Lately, I've been making use a lot of the Strategy Pattern along with the Factory Pattern. And I really mean a lot. I have a lot of "algorithms" for everything and factories that retrieve algorithms based on parameters.
Even though the code seems very extensible, and it is, having N factories seems a bit of an abuse.
I know this is pretty subjective, and we're talking without seeing code, but is this acceptable in real world code? Would you change something ?
OK- ask yourself a question. does/will this algorithm implementation ever change? If no then remove the strategy.
I am maintence.
I once was forced (by my pattern lovin' boss) to write a set of 16 "buffer interpreter tuxedo services" using an AbstractFactory and a double DAO pattern in C++ (no reflections, no code gen). All up it something like 20,000 lines of the nastiest code I've even seen (not the least because I didn't really know C++ when I started) and it took about three months.
Since my old boss has moved on I've rewritten them using good 'ole "straight up and down" procedural style C++, with couple of funky-macros... each service is like 60 lines of code, times 16... all up less than a 1000 lines of really SIMPLE code; so simple that even I can follow it.
Cheers. Keith.
Whenever I'm implementing code in this fashion, some questions I ask are:
what components do I need to substitute to test ?
what components will I expect users/admins to disable or substitute (e.g. via Spring configs or similar) ?
what components do I expect or suspect will not be required in the future due to (possibly) changing requirements ?
This all drives how I construct object or components (via factories) and how I implement algorithms. The above is vague, but (of course) the requirements can be similarly difficult to pin down. Without seeing your implementation and your requirements, I can't comment further, but the above should act as some guideline to help you determine whether what you've done is overkill.
If you're using the same design pattern all over the place, perhaps you should either switch to a language that has better support for what you're trying to do or rethink your code to be more idiomatic in your language of choice. After all, that's why we have more than one programming language.
Would depends on the kind of software I'm working on.
Maintenance asks for simple code and factories is NOT simple code.
But extensibility asks sometimes for factories...
So you have to take both in consideration.
Just have in mind that most of the time, you will have to maintain a source file MANY times a year and you will NOT have to extend it.
IMO, patterns should only be used when absolutely needed. If you think it can be handy in two years, you are better to use them... in two years.
How complex is the work the factory is handling? Does object creation really need to be abstracted to a different class? A variation of the factory method is having a simple, in-class factory. This really works best if any dependencies have already been injected.
For instance,
public class Customer
{
public Customer CreateNewCustomer()
{
// handle minimally complex create logic here
}
}
As far as Strategy overuse... Again, as #RichardOD explained, will the algorithm ever really change?
Always keep in mind the YAGNI principle. You Aren't Gonna Need It.
Can't you make an AbstractFactory instead off different standalone factories?
AlgorithmFactory creates the algorithms you need based on the concrete factory.
What are the possible and critical disadvantages of Aspect-Oriented Programming?
For example: cryptic debugging for newbies (readability impact)
I think the biggest problem is that nobody knows how to define the semantics of an aspect, or how to declare join points non-procedurally.
If you can't define what an aspect does independently of the context in which it will be embedded, or define the effects that it has in such a way that it doesn't damage the context in which it is embedded, you (and tools) can't reason about what it does reliably. (You'll note the most common example of aspects is "logging", which is defined as "write some stuff to a log stream the application doesn't know anything about", because this is pretty safe). This violates David Parnas' key notion of information hiding. One of the worst examples of aspects that I see are ones that insert synchronization primitives into code; this effects the sequence of possible interactions that code can have. How can you possibly know this is safe (won't deadlock? won't livelock? won't fail to protect? is recoverable in the face of a thrown exception in a synchronization partner?) unless the application only does trivial things already.
Join points are now typically defined by providing some kind of identifier wildcarding (e.g, "put this aspect into any method named "DataBaseAccess*". For this to work, the folks writing the affected methods have to signal the intention for their code to be aspectized by naming their code in a funny way; that's hardly modular. Worse, why should the victim of an aspect even have to know that it exists? And consider what happens if you simply rename some methods; the aspect is no longer injected where it is needed, and your application breaks. What is needed are join point specifications which are intentional; somehow, the aspect has to know where it is needed without the programmers placing a neon sign at each usage point. (AspectJ has some control-flow related join points which seem a bit better in this regard).
So aspects are kind of interesting idea, but I think they are technologically immature. And that immaturity makes their usage fragile. And that's where the problem lies.
(I'm a big fan of automated software engineering tools [see my bio] just not this way).
Poor toolchain support - debuggers, profilers etc may not know about the AOP and so may work on code as if all the aspects had been replaced by procedural code
Code bloat - small source can lead to much larger object code as code is "weaved" throughout the code base
I think the biggest disadvantage is using AOP well. People use it in places where it doesn't make sense, for example, and will not use it where it does.
For example, a factory pattern is obviously something that AOP can do better, as DI can also do it well, but the observer pattern is simpler when using AOP, as is the strategy pattern.
It will be harder to unit test, esp if you do the weaving at runtime.
If weaving at runtime then you also take a performance hit.
Having a good way to model what is going on when mixing AOP with classes is a problem, as UML I don't think is a good model at that point.
Unless you are using Eclipse then tools do have problems, but, with Eclipse and AJDT AOP is much easier.
We still use junit and nunit for example, and so have to modify our code to allow unit tests to run, when using priviledged mode AOP could do better unit tests by testing private methods also, and we don't have to change our programs just to make them work with unit-testing. This is another example of not really understanding how AOP can be helpful, we are still chained in many ways to the past with unit-test frameworks and current design pattern implementations and don't see how AOP could help us do better coding.
Maintenance and Debugging. With aop, you suddenly have code that is being run at a given point ( method entry, exit, whatever ) but in just looking at the code, you have no clue that it's even getting called, especially if the aop configuration is in another file, like xml config. If the advice causes some changes, then while debugging an application, things may look strange with no explanation. This doesn't affect only newbies.
I would not call it a critical disadvantage, but the biggest problem I have seen is one of developer experience and ability to adapt. Not all developers yet understand the difference between declarative and imperative programming.
We make use of the policy injection application block in EntLib 4.1 pretty extensively as well as Unity for DI and it's just not something that sinks in quickly for some people. I find my self explaining over and over again to the same people why the application is not behaving in the way they expect. It generally starts off them explaining something and me saying "see that declaration above the method." :) Some people get it right away, love it and become extremely productive -- others struggle.
Learning curve is not unique to AOP, but it seems to have a higher learning curve that other things that your average developer encounters.
With regard to the maintenance/debugging argument, aspect-oriented programming tends to go hand-in-hand with all the other aspects of agile software-development practices.
These practices tend to remove debugging from the picture, replacing it with unit testing and test-driven development.
In addition, it may be much easier to maintain a small, clear code footprint with advice than a large, incomprehensible code footprint without advice (the advice being the thing that transforms a large, incomprehensible code footprint into a small, clear code footprint).
Because the power of AOP, if there is a bug your crosscutting, it can cause to widespread problems. On the otherhand, someone might change the join points in a program – e.g., by renaming or moving methods – in ways that the aspect writer did not expect, with unintended consequences. One advantage of modularizing crosscutting concerns is enabling one programmer to affect the entire system easily.
As it currently stands, this question is not a good fit for our Q&A format. We expect answers to be supported by facts, references, or expertise, but this question will likely solicit debate, arguments, polling, or extended discussion. If you feel that this question can be improved and possibly reopened, visit the help center for guidance.
Closed 10 years ago.
So I'm working on this class that's supposed to request help documentation from a vendor through a web service. I try to name it DocumentRetriever, VendorDocRequester, DocGetter, but they just don't sound right. I ended up browsing through dictionary.com for half an hour trying to come up with an adequate word.
Start programming with bad names is like having a very bad hair day in the morning, the rest of the day goes downhill from there. Feel me?
What you are doing now is fine, and I highly recommend you stick with your current syntax, being:
context + verb + how
I use this method to name functions/methods, SQL stored procs, etc. By keeping with this syntax, it will keep your Intellisense/Code Panes much more neat. So you want EmployeeGetByID() EmployeeAdd(), EmployeeDeleteByID(). When you use a more grammatically correct syntax such as GetEmployee(), AddEmployee() you'll see that this gets really messy if you have multiple Gets in the same class as unrelated things will be grouped together.
I akin this to naming files with dates, you want to say 2009-01-07.log not 1-7-2009.log because after you have a bunch of them, the order becomes totally useless.
One lesson I have learned, is that if you can't find a name for a class, there is almost always something wrong with that class:
you don't need it
it does too much
A good naming convention should minimize the number of possible names you can use for any given variable, class, method, or function. If there is only one possible name, you'll never have trouble remembering it.
For functions and for singleton classes, I scrutinize the function to see if its basic function is to transform one kind of thing into another kind of thing. I'm using that term very loosely, but you'll discover that a HUGE number of functions that you write essentially take something in one form and produce something in another form.
In your case it sounds like your class transforms a Url into a Document. It's a little bit weird to think of it that way, but perfectly correct, and when you start looking for this pattern, you'll see it everywhere.
When I find this pattern, I always name the function xFromy.
Since your function transforms a Url into a Document, I would name it
DocumentFromUrl
This pattern is remarkably common. For example:
atoi -> IntFromString
GetWindowWidth -> WidthInPixelsFromHwnd // or DxFromWnd if you like Hungarian
CreateProcess -> ProcessFromCommandLine
You could also use UrlToDocument if you're more comfortable with that order. Whether you say xFromy or yTox is probably a matter of taste, but I prefer the From order because that way the beginning of the function name already tells you what type it returns.
Pick one convention and stick to it. If you are careful to use the same names as your class names in your xFromy functions, it'll be a lot easier to remember what names you used. Of course, this pattern doesn't work for everything, but it does work where you're writing code that can be thought of as "functional."
Sometimes there isn't a good name for a class or method, it happens to us all. Often times, however, the inability to come up with a name may be a hint to something wrong with your design. Does your method have too many responsibilities? Does your class encapsulate a coherent idea?
Thread 1:
function programming_job(){
while (i make classes){
Give each class a name quickly; always fairly long and descriptive.
Implement and test each class to see what they really are.
while (not satisfied){
Re-visit each class and make small adjustments
}
}
}
Thread 2:
while(true){
if (any code smells bad){
rework, rename until at least somewhat better
}
}
There's no Thread.sleep(...) anywhere here.
I do spend a lot of time as well worrying about the names of anything that can be given a name when I am programming. I'd say it pays off very well though. Sometimes when I am stuck I leave it for a while and during a coffee break I ask around a bit if someone has a good suggestion.
For your class I'd suggest VendorHelpDocRequester.
The book Code Complete by Steve Mcconnell has a nice chapter on naming variables/classes/functions/...
I think this is a side effect.
It's not the actual naming that's hard. What's hard is that the process of naming makes you face the horrible fact that you have no idea what the hell you're doing.
I actually just heard this quote yesterday, through the Signal vs. Noise blog at 37Signals, and I certainly agree with it:
"There are only two hard things in Computer Science: cache invalidation and naming things."
— Phil Karlton
It's good that it's difficult. It's forcing you to think about the problem, and what the class is actually supposed to do. Good names can help lead to good design.
Agreed. I like to keep my type names and variables as descriptive as possible without being too horrendously long, but sometimes there's just a certain concept that you can't find a good word for.
In that case, it always helps me to ask a coworker for input - even if they don't ultimately help, it usually helps me to at least explain it out loud and get my wheels turning.
I was just writing on naming conventions last month: http://caseysoftware.com/blog/useful-naming-conventions
The gist of it:
verbAdjectiveNounStructure - with Structure and Adjective as optional parts
For verbs, I stick to action verbs: save, delete, notify, update, or generate. Once in a while, I use "process" but only to specifically refer to queues or work backlogs.
For nouns, I use the class or object being interacted with. In web2project, this is often Tasks or Projects. If it's Javascript interacting with the page, it might be body or table. The point is that the code clearly describes the object it's interacting with.
The structure is optional because it's unique to the situation. A listing screen might request a List or an Array. One of the core functions used in the Project List for web2project is simply getProjectList. It doesn't modify the underlying data, just the representation of the data.
The adjectives are something else entirely. They are used as modifiers to the noun. Something as simple as getOpenProjects might be easily implemented with a getProjects and a switch parameter, but this tends to generate methods which require quite a bit of understanding of the underlying data and/or structure of the object... not necessarily something you want to encourage. By having more explicit and specific functions, you can completely wrap and hide the implementation from the code using it. Isn't that one of the points of OO?
More so than just naming a class, creating an appropriate package structure can be a difficult but rewarding challenge. You need to consider separating the concerns of your modules and how they relate to the vision of the application.
Consider the layout of your app now:
App
VendorDocRequester (read from web service and provide data)
VendorDocViewer (use requester to provide vendor docs)
I would venture to guess that there's a lot going on inside a few classes. If you were to refactor this into a more MVC-ified approach, and allow small classes to handle individual duties, you might end up with something like:
App
VendorDocs
Model
Document (plain object that holds data)
WebServiceConsumer (deal with nitty gritty in web service)
Controller
DatabaseAdapter (handle persistance using ORM or other method)
WebServiceAdapter (utilize Consumer to grab a Document and stick it in database)
View
HelpViewer (use DBAdapter to spit out the documention)
Then your class names rely on the namespace to provide full context. The classes themselves can be inherently related to application without needing to explicitly say so. Class names are simpler and easier to define as a result!
One other very important suggestion: please do yourself a favor and pick up a copy of Head First Design Patterns. It's a fantastic, easy-reading book that will help you organize your application and write better code. Appreciating design patterns will help you to understanding that many of the problems you encounter have already been solved, and you'll be able to incorporate the solutions into your code.
Leo Brodie, in his book "Thinking Forth", wrote that the most difficult task for a programmer was naming things well, and he stated that the most important programming tool is a thesaurus.
Try using the thesaurus at http://thesaurus.reference.com/.
Beyond that, don't use Hungarian Notation EVER, avoid abbreviations, and be consistent.
Best wishes.
In short:
I agree that good names are important, but I don't think you have to find them before implementing at all costs.
Of course its better to have a good name right from the start. But if you can't come up with one in 2 minutes, renaming later will cost less time and is the right choice from a productivity point of view.
Long:
Generally it's often not worth to think too long about a name before implementing. If you implement your class, naming it "Foo" or "Dsnfdkgx", while implementing you see what you should have named it.
Especially with Java+Eclipse, renaming things is no pain at all, as it carefully handles all references in all classes, warns you of name collisions, etc. And as long as the class is not yet in the version control repository, I don't think there's anything wrong with renaming it 5 times.
Basically, it's a question of how you think about refactoring. Personally, I like it, though it annoys my team mates sometimes, as they believe in never touch a running system. And from everything you can refactor, changing names is one of the most harmless things you can do.
Why not HelpDocumentServiceClient kind of a mouthful, or HelpDocumentClient...it doesn't matter it's a vendor the point is it's a client to a webservice that deals with Help documents.
And yes naming is hard.
There is only one sensible name for that class:
HelpRequest
Don't let the implementation details distract you from the meaning.
Invest in a good refactoring tool!
I stick to basics: VerbNoun(arguments). Examples: GetDoc(docID).
There's no need to get fancy. It will be easy to understand a year from now, whether it's you or someone else.
For me I don't care how long a method or class name is as long as its descriptive and in the correct library. Long gone are the days where you should remember where each part of the API resides.
Intelisense exists for all major languages. Therefore when using a 3rd party API I like to use its intelisense for the documentation as opposed to using the 'actual' documentation.
With that in mind I am fine to create a method name such as
StevesPostOnMethodNamesBeingLongOrShort
Long - but so what. Who doesnt use 24inch screens these days!
I have to agree that naming is an art. It gets a little easier if your class is following a certain "desigh pattern" (factory etc).
This is one of the reasons to have a coding standard. Having a standard tends to assist coming up with names when required. It helps free up your mind to use for other more interesting things! (-:
I'd recommend reading the relevant chapter of Steve McConnell's Code Complete (Amazon link) which goes into several rules to assist readability and even maintainability.
HTH
cheers,
Rob
Nope, debugging is the most difficult thing thing for me! :-)
DocumentFetcher? It's hard to say without context.
It can help to act like a mathematician and borrow/invent a lexicon for your domain as you go: settle on short plain words that suggest the concept without spelling it out every time. Too often I see long latinate phrases that get turned into acronyms, making you need a dictionary for the acronyms anyway.
The language you use to describe the problem, is the language you should use for the variables, methods, objects, classes, etc. Loosely, nouns match objects and verbs match methods. If you're missing words to describe the problem, you're also missing a full understanding (specification) of the problem.
If it's just choosing between a set of names, then it should be driven by the conventions you are using to build the system. If you've come to a new spot, uncovered by previous conventions, then it's always worth spending some effort on trying extend them (properly, consistently) to cover this new case.
If in doubt, sleep on it, and pick the first most obvious name, the next morning :-)
If you wake up one day and realize you were wrong, then change it right away.
Paul.
BTW: Document.fetch() is pretty obvious.
I find I have the most trouble in local variables. For example, I want to create an object of type DocGetter. So I know it's a DocGetter. Why do I need to give it another name? I usually end up giving it a name like dg (for DocGetter) or temp or something equally nondescriptive.
Don't forget design patterns (not just the GoF ones) are a good way of providing a common vocabulary and their names should be used whenever one fits the situation. That will even help newcomers that are familiar with the nomenclature to quickly understand the architecture. Is this class you're working on supposed to act like a Proxy, or even a Façade ?
Shouldn't the vendor documentation be the object? I mean, that one is tangible, and not just as some anthropomorphization of a part of your program. So, you might have a VendorDocumentation class with a constructor that fetches the information. I think that if a class name contains a verb, often something has gone wrong.
I definitely feel you. And I feel your pain. Every name I think of just seems rubbish to me. It all seems so generic and I want to eventually learn how to inject a bit of flair and creativity into my names, making them really reflect what they describe.
One suggestion I have is to consult a Thesaurus. Word has a good one, as does Mac OS X. That can really help me get my head out of the clouds and gives me a good starting place as well as some inspiration.
If the name would explain itself to a lay programmer then there's probably no need to change it.