I am struggling to see the real-world benefits of loosely coupled code. Why spend so much effort making something flexible to work with a variety of other objects? If you know what you need to achieve, why not code specifically for that purpose?
To me, this is similar to creating untyped variables: it makes it very flexible, but opens itself to problems because perhaps an unexpected value is passed in. It also makes it harder to read, because you do not explicitly know what is being passed in.
Yet I feel like strongly typed is encouraged, but loosely coupling is bad.
EDIT: I feel either my interpretation of loose coupling is off or others are reading it the wrong way.
Strong coupling to me is when a class references a concrete instance of another class. Loose coupling is when a class references an interface that another class can implement.
My question then is why not specifically call a concrete instance/definition of a class? I analogize that to specifically defining the variable type you need.
I've been doing some reading on Dependency Injection, and they seem to make it out as fact that loose coupling better design.
First of all, you're comparing apples to oranges, so let me try to explain this from two perspectives. Typing refers to how operations on values/variables are performed and if they are allowed. Coupling, as opposed to cohesion, refers to the architecture of a piece (or several pieces) of software. The two aren't directly related at all.
Strong vs Weak Typing
A strongly typed language is (usually) a good thing because behavior is well defined. Take these two examples, from Wikipedia:
Weak typing:
a = 2
b = '2'
concatenate(a, b) # Returns '22'
add(a, b) # Returns 4
The above can be slightly confusing and not-so-well-defined because some languages may use the ASCII (maybe hex, maybe octal, etc) numerical values for addition or concatenation, so there's a lot of room open for mistakes. Also, it's hard to see if a is originally an integer or a string (this may be important, but the language doesn't really care).
Strongly typed:
a = 2
b = '2'
#concatenate(a, b) # Type Error
#add(a, b) # Type Error
concatenate(str(a), b) # Returns '22'
add(a, int(b)) # Returns 4
As you can see here, everything is more explicit, you know what variables are and also when you're changing the types of any variables.
Wikipedia says:
The advantage claimed of weak typing
is that it requires less effort on the
part of the programmer than, because
the compiler or interpreter implicitly
performs certain kinds of conversions.
However, one claimed disadvantage is
that weakly typed programming systems
catch fewer errors at compile time and
some of these might still remain after
testing has been completed. Two
commonly used languages that support
many kinds of implicit conversion are
C and C++, and it is sometimes claimed
that these are weakly typed languages.
However, others argue that these
languages place enough restrictions on
how operands of different types can be
mixed, that the two should be regarded
as strongly typed languages.
Strong vs weak typing both have their advantages and disadvantages and neither is good or bad. It's important to understand the differences and similarities.
Loose vs Tight Coupling
Straight from Wikipedia:
In computer science, coupling or
dependency is the degree to which each
program module relies on each one of
the other modules.
Coupling is usually contrasted with
cohesion. Low coupling often
correlates with high cohesion, and
vice versa. The software quality
metrics of coupling and cohesion were
invented by Larry Constantine, an
original developer of Structured
Design who was also an early proponent
of these concepts (see also SSADM).
Low coupling is often a sign of a
well-structured computer system and a
good design, and when combined with
high cohesion, supports the general
goals of high readability and
maintainability.
In short, low coupling is a sign of very tight, readable and maintainable code. High coupling is preferred when dealing with massive APIs or large projects where different parts interact to form a whole. Neither is good or bad. Some projects should be tightly coupled, i.e. an embedded operating system. Others should be loosely coupled, i.e. a website CMS.
Hopefully I've shed some light here :)
The question is right to point out that weak/dynamic typing is indeed a logical extension of the concept of loose coupling, and it is inconsistent for programmers to favor one but not the other.
Loose coupling has become something of a buzzword, with many programmers unnecessarily implementing interfaces and dependency injection patterns -- or, more often than not, their own garbled versions of these patterns -- based on the possibility of some amorphous future change in requirements. There is no hiding the fact that this introduces extra complexity and makes code less maintainable for future developers. The only benefit is if this anticipatory loose coupling happens to make a future change in requirements easier to implement, or promote code reuse. Often, however, requirements changes involve enough layers of the system, from UI down to storage, that the loose coupling doesn't improve the robustness of the design at all, and makes certain types of trivial changes more tedious.
You're right that loose coupling is almost universally considered "good" in programming. To understand why, let's look at one definition of tight coupling:
You say that A is tightly coupled to B if A must change just because B changed.
This is a scale that goes from "completely decoupled" (even if B disappeared, A would stay the same) to "loosely coupled" (certain changes to B might affect A, but most evolutionary changes wouldn't), to "very tightly coupled" (most changes to B would deeply affect A).
In OOP we use a lot of techniques to get less coupling - for example, encapsulation helps decouple client code from the internal details of a class. Also, if you depend on an interface then you don't generally have to worry as much about changes to concrete classes that implement the interface.
On a side note, you're right that typing and coupling are related. In particular, stronger and more static typing tend to increase coupling. For example, in dynamic languages you can sometimes substitute a string for an array, based on the notion that a string can be seen as an array of characters. In Java you can't, because arrays and strings are unrelated. This means that if B used to return an array and now returns a string, it's guaranteed to break its clients (just one simple contrived example, but you can come up with many more that are both more complex and more compelling). So, stronger typing and more static typing are both trade-offs. While stronger typing is generally considered good, favouring static versus dynamic typing is largely a matter of context and personal tastes: try setting up a debate between Python programmers and Java programmers if you want a good fight.
So finally we can go back to your original question: why is loose coupling generally considered good? Because of unforeseen changes. When you write the system, you cannot possibly know which directions it will eventually evolve in two months, or maybe two hours. This happens both because requirements change over time, and because you don't generally understand the system completely until after you've written it. If your entire system is very tightly coupled (a situation that's sometimes referred to as "the Big Ball of Mud"), then any change in every part of the system will eventually ripple through every other part of the system (the definition of "very tight coupling"). This makes for very inflexible systems that eventually crystallize into a rigid, unmaintanable blob. If you had 100% foresight the moment you start working on a system, then you wouldn't need to decouple.
On the other hand, as you observe, decoupling has a cost because it adds complexity. Simpler systems are easier to change, so the challenge for a programmer is striking a balance between simple and flexible. Tight coupling often (not always) makes a system simpler at the cost of making it more rigid. Most developers underestimate future needs for changes, so the common heuristic is to make the system less coupled than you're tempted to, as long as this doesn't make it overly complex.
Strongly typed is good because it prevents hard to find bugs by throwing compile-time errors rather than run-time errors.
Tightly coupled code is bad because when you think you "know what you need to achieve", you are often wrong, or you don't know everything you need to know yet.
i.e. you might later find out that something you've already done could be used in another part of your code. Then maybe you decide to tightly couple 2 different versions of the same code. Then later you have to make a slight change in a business rule and you have to alter 2 different sets of tightly coupled code, and maybe you will get them both correct, which at best will take you twice as long... or at worst you will introduce a bug in one, but not in the other, and it goes undetected for a while, and then you find yourself in a real pickle.
Or maybe your business is growing much faster than you expected, and you need to offload some database components to a load-balancing system, so now you have to re-engineer everything that is tightly coupled to the existing database system to use the new system.
In a nutshell, loose coupling makes for software that is much easier to scale, maintain, and adapt to ever-changing conditions and requirements.
EDIT: I feel either my interpretation
of loose coupling is off or others are
reading it the wrong way. Strong
coupling to me is when a class
references a concrete instance of
another class. Loose coupling is when
a class references an interface that
another class can implement.
My question then is why not
specifically call a concrete
instance/definition of a class? I
analogize that to specifically
defining the variable type you need.
I've been doing some reading on
Dependency Injection, and they seem to
make it out as fact that loose
coupling better design.
I'm not really sure what your confusion is here. Let's say for instance that you have an application that makes heavy use of a database. You have 100 different parts of your application that need to make database queries. Now, you could use MySQL++ in 100 different locations, or you can create a separate interface that calls MySQL++, and reference that interface in 100 different places.
Now your customer says that he wants to use SQL Server instead of MySQL.
Which scenario do you think is going to be easier to adapt? Rewriting the code in 100 different places, or rewriting the code in 1 place?
Okay... now you say that maybe rewriting it in 100 different places isn't THAT bad.
So... now your customer says that he needs to use MySQL in some locations, and SQL Server in other locations, and Oracle in yet other locations.
Now what do you do?
In a loosely coupled world, you can have 3 separate database components that all share the same interface with different implementations. In a tightly coupled world, you'd have 100 sets of switch statements strewn with 3 different levels of complexity.
If you know what you need to achieve, why not code specifically for that purpose.
Short answer: You almost never know exactly what you need to achieve. Requirements change, and if your code is loosely coupled in the first place, it will be less of a nightmare to adapt.
Yet I feel like strongly typed is encouraged, but loosely coupling is bad.
I don't think it is fair to say that strong typing is good or encouraged. Certainly lots of people prefer strongly typed languages because it comes with compile-time checking. But plenty of people would say that weak typing is good. It sounds like since you've heard "strong" is good, how can "loose" be good too. The merits of a language's typing system isn't even in the realm of a similar concept as class design.
Side note: don't confuse strong and static typing
strong typing will help reduce errors while typically aiding performance. the more information the code-generation tools can gather about acceptable value ranges for variables, the more these tools can do to generate fast code.
when combined with type inference and feature's like traits (perl6 and others) or type classes (haskell), strongly typed code can continue to be compact and elegant.
I think that tight/loose coupling (to me: Interface declaration and assignment of an object instance) is related to the Liskov Principle. Using loose coupling enables some of the advantages of the Liskov Principle.
However, as soon as instanceof, cast or copying operations are executed, the usage of loose coupling starts being questionable. Furthermore, for local variables withing a method or block, it is non-sense.
If any modification done in our function, which is in a derived class, will change the code in the base abstract class, then this shows the full dependency and it means this is tight coupled.
If we don't write or recompile the code again then it showes the less dependency, hence it is loose coupled.
Related
Is it at all practical to implement an object model in functional style?
One problem that OOP seems to excel at is describing object models.
For instance, an HTML DOM is a complicated, stateful beast which interfaces directly with the UI and requires programmability from dynamic languages. OOP features tend to come in useful in a lot of ways:
Member access constraints make interfacing with untrusted code (e.g. javascript) safe
Accessor functions and properties make binding to the UI much more convenient
Not having to pass around the object model all the time makes methods a lot simpler.
The UI side of the story might be a bit moot if you're projecting the model via MVVM, but you're still constantly wrestling with state internally.
I'm working in F# for this project so I could easily resort to OOP, but I'm curious as to how far I can push it before it becomes impractical. Are there maybe design patterns or anything?
This is a bit philosophical to have a "correct" answer but okay I'll bite.
In my opinion the problem comes because you consider FP and OO to be juxtapose, they are not. FP and imperative programming are juxtaposed, i.e. using expressions versus using statements.
Part of the problem is that OO lacks a clear definition, well in my opinion anyway. To support this I'd point to Alan Kay who said “Actually I made up the term "object-oriented", and I can tell you I did not have C++ in mind.”, yet most language we consider OO i.e. java/C# take more after C++ than smalltalk.
What OO C++/java/C# style does give us is a nice way to organize our code into models, create data contains add properties to them etc. Non of this is practically un-functional and can be used nice with functional programming.
As you point out a lot of C++/java/C# tend to be stateful, but they don’t have to be, both java and C# have fundamental types such as their string classes that are immutable. It’s true java and C# don’t make it easy to create immutable class but with a bit of effort you can do it.
Which brings us to where is immutable appropriates? In my designs usually start of by making everything immutable, since this makes getting things correct easier, and if I see this causing performance problems I start adding some mutability on the critical paths. The one place immutability is never going to work is GUI controls, which generally contain far too much state to be immutable. Having said that you can get quite a long way building GUI using a immutable “combinator” approach then is then interpreted by mutable gui controls. This is more or less what the WebSharper guy’s do: http://www.intellifactory.com/products/wsp/Home.aspx
Another great resource for the FP/OO debate is Brain’s “How does functional programming affect the structure of your code?” (which greatly influenced my thinking about FP/OO): http://lorgonblog.wordpress.com/2008/09/22/how-does-functional-programming-affect-the-structure-of-your-code/
I've always reversed names so that they naturally group in intellisense. I am wondering if this is a bad idea.
For example, I run a pet store and I have invoicing pages add, edit, delete, and store pages display, preview, edit. To get the URL for these, I would call the methods (in a suitable class like GlobalUrls.cs
InvoicingAddUrl()
InvoicingEditUrl()
InvoicingDeleteUrl()
StoreDisplayUrl()
StorePreviewUrl()
StoreEditUrl()
This groups them nicely in intellisense. More logical naming would be:
AddInvoiceUrl()
EditInvoiceUrl()
DeleteInvoiceUrl()
DisplayStoreUrl()
PreviewStoreUrl()
EditStoreUrl()
Is it better (better being, more of an industry standard way) to group them for intellisense, or logically?
Grouping in Intellisense is just one factor in creating a naming scheme, but logically grouping by category rather than function is a common practice as well.
Most naming "conventions" dictate usage of characters, casing, underscores, etc. I think it is a matter of personal preference (company, team or otherwise) as to whether you use NounVerb or VerbNoun formatting for your method names.
Here are some resources:
Microsoft - General Naming Conventions
Wikibooks C# Programming/Naming
Akadia .NET Naming Conventions
Related questions:
Naming Conventions - Guidelines for Verbs, Nouns and English Grammar Usage
Do vs. Run vs. Execute vs. Perform verbs
Events - naming convention and style
Check out how the military names things. For example, MREs are Meals, Ready to Eat. They do this because of sort order, efficiency and not making mistakes. They are ready to ignore the standard naming conventions of the language (i.e., English) used outside of their organization because they are not impressed with the quality of operations outside of their organization. In the military, the quality of operations is literally a matter of life and death. Also, by doing things their own way they have a way of identifying who is inside and who is outside of the organization. Anyone unable or unwilling to learn the military way, which is different but not impossibly difficult, is not their first choice for recruitment or promotion.
So, if you are impressed with the standard quality of software out there, then by all means keep doing what everyone else is doing. But, if you wish to do better than you have in the past, or better than your competitor, then I suggest looking at other fields for lessons learned the hard way, such as the military. Then make some choices for your organization, that are not impossible but are for you and your competitiveness. You can choose big-endian names (most significant information comes last) or the military-style little-endian names (most significant information comes first), or you can use the dominant style your competitors probably use, which is doing whatever you feel like whenever you feel like it.
Personally, I prefer little-endian Hungarian (Apps) naming, which was widely seen as superior when it first came out, but then lost favor because Hungarian (Sys) naming destroyed the advantage due to a mistranslation of the basic idea, and because of rampant abbreviations. The original intent was to start a name with what kind of a thing it is, then become increasingly specific until you end with a unique qualification. This is also the order that most array dimensions and object qualifiers are in, so in most languages little-endian naming flows into the larger scheme of the language.
You are on to something. Forward, march.
It's not intrinsically bad. It has the upside of being easier to identify the type while scanning, and groups the options together in Intellisense like you said. As long as you and everyone else on your team picks a way of doing things and stays consistent about it there shouldn't be any big problems.
Based on the methods listed, you might be able to refactor Invoicing and Store out into their own classes, which would be closer to the mythical "industry standard" way.
That said, whatever your programming team can agree on for naming convention should be fine. The important thing is to be consistent within the project.
I don't think it's a good idea to develop a coding standard around a tool (as least not as the first consideration). Even though most IDEs will have Intellisense these days, and most people will be using said IDEs, I think that first and foremost a coding standard should be about making the code legible and navigable on its own merits.
I would opt for most logical naming, personally. When I write code and I have some object I'm about to call a member function on, I'm usually thinking about what member function to call based on the action I'm about to do, because I already know the object I'm manipulating. So my first impulse would be to start typing "Add" if I wanted to add something, and see what Intellisense showed me. This is, of course, subjective.
I have never actually seen anybody using your alphabetical, Intellisense grouping anywhere -- at least not in code that is not worth using as a basis for comparison because it was so horrid in other ways.
That said, if it's your standard, do what you want -- consistency is the important part.
It seems to me that the most invaluable thing about a static/strongly-typed programming language is that it helps refactoring: if/when you change any API, then the compiler will tell you what that change has broken.
I can imagine writing code in a runtime/weakly-typed language ... but I can't imagine refactoring without the compiler's help, and I can't imagine writing tens of thousands of lines of code without refactoring.
Is this true?
I think you're conflating when types are checked with how they're checked. Runtime typing isn't necessarily weak.
The main advantage of static types is exactly what you say: they're exhaustive. You can be confident all call sites conform to the type just by letting the compiler do it's thing.
The main limitation of static types is that they're limited in the constraints they can express. This varies by language, with most languages having relatively simple type systems (c, java), and others with extremely powerful type systems (haskell, cayenne).
Because of this limitation types on their own are not sufficient. For example, in java types are more or less restricted to checking type names match. This means the meaning of any constraint you want checked has to be encoded into a naming scheme of some sort, hence the plethora of indirections and boiler plate common to java code. C++ is a little better in that templates allow a bit more expressiveness, but don't come close to what you can do with dependent types. I'm not sure what the downsides to the more powerful type systems are, though clearly there must be some or more people would be using them in industry.
Even if you're using static typing, chances are it's not expressive enough to check everything you care about, so you'll need to write tests too. Whether static typing saves you more effort than it requires in boilerplate is a debate that's raged for ages and that I don't think has a simple answer for all situations.
As to your second question:
How can we re-factor safely in a runtime typed language?
The answer is tests. Your tests have to cover all the cases that matter. Tools can help you in gauging how exhaustive your tests are. Coverage checking tools let you know wether lines of code are covered by the tests or not. Test mutation tools (jester, heckle) can let you know if your tests are logically incomplete. Acceptance tests let you know what you've written matches requirements, and lastly regression and performance tests ensure that each new version of the product maintains the quality of the last.
One of the great things about having proper testing in place vs relying on elaborate type indirections is that debugging becomes much simpler. When running the tests you get specific failed assertions within tests that clearly express what they're doing, rather than obtuse compiler error statements (think c++ template errors).
No matter what tools you use: writing code you're confident in will require effort. It most likely will require writing a lot of tests. If the penalty for bugs is very high, such as aerospace or medical control software, you may need to use formal mathematical methods to prove the behavior of your software, which makes such development extremely expensive.
I totally agree with your sentiment. The very flexibility that dynamically typed languages are supposed to be good at is actually what makes the code very hard to maintain. Really, is there such a thing as a program that continues to work if the data types are changed in a non trivial way without actually changing the code?
In the mean time, you could check the type of variable being passed, and somehow fail if its not the expected type. You'd still have to run your code to root out those cases, but at least something would tell you.
I think Google's internal tools actually do a compilation and probably type checking to their Javascript. I wish I had those tools.
To start, I'm a native Perl programmer so on the one hand I've never programmed with the net of static types. OTOH I've never programmed with them so I can't speak to their benefits. What I can speak to is what its like to refactor.
I don't find the lack of static types to be a problem wrt refactoring. What I find a problem is the lack of a refactoring browser. Dynamic languages have the problem that you don't really know what the code is really going to do until you actually run it. Perl has this more than most. Perl has the additional problem of having a very complicated, almost unparsable, syntax. Result: no refactoring tools (though they're working very rapidly on that). The end result is I have to refactor by hand. And that is what introduces bugs.
I have tests to catch them... usually. I do find myself often in front of a steaming pile of untested and nigh untestable code with the chicken/egg problem of having to refactor the code in order to test it, but having to test it in order to refactor it. Ick. At this point I have to write some very dumb, high level "does the program output the same thing it did before" sort of tests just to make sure I didn't break something.
Static types, as envisioned in Java or C++ or C#, really only solve a small class of programming problems. They guarantee your interfaces are passed bits of data with the right label. But just because you get a Collection doesn't mean that Collection contains the data you think it does. Because you get an integer doesn't mean you got the right integer. Your method takes a User object, but is that User logged in?
Classic example: public static double sqrt(double a) is the signature for the Java square root function. Square root doesn't work on negative numbers. Where does it say that in the signature? It doesn't. Even worse, where does it say what that function even does? The signature only says what types it takes and what it returns. It says nothing about what happens in between and that's where the interesting code lives. Some people have tried to capture the full API by using design by contract, which can broadly be described as embedding run-time tests of your function's inputs, outputs and side effects (or lack thereof)... but that's another show.
An API is far more than just function signatures (if it wasn't, you wouldn't need all that descriptive prose in the Javadocs) and refactoring is far more even than just changing the API.
The biggest refactoring advantage a statically typed, statically compiled, non-dynamic language gives you is the ability to write refactoring tools to do quite complex refactorings for you because it knows where all the calls to your methods are. I'm pretty envious of IntelliJ IDEA.
I would say refactoring goes beyond what the compiler can check, even in statically-typed languages. Refactoring is just changing a programs internal structure without affecting the external behavior. Even in dynamic languages, there are still things that you can expect to happen and test for, you just lose a little bit of assistance from the compiler.
One of the benefits of using var in C# 3.0 is that you can often change the type without breaking any code. The type needs to still look the same - properties with the same names must exist, methods with the same or similar signature must still exist. But you can really change to a very different type, even without using something like ReSharper.
I often heard from professionals blog something like refactoring your code whenever the chance you get. What is it exactly? Rewriting your code in simpler and fewer lines? What is the purpose of doing this?
Refactoring code is a process of cleaning up your code, reducing the clutter and improving the readability without causing any side effects or changes to features.
Basically, you refactor by applying a series of code change rules that improve code readability and re-usability, without affecting the logic.
Always unit test before and after refactoring to ensure your logic isn't affected.
This Wikipedia article will give you an idea of the types of things included in the general concept of Refactoring.
The idea is adapt / evolve your code as you go. Simple things may be to rename variables or method parameters, but others may be to pass an additional parameter or to drop one, or to change its type. The data model may evolve as well. etc.
Often refactoring, works hand-in-hand with unit-testing, whereby the risk of "breaking something" is offset by the fact that such an issue may likely be discovered by the automatic testing (provide a good coverage and relevant test cases...).
In a nutshell, the ability to refactor (and btw, most IDE or add-ons to the IDEs, offer various tools that make refactoring easier and less error prone) allows one to write more quickly without stressing about some decisions ("should this object include an array or a list etc...) letting the programmer change some of these decisions as times goes, and with the added insight offered by having a workable, if not perfect solution. See a related concept: agile development.
Beware, refactoring doesn't give you license to start coding without putting any thought in design, in the object model, the APIs etc., however it lessens the stiffness of some of these decisions.
Martin Fowler has probably done the most to popularize refactoring, but I think good developers have always done these sorts of restructurings. Check out Fowler'srefactoring web site, and his 1999 Refactoring, which is an excellent introduction and catalog of specific refactorings using Java.
And I see he's a co-author of the brand new Refactoring, Ruby Edition, which should be a great resource.
I find that regularly cleaning up your code like this makes it a lot clearer and more maintainable.
To take one example, I wrote a small (Java 1.6) client library for accessing remote web services (using the REST architectural style). The bulk of this library is in one source file, and about half of that deals with the web services, while the other half is a simple in-memory cache of the responses (for performance). Over time both halves have grown in functionality, to the point where the source file was getting too complex. So today I used Fowler's "Extract Class" refactoring to move the cache logic into a new class. Before that I had to do some "Extract Methods" to isolate the caching logic. Along the way I did a few "Rename Methods" and an "Introduce Explaining Variable".
As other folks have noted, it's very important to have a good set of unit tests to apply after you make each change. They help ensure that you're not introducing new bugs, among other good things.
In a nutshell, refactoring means improving the design and/or implementation of software, usually without changing its behavior. This is normally done to make the code easier to understand and work with going forward, thereby making future development faster and less bug-prone.
Refactoring is a long-term investment in your code - since it doesn't affect the outward "appearance" of the software, there is very often pressure (from management, etc.) to "just get it working and move on to the next thing." While this may sometimes be the right decision, depending on business drivers, a codebase that undergoes change but never gets refactored will decay into a difficult, buggy mess (See also Technical Debt).
Specifically, the top reasons to refactor are usually the following:
Getting rid of duplicated code
Breaking up a long method into smaller pieces by extracting new methods from sections of the longer method
Breaking up a class that has too many responsibilities into smaller, more targeted classes or subclasses
Moving methods from one class to another. Often this is done so the methods reside in the same class as the data they operate on.
In the simplest terms, refactoring code is optimizing code. The criteria for what is "better" code is open to much interpretation as there are various coding styles and patterns out there. A central idea with refactoring is the question of, "Could this code be made better?" A few examples of that criteria can include scalability, maintainability, readablity, performance, size of executable, or minimizing memory used in executing the code.
"Refactoring is the process of changing a software system in such a way that it does not alter the external behavior of the code yet improves its internal structure." -- MartinFowler in RefactoringImprovingTheDesignOfExistingCode
see this WhatIsRefactoring for more explanation.
Refactoring code generally means taking code that has been patched multiple times and re-writing it so that the needs of the later patches are taken into account.
In day to day programs I wouldn't even bother thinking about the possible performance hit for coding against interfaces rather than implementations. The advantages largely outweigh the cost. So please no generic advice on good OOP.
Nevertheless in this post, the designer of the XNA (game) platform gives as his main argument to not have designed his framework's core classes against an interface that it would imply a performance hit. Seeing it is in the context of a game development where every fps possibly counts, I think it is a valid question to ask yourself.
Does anybody have any stats on that? I don't see a good way to test/measure this as don't know what implications I should bear in mind with such a game (graphics) object.
Coding to an interface is always going to be easier, simply because interfaces, if done right, are much simpler. Its palpably easier to write a correct program using an interface.
And as the old maxim goes, its easier to make a correct program run fast than to make a fast program run correctly.
So program to the interface, get everything working and then do some profiling to help you meet whatever performance requirements you may have.
What Things Cost in Managed Code
"There does not appear to be a significant difference in the raw cost of a static call, instance call, virtual call, or interface call."
It depends on how much of your code gets inlined or not at compile time, which can increase performance ~5x.
It also takes longer to code to interfaces, because you have to code the contract(interface) and then the concrete implementation.
But doing things the right way always takes longer.
First I'd say that the common conception is that programmers time is usually more important, and working against implementation will probably force much more work when the implementation changes.
Second with proper compiler/Jit I would assume that working with interface takes a ridiculously small amount of extra time compared to working against the implementation itself.
Moreover, techniques like templates can remove the interface code from running.
Third to quote Knuth : "We should forget about small efficiencies, say about 97% of the time: premature optimization is the root of all evil."
So I'd suggest coding well first, and only if you are sure that there is a problem with the Interface, only then I would consider changing.
Also I would assume that if this performance hit was true, most games wouldn't have used an OOP approach with C++, but this is not the case, this Article elaborates a bit about it.
It's hard to talk about tests in a general form, naturally a bad program may spend a lot of time on bad interfaces, but I doubt if this is true for all programs, so you really should look at each particular program.
Interfaces generally imply a few hits to performance (this however may change depending on the language/runtime used):
Interface methods are usually implemented via a virtual call by the compiler. As another user points out, these can not be inlined by the compiler so you lose that potential gain. Additionally, they add a few instructions (jumps and memory access) at a minimum to get the proper PC in the code segment.
Interfaces, in a number of languages, also imply a graph and require a DAG (directed acyclic graph) to properly manage memory. In various languages/runtimes you can actually get a memory 'leak' in the managed environment by having a cyclic graph. This imposes great stress (obviously) on the garbage collector/memory in the system. Watch out for cyclic graphs!
Some languages use a COM style interface as their underlying interface, automatically calling AddRef/Release whenever the interface is assigned to a local, or passed by value to a function (used for life cycle management). These AddRef/Release calls can add up and be quite costly. Some languages have accounted for this and may allow you to pass an interface as 'const' which will not generate the AddRef/Release pair automatically cutting down on these calls.
Here is a small example of a cyclic graph where 2 interfaces reference each other and neither will automatically be collected as their refcounts will always be greater than 1.
interface Parent {
Child c;
}
interface Child {
Parent p;
}
function createGraph() {
...
Parent p = ParentFactory::CreateParent();
Child c = ChildFactory::CreateChild();
p.c = c;
c.p = p;
... // do stuff here
// p has a reference to c and c has a reference to p.
// When the function goes out of scope and attempts to clean up the locals
// it will note that p has a refcount of 1 and c has a refcount of 1 so neither
// can be cleaned up (of course, this is depending on the language/runtime and
// if DAGS are allowed for interfaces). If you were to set c.p = null or
// p.c = null then the 2 interfaces will be released when the scope is cleaned up.
}
I think object lifetime and the number of instances you're creating will provide a coarse-grain answer.
If you're talking about something which will have thousands of instances, with short lifetimes, I would guess that's probably better done with a struct rather than a class, let alone a class implementing an interface.
For something more component-like, with low numbers of instances and moderate-to-long lifetime, I can't imagine it's going to make much difference.
IMO yes, but for a fundamental design reason far more subtle and complex than virtual dispatch or COM-like interface queries or object metadata required for runtime type information or anything like that. There is overhead associated with all of that but it depends a lot on the language and compiler(s) used, and also depends on whether the optimizer can eliminate such overhead at compile-time or link-time. Yet in my opinion there's a broader conceptual reason why coding to an interface implies (not guarantees) a performance hit:
Coding to an interface implies that there is a barrier between you and
the concrete data/memory you want to access and transform.
This is the primary reason I see. As a very simple example, let's say you have an abstract image interface. It fully abstracts away its concrete details like its pixel format. The problem here is that often the most efficient image operations need those concrete details. We can't implement our custom image filter with efficient SIMD instructions, for example, if we had to getPixel one at a time and setPixel one at a time and while oblivious to the underlying pixel format.
Of course the abstract image could try to provide all these operations, and those operations could be implemented very efficiently since they have access to the private, internal details of the concrete image which implements that interface, but that only holds up as long as the image interface provides everything the client would ever want to do with an image.
Often at some point an interface cannot hope to provide every function imaginable to the entire world, and so such interfaces, when faced with performance-critical concerns while simultaneously needing to fulfill a wide range of needs, will often leak their concrete details. The abstract image might still provide, say, a pointer to its underlying pixels through a pixels() method which largely defeats a lot of the purpose of coding to an interface, but often becomes a necessity in the most performance-critical areas.
Just in general a lot of the most efficient code often has to be written against very concrete details at some level, like code written specifically for single-precision floating-point, code written specifically for 32-bit RGBA images, code written specifically for GPU, specifically for AVX-512, specifically for mobile hardware, etc. So there's a fundamental barrier, at least with the tools we have so far, where we cannot abstract that all away and just code to an interface without an implied penalty.
Of course our lives would become so much easier if we could just write code, oblivious to all such concrete details like whether we're dealing with 32-bit SPFP or 64-bit DPFP, whether we're writing shaders on a limited mobile device or a high-end desktop, and have all of it be the most competitively efficient code out there. But we're far from that stage. Our current tools still often require us to write our performance-critical code against concrete details.
And lastly this is kind of an issue of granularity. Naturally if we have to work with things on a pixel-by-pixel basis, then any attempts to abstract away concrete details of a pixel could lead to a major performance penalty. But if we're expressing things at the image level like, "alpha blend these two images together", that could be a very negligible cost even if there's virtual dispatch overhead and so forth. So as we work towards higher-level code, often any implied performance penalty of coding to an interface diminishes to a point of becoming completely trivial. But there's always that need for the low-level code which does do things like process things on a pixel-by-pixel basis, looping through millions of them many times per frame, and there the cost of coding to an interface can carry a pretty substantial penalty, if only because it's hiding the concrete details necessary to write the most efficient implementation.
In my personal opinion, all the really heavy lifting when it comes to graphics is passed on to the GPU anwyay. These frees up your CPU to do other things like program flow and logic. I am not sure if there is a performance hit when programming to an interface but thinking about the nature of games, they are not something that needs to be extendable. Maybe certain classes but on the whole I wouldn't think that a game needs to programmed with extensibility in mind. So go ahead, code the implementation.
it would imply a performance hit
The designer should be able to prove his opinion.