher.Wash(Car).
If you use a language that does not support multiple dispatch and you do need it (I can honoustly say I have never encountered such a situation though) then you can use the visitor design pattern to enable this. For this two things need to be done. First of all add an Accept method to the Vehicle class (the visitee) that accepts a VehicleWasher as a visitor and then call its operation Wash:
Accept(VehicleWasher washer)
washer.Wash(this);
The second thing is to modify the calling code and replace the washer.Wash(x); line with the following:
x.Accept(washer);
Now for the call to the Accept method the actual type of x is considered (and only that of x since we are assuming to be using a single dispatch language). In the implementation of the Accept method the Wash method is called on the washer object (the visitor). For this the actual type of the washer is considered and this will invoke CarWasher.Wash(Car). By combining two single dispatches a double dispatch is implemented.
Now to eleborate on your remark of the terms like Accept and Visit and Visitor being very unspecific. That is absolutely true. But it is for a reason.
Consider the requirement in this example to implement a new class that is able to repair vehicles: a VehicleRepairer. This class can only be used as a visitor in this example if it would inherit from VehicleWasher and have its repair logic inside a Wash method. But that ofcourse doesn't make any sense and would be confusing. So I totally agree that design patterns tend to have very vague and unspecific naming but it does make them applicable to many situations. The more specific your naming is, the more restrictive it can be.
Your switch statement only considers one type which is actually a manual way of single dispatch. Applying the visitor design pattern in the above way will provide double dispatch.
This way you do not necessarily need additional Visit methods when adding additional types to your hierarchy. Ofcourse it does add some complexity as it makes the code less readable. But ofcourse all patterns come at a price.
Ofcourse this pattern cannot always be used. If you expect lots of complex operations with multiple parameters then this will not be a good option.
An alternative is to use a language that does support multiple dispatch. For instance .NET did not support it until version 4.0 which introduced the dynamic keyword. Then in C# you can do the following:
washer.Wash((dynamic)x);
Because x is then converted to a dynamic type its actual type will be considered for the dispatch and so both x and washer will be used to select the correct method so that CarWasher.Wash(Car) will be called (making the code work correctly and staying intuitive).
Separate data structures and operations
The other benefit and requirement is that it can separate the data structures from the operations. This can be an advantage because it allows new visitors to be added that have there own operations while it also allows data structures to be added that 'inherit' these operations. It can however be only applied if this seperation can be done / makes sense. The classes that perform the operations (the visitors) do not know the structure of the data structures nor do they have to know that which makes code more maintainable and reusable. When applied for this reason the visitors have operations for the different elements in the data structures.
Say you have different data structures and they all consist of elements of class Item. The structures can be lists, stacks, trees, queues etc.
You can then implement visitors that in this case will have the following method:
Visit(Item)
The data structures need to accept visitors and then call the Visit method for each Item.
This way you can implement all kinds of visitors and you can still add new data structures as long as they consist of elements of type Item.
For more specific data structures with additional elements (e.g. a Node) you might consider a specific visitor (NodeVisitor) that inherits from your conventional Visitor and have your new data structures accept that visitor (Accept(NodeVisitor)). The new visitors can be used for the new data structures but also for the old data structures due to inheritence and so you do not need to modify your existing 'interface' (the super class in this case).
In my personal opinion, the visitor pattern is only useful if the interface you want implemented is rather static and doesn't change a lot, while you want to give anyone a chance to implement their own functionality.
Note that you can avoid changing everything every time you add a new method by creating a new interface instead of modifying the old one - then you just have to have some logic handling the case when the visitor doesn't implement all the interfaces.
Basically, the benefit is that it allows you to choose the correct method to call at runtime, rather than at compile time - and the available methods are actually extensible.
For more info, have a look at this article - http://rgomes-info.blogspot.co.uk/2013/01/a-better-implementation-of-visitor.html
By experience, I would say that "Adding a new type to the type hierarchy requires changes to all visitors" is an advantage. Because it definitely forces you to consider the new type added in ALL places where you did some type-specific stuff. It prevents you from forgetting one....
This is an old question but i would like to answer.
The visitor pattern is useful mostly when you have a composite pattern in place in which you build a tree of objects and such tree arrangement is unpredictable.
Type checking may be one thing that a visitor can do, but say you want to build an expression based on a tree that can vary its form according to a user input or something like that, a visitor would be an effective way for you to validate the tree, or build a complex object according to the items found on the tree.
The visitor may also carry an object that does something on each node it may find on that tree. this visitor may be a composite itself chaining lots of operations on each node, or it can carry a mediator object to mediate operations or dispatch events on each node.
You imagination is the limit of all this. you can filter a collection, build an abstract syntax tree out of an complete tree, parse a string, validate a collection of things, etc.
This question already has answers here:
Closed 12 years ago.
Possible Duplicate:
Specific functions vs many Arguments vs context dependent
So I've been developing for 3-4 years now, know a wide range of languages, know some impressive (to the small minded :P ) stuff.
But something I've always wondered; when I make a function should it be for a specific purpose, or should it be moulded to be re-usable, even if I have no need for it to be?
E.G:
//JS, but could be any language really
//specific
function HAL(){
alert("I'm afraid I can't let you do that, " + document.getElementById("Name").value + ".");
}
//generic
function HAL(nme){
alert("I'm afraid I can't let you do that, " + nme + ".");
}
//more generic
function HAL(msg, nme){
alert(msg + " " + nme + ".");
}
Yes, very simple example, but conveys the point I want to make. If we take this example, would I ever use it outside of the first? Probably not, so I'd be tempted to make it this way, but then common sense would (now) convince me to make it the second, yet I can't see any benefit of this way, if I know it's not going to be used in any other way, i.e. It's always going to use the input's value (Yes I would put that into a global variable normally).
Is it just a case of whatever I feel makes the most sense at the time, or should I follow the 2nd pattern as best I can?
In that particular case, I would write the first function for now (YAGNI, right?), and probably never need to change it. Then, if it turned out I did need to support alternate names, I'd make the current behavior the default, but allow an optional parameter to specify a name. Likewise with the message.
# In Ruby, but like you say, could be in anything:
// specific
def hal()
puts "I'm afraid I can't let you do that, #{fetch_name}."
end
// genericized refactoring
def hal( name = fetch_name )
puts "I'm afraid I can't let you do that, #{name}."
end
Typically, that's the approach I prefer to take: create functions at whatever is the most convenient degree of specificity for my current needs, but leave the door open for a more generalized approach later.
It helps that I use languages like Ruby that make this easy, but you can take the same approach to some extent even in Java or C. For example, in Java you might make a specific method with no parameters first, and then later refactor to a more generalized method with a "name" parameter and a no-parameter wrapper that filled in the default name.
A rule of thumb is that a function should have minimal side effects.
So, really, it would look something like this:
//By the way - don't call functions nouns. functions are verbs. data are nouns
void HAL(string s)
{
voicetype_t vt = voice.type();
voice.type(VOICE_OF_DOOM);
voice.say(s);
voice.type(vt);
}
A function shouldn't be just a series of statements to call them in some other context. It should be a unit of functionality that you want to abstract. Making a function to be specific is good, but making it context sensitive is bad. What you should do, is to use the generic way(last one) presented in your post, but provide the messages as constants. The language you use has some way to declare constants right?
In your example, I wouldn't make it generic. If a functionality can be used in many cases, make it generic so you can use it all the time without "copy, paste, make minor change, repeat". But telling the user he can't do that and adressing it as [contents of certain input field] is useful for only one case. Plus, the last shot is pointless.
However, I generally prefer my code to be as generic as feasible. Well, as long as the odds are I will need it one day... let's not violate YANGI too hard. But if it can be generic without hassle, why not?
In my opinion, functions should be genericized only to the extent that their purposes need to be. In other words, you should concede to the fact that, although we want to think differently, not everything is reusable, and thus, you shouldn't go out of your way to implement everything to be like that. Programmers should be conscious of the scope (and possibly the future development) of the product, so ultimately one should use their intuition as to how far to take generalizations of functions.
As for your examples, #3 is completely worthless as it only affixes a space between two strings and appends a period at the end--why would someone do this with a special function? I know that's only an example, but if we're talking about how far to generalize a method, something like that's taking it too far--almost to the point where it's just wasted LOC, and that is never something to sacrifice for the sake of generalizing.
Closed. This question is opinion-based. It is not currently accepting answers.
Closed 11 months ago.
Locked. This question and its answers are locked because the question is off-topic but has historical significance. It is not currently accepting new answers or interactions.
Sorry for the waffly title - if I could come up with a concise title, I wouldn't have to ask the question.
Suppose I have an immutable list type. It has an operation Foo(x) which returns a new immutable list with the specified argument as an extra element at the end. So to build up a list of strings with values "Hello", "immutable", "world" you could write:
var empty = new ImmutableList<string>();
var list1 = empty.Foo("Hello");
var list2 = list1.Foo("immutable");
var list3 = list2.Foo("word");
(This is C# code, and I'm most interested in a C# suggestion if you feel the language is important. It's not fundamentally a language question, but the idioms of the language may be important.)
The important thing is that the existing lists are not altered by Foo - so empty.Count would still return 0.
Another (more idiomatic) way of getting to the end result would be:
var list = new ImmutableList<string>().Foo("Hello")
.Foo("immutable")
.Foo("word");
My question is: what's the best name for Foo?
EDIT 3: As I reveal later on, the name of the type might not actually be ImmutableList<T>, which makes the position clear. Imagine instead that it's TestSuite and that it's immutable because the whole of the framework it's a part of is immutable...
(End of edit 3)
Options I've come up with so far:
Add: common in .NET, but implies mutation of the original list
Cons: I believe this is the normal name in functional languages, but meaningless to those without experience in such languages
Plus: my favourite so far, it doesn't imply mutation to me. Apparently this is also used in Haskell but with slightly different expectations (a Haskell programmer might expect it to add two lists together rather than adding a single value to the other list).
With: consistent with some other immutable conventions, but doesn't have quite the same "additionness" to it IMO.
And: not very descriptive.
Operator overload for + : I really don't like this much; I generally think operators should only be applied to lower level types. I'm willing to be persuaded though!
The criteria I'm using for choosing are:
Gives the correct impression of the result of the method call (i.e. that it's the original list with an extra element)
Makes it as clear as possible that it doesn't mutate the existing list
Sounds reasonable when chained together as in the second example above
Please ask for more details if I'm not making myself clear enough...
EDIT 1: Here's my reasoning for preferring Plus to Add. Consider these two lines of code:
list.Add(foo);
list.Plus(foo);
In my view (and this is a personal thing) the latter is clearly buggy - it's like writing "x + 5;" as a statement on its own. The first line looks like it's okay, until you remember that it's immutable. In fact, the way that the plus operator on its own doesn't mutate its operands is another reason why Plus is my favourite. Without the slight ickiness of operator overloading, it still gives the same connotations, which include (for me) not mutating the operands (or method target in this case).
EDIT 2: Reasons for not liking Add.
Various answers are effectively: "Go with Add. That's what DateTime does, and String has Replace methods etc which don't make the immutability obvious." I agree - there's precedence here. However, I've seen plenty of people call DateTime.Add or String.Replace and expect mutation. There are loads of newsgroup questions (and probably SO ones if I dig around) which are answered by "You're ignoring the return value of String.Replace; strings are immutable, a new string gets returned."
Now, I should reveal a subtlety to the question - the type might not actually be an immutable list, but a different immutable type. In particular, I'm working on a benchmarking framework where you add tests to a suite, and that creates a new suite. It might be obvious that:
var list = new ImmutableList<string>();
list.Add("foo");
isn't going to accomplish anything, but it becomes a lot murkier when you change it to:
var suite = new TestSuite<string, int>();
suite.Add(x => x.Length);
That looks like it should be okay. Whereas this, to me, makes the mistake clearer:
var suite = new TestSuite<string, int>();
suite.Plus(x => x.Length);
That's just begging to be:
var suite = new TestSuite<string, int>().Plus(x => x.Length);
Ideally, I would like my users not to have to be told that the test suite is immutable. I want them to fall into the pit of success. This may not be possible, but I'd like to try.
I apologise for over-simplifying the original question by talking only about an immutable list type. Not all collections are quite as self-descriptive as ImmutableList<T> :)
In situations like that, I usually go with Concat. That usually implies to me that a new object is being created.
var p = listA.Concat(listB);
var k = listA.Concat(item);
I'd go with Cons, for one simple reason: it means exactly what you want it to.
I'm a huge fan of saying exactly what I mean, especially in source code. A newbie will have to look up the definition of Cons only once, but then read and use that a thousand times. I find that, in the long term, it's nicer to work with systems that make the common case easier, even if the up-front cost is a little bit higher.
The fact that it would be "meaningless" to people with no FP experience is actually a big advantage. As you pointed out, all of the other words you found already have some meaning, and that meaning is either slightly different or ambiguous. A new concept should have a new word (or in this case, an old one). I'd rather somebody have to look up the definition of Cons, than to assume incorrectly he knows what Add does.
Other operations borrowed from functional languages often keep their original names, with no apparent catastrophes. I haven't seen any push to come up with synonyms for "map" and "reduce" that sound more familiar to non-FPers, nor do I see any benefit from doing so.
(Full disclosure: I'm a Lisp programmer, so I already know what Cons means.)
Actually I like And, especially in the idiomatic way. I'd especially like it if you had a static readonly property for the Empty list, and perhaps make the constructor private so you always have to build from the empty list.
var list = ImmutableList<string>.Empty.And("Hello")
.And("Immutable")
.And("Word");
Whenever I'm in a jam with nomenclature, I hit up the interwebs.
thesaurus.com returns this for "add":
Definition: adjoin, increase; make
further comment
Synonyms: affix,
annex, ante, append, augment, beef
up, boost, build up, charge up,
continue, cue in, figure in, flesh
out, heat up, hike, hike up, hitch on,
hook on, hook up with, include, jack
up, jazz up, join together, pad,
parlay, piggyback, plug into, pour it
on, reply, run up, say further, slap
on, snowball, soup up, speed up,
spike, step up, supplement, sweeten,
tack on, tag
I like the sound of Adjoin, or more simply Join. That is what you're doing, right? The method could also apply to joining other ImmutableList<>'s.
Personally, I like .With(). If I was using the object, after reading the documentation or the code comments, it would be clear what it does, and it reads ok in the source code.
object.With("My new item as well");
Or, you add "Along" with it.. :)
object.AlongWith("this new item");
I ended up going with Add for all of my Immutable Collections in BclExtras. The reason being is that it's an easy predictable name. I'm not worried so much about people confusing Add with a mutating add since the name of the type is prefixed with Immutable.
For awhile I considered Cons and other functional style names. Eventually I discounted them because they're not nearly as well known. Sure functional programmers will understand but they're not the majority of users.
Other Names: you mentioned:
Plus: I'm wishy/washing on this one. For me this doesn't distinguish it as being a non-mutating operation anymore than Add does
With: Will cause issues with VB (pun intended)
Operator overloading: Discoverability would be an issue
Options I considered:
Concat: String's are Immutable and use this. Unfortunately it's only really good for adding to the end
CopyAdd: Copy what? The source, the list?
AddToNewList: Maybe a good one for List. But what about a Collection, Stack, Queue, etc ...
Unfortunately there doesn't really seem to be a word that is
Definitely an immutable operation
Understandable to the majority of users
Representable in less than 4 words
It gets even more odd when you consider collections other than List. Take for instance Stack. Even first year programmers can tell you that Stacks have a Push/Pop pair of methods. If you create an ImmutableStack and give it a completely different name, lets call it Foo/Fop, you've just added more work for them to use your collection.
Edit: Response to Plus Edit
I see where you're going with Plus. I think a stronger case would actually be Minus for remove. If I saw the following I would certainly wonder what in the world the programmer was thinking
list.Minus(obj);
The biggest problem I have with Plus/Minus or a new pairing is it feels like overkill. The collection itself already has a distinguishing name, the Immutable prefix. Why go further by adding vocabulary whose intent is to add the same distinction as the Immutable prefix already did.
I can see the call site argument. It makes it clearer from the standpoint of a single expression. But in the context of the entire function it seems unnecessary.
Edit 2
Agree that people have definitely been confused by String.Concat and DateTime.Add. I've seen several very bright programmers hit this problem.
However I think ImmutableList is a different argument. There is nothing about String or DateTime that establishes it as Immutable to a programmer. You must simply know that it's immutable via some other source. So the confusion is not unexpected.
ImmutableList does not have that problem because the name defines it's behavior. You could argue that people don't know what Immutable is and I think that's also valid. I certainly didn't know it till about year 2 in college. But you have the same issue with whatever name you choose instead of Add.
Edit 3: What about types like TestSuite which are immutable but do not contain the word?
I think this drives home the idea that you shouldn't be inventing new method names. Namely because there is clearly a drive to make types immutable in order to facilitate parallel operations. If you focus on changing the name of methods for collections, the next step will be the mutating method names on every type you use that is immutable.
I think it would be a more valuable effort to instead focus on making types identifiable as Immutable. That way you can solve the problem without rethinking every mutating method pattern out there.
Now how can you identify TestSuite as Immutable? In todays environment I think there are a few ways
Prefix with Immutable: ImmutableTestSuite
Add an Attribute which describes the level of Immutablitiy. This is certainly less discoverable
Not much else.
My guess/hope is development tools will start helping this problem by making it easy to identify immutable types simply by sight (different color, stronger font, etc ...). But I think that's the answer though over changing all of the method names.
I think this may be one of those rare situations where it's acceptable to overload the + operator. In math terminology, we know that + doesn't append something to the end of something else. It always combines two values together and returns a new resulting value.
For example, it's intuitively obvious that when you say
x = 2 + 2;
the resulting value of x is 4, not 22.
Similarly,
var empty = new ImmutableList<string>();
var list1 = empty + "Hello";
var list2 = list1 + "immutable";
var list3 = list2 + "word";
should make clear what each variable is going to hold. It should be clear that list2 is not changed in the last line, but instead that list3 is assigned the result of appending "word" to list2.
Otherwise, I would just name the function Plus().
To be as clear as possible, you might want to go with the wordier CopyAndAdd, or something similar.
I would call it Extend() or maybe ExtendWith() if you feel like really verbose.
Extends means adding something to something else without changing it. I think this is very relevant terminology in C# since this is similar to the concept of extension methods - they "add" a new method to a class without "touching" the class itself.
Otherwise, if you really want to emphasize that you don't modify the original object at all, using some prefix like Get- looks like unavoidable to me.
Added(), Appended()
I like to use the past tense for operations on immutable objects. It conveys the idea that you aren't changing the original object, and it's easy to recognize when you see it.
Also, because mutating method names are often present-tense verbs, it applies to most of the immutable-method-name-needed cases you run into. For example an immutable stack has the methods "pushed" and "popped".
I like mmyers suggestion of CopyAndAdd. In keeping with a "mutation" theme, maybe you could go with Bud (asexual reproduction), Grow, Replicate, or Evolve? =)
EDIT: To continue with my genetic theme, how about Procreate, implying that a new object is made which is based on the previous one, but with something new added.
This is probably a stretch, but in Ruby there is a commonly used notation for the distinction: add doesn't mutate; add! mutates. If this is an pervasive problem in your project, you could do that too (not necessarily with non-alphabetic characters, but consistently using a notation to indicate mutating/non-mutating methods).
Join seems appropriate.
Maybe the confusion stems from the fact that you want two operations in one. Why not separate them? DSL style:
var list = new ImmutableList<string>("Hello");
var list2 = list.Copy().With("World!");
Copy would return an intermediate object, that's a mutable copy of the original list. With would return a new immutable list.
Update:
But, having an intermediate, mutable collection around is not a good approach. The intermediate object should be contained in the Copy operation:
var list1 = new ImmutableList<string>("Hello");
var list2 = list1.Copy(list => list.Add("World!"));
Now, the Copy operation takes a delegate, which receives a mutable list, so that it can control the copy outcome. It can do much more than appending an element, like removing elements or sorting the list. It can also be used in the ImmutableList constructor to assemble the initial list without intermediary immutable lists.
public ImmutableList<T> Copy(Action<IList<T>> mutate) {
if (mutate == null) return this;
var list = new List<T>(this);
mutate(list);
return new ImmutableList<T>(list);
}
Now there's no possibility of misinterpretation by the users, they will naturally fall into the pit of success.
Yet another update:
If you still don't like the mutable list mention, even now that it's contained, you can design a specification object, that will specify, or script, how the copy operation will transform its list. The usage will be the same:
var list1 = new ImmutableList<string>("Hello");
// rules is a specification object, that takes commands to run in the copied collection
var list2 = list1.Copy(rules => rules.Append("World!"));
Now you can be creative with the rules names and you can only expose the functionality that you want Copy to support, not the entire capabilities of an IList.
For the chaining usage, you can create a reasonable constructor (which will not use chaining, of course):
public ImmutableList(params T[] elements) ...
...
var list = new ImmutableList<string>("Hello", "immutable", "World");
Or use the same delegate in another constructor:
var list = new ImmutableList<string>(rules =>
rules
.Append("Hello")
.Append("immutable")
.Append("World")
);
This assumes that the rules.Append method returns this.
This is what it would look like with your latest example:
var suite = new TestSuite<string, int>(x => x.Length);
var otherSuite = suite.Copy(rules =>
rules
.Append(x => Int32.Parse(x))
.Append(x => x.GetHashCode())
);
A few random thoughts:
ImmutableAdd()
Append()
ImmutableList<T>(ImmutableList<T> originalList, T newItem) Constructor
DateTime in C# uses Add. So why not use the same name? As long the users of your class understand the class is immutable.
I think the key thing you're trying to get at that's hard to express is the nonpermutation, so maybe something with a generative word in it, something like CopyWith() or InstancePlus().
I don't think the English language will let you imply immutability in an unmistakable way while using a verb that means the same thing as "Add". "Plus" almost does it, but people can still make the mistake.
The only way you're going to prevent your users from mistaking the object for something mutable is by making it explicit, either through the name of the object itself or through the name of the method (as with the verbose options like "GetCopyWith" or "CopyAndAdd").
So just go with your favourite, "Plus."
First, an interesting starting point:
http://en.wikipedia.org/wiki/Naming_conventions_(programming) ...In particular, check the "See Also" links at the bottom.
I'm in favor of either Plus or And, effectively equally.
Plus and And are both math-based in etymology. As such, both connote mathematical operation; both yield an expression which reads naturally as expressions which may resolve into a value, which fits with the method having a return value. And bears additional logic connotation, but both words apply intuitively to lists. Add connotes action performed on an object, which conflicts with the method's immutable semantics.
Both are short, which is especially important given the primitiveness of the operation. Simple, frequently-performed operations deserve shorter names.
Expressing immutable semantics is something I prefer to do via context. That is, I'd rather simply imply that this entire block of code has a functional feel; assume everything is immutable. That might just be me, however. I prefer immutability to be the rule; if it's done, it's done a lot in the same place; mutability is the exception.
How about Chain() or Attach()?
I prefer Plus (and Minus). They are easily understandable and map directly to operations involving well known immutable types (the numbers). 2+2 doesn't change the value of 2, it returns a new, equally immutable, value.
Some other possibilities:
Splice()
Graft()
Accrete()
How about mate, mateWith, or coitus, for those who abide. In terms of reproducing mammals are generally considered immutable.
Going to throw Union out there too. Borrowed from SQL.
Apparently I'm the first Obj-C/Cocoa person to answer this question.
NNString *empty = [[NSString alloc] init];
NSString *list1 = [empty stringByAppendingString:#"Hello"];
NSString *list2 = [list1 stringByAppendingString:#"immutable"];
NSString *list3 = [list2 stringByAppendingString:#"word"];
Not going to win any code golf games with this.
I think "Add" or "Plus" sounds fine. The name of the list itself should be enough to convey the list's immutability.
Maybe there are some words which remember me more of making a copy and add stuff to that instead of mutating the instance (like "Concatenate"). But i think having some symmetry for those words for other actions would be good to have too. I don't know of a similar word for "Remove" that i think of the same kind like "Concatenate". "Plus" sounds little strange to me. I wouldn't expect it being used in a non-numerical context. But that could aswell come from my non-english background.
Maybe i would use this scheme
AddToCopy
RemoveFromCopy
InsertIntoCopy
These have their own problems though, when i think about it. One could think they remove something or add something to an argument given. Not sure about it at all. Those words do not play nice in chaining either, i think. Too wordy to type.
Maybe i would just use plain "Add" and friends too. I like how it is used in math
Add 1 to 2 and you get 3
Well, certainly, a 2 remains a 2 and you get a new number. This is about two numbers and not about a list and an element, but i think it has some analogy. In my opinion, add does not necessarily mean you mutate something. I certainly see your point that having a lonely statement containing just an add and not using the returned new object does not look buggy. But I've now also thought some time about that idea of using another name than "add" but i just can't come up with another name, without making me think "hmm, i would need to look at the documentation to know what it is about" because its name differs from what I would expect to be called "add". Just some weird thought about this from litb, not sure it makes sense at all :)
Looking at http://thesaurus.reference.com/browse/add and http://thesaurus.reference.com/browse/plus I found gain and affix but I'm not sure how much they imply non-mutation.
I think that Plus() and Minus() or, alternatively, Including(), Excluding() are reasonable at implying immutable behavior.
However, no naming choice will ever make it perfectly clear to everyone, so I personally believe that a good xml doc comment would go a very long way here. VS throws these right in your face when you write code in the IDE - they're hard to ignore.
Append - because, note that names of the System.String methods suggest that they mutate the instance, but they don't.
Or I quite like AfterAppending:
void test()
{
Bar bar = new Bar();
List list = bar.AfterAppending("foo");
}
list.CopyWith(element)
As does Smalltalk :)
And also list.copyWithout(element) that removes all occurrences of an element, which is most useful when used as list.copyWithout(null) to remove unset elements.
I would go for Add, because I can see the benefit of a better name, but the problem would be to find different names for every other immutable operation which might make the class quite unfamiliar if that makes sense.
Similar to Is hard-coding literals ever acceptable?, but I'm specifically thinking of "magic strings" here.
On a large project, we have a table of configuration options like these:
Name Value
---- -----
FOO_ENABLED Y
BAR_ENABLED N
...
(Hundreds of them).
The common practice is to call a generic function to test an option like this:
if (config_options.value('FOO_ENABLED') == 'Y') ...
(Of course, this same option may need to be checked in many places in the system code.)
When adding a new option, I was considering adding a function to hide the "magic string" like this:
if (config_options.foo_enabled()) ...
However, colleagues thought I'd gone overboard and objected to doing this, preferring the hard-coding because:
That's what we normally do
It makes it easier to see what's going on when debugging the code
The trouble is, I can see their point! Realistically, we are never going to rename the options for any reason, so about the only advantage I can think of for my function is that the compiler would catch any typo like fo_enabled(), but not 'FO_ENABLED'.
What do you think? Have I missed any other advantages/disadvantages?
If I use a string once in the code, I don't generally worry about making it a constant somewhere.
If I use a string twice in the code, I'll consider making it a constant.
If I use a string three times in the code, I'll almost certainly make it a constant.
if (config_options.isTrue('FOO_ENABLED')) {...
}
Restrict your hard coded Y check to one place, even if it means writing a wrapper class for your Map.
if (config_options.isFooEnabled()) {...
}
Might seem okay until you have 100 configuration options and 100 methods (so here you can make a judgement about future application growth and needs before deciding on your implementation). Otherwise it is better to have a class of static strings for parameter names.
if (config_options.isTrue(ConfigKeys.FOO_ENABLED)) {...
}
I realise the question is old, but it came up on my margin.
AFAIC, the issue here has not been identified accurately, either in the question, or the answers. Forget about 'harcoding strings" or not, for a moment.
The database has a Reference table, containing config_options. The PK is a string.
There are two types of PKs:
Meaningful Identifiers, that the users (and developers) see and use. These PKs are supposed to be stable, they can be relied upon.
Meaningless Id columns which the users should never see, that the developers have to be aware of, and code around. These cannot be relied upon.
It is ordinary, normal, to write code using the absolute value of a meaningful PK IF CustomerCode = "IBM" ... or IF CountryCode = "AUS" etc.
referencing the absolute value of a meaningless PK is not acceptable (due to auto-increment; gaps being changed; values being replaced wholesale).
.
Your reference table uses meaningful PKs. Referencing those literal strings in code is unavoidable. Hiding the value will make maintenance more difficult; the code is no longer literal; your colleagues are right. Plus there is the additional redundant function that chews cycles. If there is a typo in the literal, you will soon find that out during Dev testing, long before UAT.
hundreds of functions for hundreds of literals is absurd. If you do implement a function, then Normalise your code, and provide a single function that can be used for any of the hundreds of literals. In which case, we are back to a naked literal, and the function can be dispensed with.
the point is, the attempt to hide the literal has no value.
.
It cannot be construed as "hardcoding", that is something quite different. I think that is where your issue is, identifying these constructs as "hardcoded". It is just referencing a Meaningfull PK literally.
Now from the perspective of any code segment only, if you use the same value a few times, you can improve the code by capturing the literal string in a variable, and then using the variable in the rest of the code block. Certainly not a function. But that is an efficiency and good practice issue. Even that does not change the effect IF CountryCode = #cc_aus
I really should use constants and no hard coded literals.
You can say they won't be changed, but you may never know. And it is best to make it a habit. To use symbolic constants.
In my experience, this kind of issue is masking a deeper problem: failure to do actual OOP and to follow the DRY principle.
In a nutshell, capture the decision at startup time by an appropriate definition for each action inside the if statements, and then throw away both the config_options and the run-time tests.
Details below.
The sample usage was:
if (config_options.value('FOO_ENABLED') == 'Y') ...
which raises the obvious question, "What's going on in the ellipsis?", especially given the following statement:
(Of course, this same option may need to be checked in many places in the system code.)
Let's assume that each of these config_option values really does correspond to a single problem domain (or implementation strategy) concept.
Instead of doing this (repeatedly, in various places throughout the code):
Take a string (tag),
Find its corresponding other string (value),
Test that value as a boolean-equivalent,
Based on that test, decide whether to perform some action.
I suggest encapsulating the concept of a "configurable action".
Let's take as an example (obviously just as hypthetical as FOO_ENABLED ... ;-) that your code has to work in either English units or metric units. If METRIC_ENABLED is "true", convert user-entered data from metric to English for internal computation, and convert back prior to displaying results.
Define an interface:
public interface MetricConverter {
double toInches(double length);
double toCentimeters(double length);
double toPounds(double weight);
double toKilograms(double weight);
}
which identifies in one place all the behavior associated with the concept of METRIC_ENABLED.
Then write concrete implementations of all the ways those behaviors are to be carried out:
public class NullConv implements MetricConverter {
double toInches(double length) {return length;}
double toCentimeters(double length) {return length;}
double toPounds(double weight) {return weight;}
double toKilograms(double weight) {return weight;}
}
and
// lame implementation, just for illustration!!!!
public class MetricConv implements MetricConverter {
public static final double LBS_PER_KG = 2.2D;
public static final double CM_PER_IN = 2.54D
double toInches(double length) {return length * CM_PER_IN;}
double toCentimeters(double length) {return length / CM_PER_IN;}
double toPounds(double weight) {return weight * LBS_PER_KG;}
double toKilograms(double weight) {return weight / LBS_PER_KG;}
}
At startup time, instead of loading a bunch of config_options values, initialize a set of configurable actions, as in:
MetricConverter converter = (metricOption()) ? new MetricConv() : new NullConv();
(where the expression metricOption() above is a stand-in for whatever one-time-only check you need to make, including looking at the value of METRIC_ENABLED ;-)
Then, wherever the code would have said:
double length = getLengthFromGui();
if (config_options.value('METRIC_ENABLED') == 'Y') {
length = length / 2.54D;
}
// do some computation to produce result
// ...
if (config_options.value('METRIC_ENABLED') == 'Y') {
result = result * 2.54D;
}
displayResultingLengthOnGui(result);
rewrite it as:
double length = converter.toInches(getLengthFromGui());
// do some computation to produce result
// ...
displayResultingLengthOnGui(converter.toCentimeters(result));
Because all of the implementation details related to that one concept are now packaged cleanly, all future maintenance related to METRIC_ENABLED can be done in one place. In addition, the run-time trade-off is a win; the "overhead" of invoking a method is trivial compared with the overhead of fetching a String value from a Map and performing String#equals.
I believe that the two reasons you have mentioned, Possible misspelling in string, that cannot be detected until run time and the possibility (although slim) of a name change would justify your idea.
On top of that you can get typed functions, now it seems you only store booleans, what if you need to store an int, a string etc. I would rather use get_foo() with a type, than get_string("FOO") or get_int("FOO").
I think there are two different issues here:
In the current project, the convention of using hard-coded strings is already well established, so all the developers working on the project are familiar with it. It might be a sub-optimal convention for all the reasons that have been listed, but everybody familiar with the code can look at it and instinctively knows what the code is supposed to do. Changing the code so that in certain parts, it uses the "new" functionality will make the code slightly harder to read (because people will have to think and remember what the new convention does) and thus a little harder to maintain. But I would guess that changing over the whole project to the new convention would potentially be prohibitively expensive unless you can quickly script the conversion.
On a new project, symbolic constants are the way IMO, for all the reasons listed. Especially because anything that makes the compiler catch errors at compile time that would otherwise be caught by a human at run time is a very useful convention to establish.
Another thing to consider is intent. If you are on a project that requires localization hard coded strings can be ambiguous. Consider the following:
const string HELLO_WORLD = "Hello world!";
print(HELLO_WORLD);
The programmer's intent is clear. Using a constant implies that this string does not need to be localized. Now look at this example:
print("Hello world!");
Here we aren't so sure. Did the programmer really not want this string to be localized or did the programmer forget about localization while he was writing this code?
I too prefer a strongly-typed configuration class if it is used through-out the code. With properly named methods you don't lose any readability. If you need to do conversions from strings to another data type (decimal/float/int), you don't need to repeat the code that does the conversion in multiple places and can cache the result so the conversion only takes place once. You've already got the basis of this in place already so I don't think it would take much to get used to the new way of doing things.