What are some authoritative/respected "best known implementation" websites/resources?_问答_开发者

EDIT:

Wow, the initial response to this question was quite negative. I think I might have triggered some pretty strong emotions by using the word "bes开发者_StackOverflow中文版t"; it seems like a few people latched onto that word and decided to dismiss my question right away.

Obviously, there are many, many situations in which no single approach is "best", or at least, what ends up being the best solution to one problem will often not be the best solution for other, even similar, problems. I get that. But now let me try to elaborate on the reasoning behind what I'm actually asking.

I tend to find it easiest to explain myself using analogies, so here goes. In my current job I work almost exclusively in .NET. .NET has a lot of functionality built into the framework. A prime example is the System.Collections.Generic namespace, which has a bunch of collection classes that (almost) no .NET developer in his/her right mind would bother re-developing from scratch, because very good implementations are already there. If I am working on a problem that requires a doubly linked list, I'm not going to decide, "Okay, time to write a doubly linked list class"; I'm just going to use the LinkedList<T> that's already there, or, at most, extend it or wrap it with my own class that adds some extra functionality.

Am I saying the "best" version of a doubly linked list is LinkedList<T> from .NET? Of course not. That would be absurd. But I highly doubt .NET's implementation of LinkedList<T> is drastically different from most other established libraries' implementations of collections that are intended to serve the same purpose (that of a doubly linked list). On the other hand, I am relatively confident that if I were to write my own implementation from scratch, there'd be a considerable number of issues with it, in terms of robustness, performance, flexibility, etc. for one simple reason: not that I'm stupid, or lazy, or don't care about good code--simply that I'm one person, and I'm not an expert on linked lists, and I haven't thought of everything that needs to be taken into consideration when designing one.

But I happen to be a developer who does take an interest in how things are implemented internally. And so it would be nice if I could check out a page where some variant of a well thought-out design for a linked list--or for any fairly established concept for which robust, efficient implementations have been written--were available to view. (By the way, yes I am aware that the source code for .NET's LinkedList<T> is available. I'm just using that as an example; really I am talking about all problems with solutions for which good, working implementations exist.)

Now, I talked about this being something that is open; let me elaborate on that. I am not talking about sites like SourceForge.net, or CodePlex, or Google Code. These are all sites for hosting projects, i.e., applications or libraries tailored for some specific industry or field or otherwise categorizable purpose. What I'm talking about is something like this:

http://en.wikibooks.org/wiki/Category:Algorithms_and_data_structures

Maybe I should have just provided that link in the first place, as it probably illustrates what I'm getting at better than anything I've written so far. But I think the main point that differentiates what I'm asking about from any other site I've seen is that I was specifically wondering if there could be some way to work on a new problem--so, something for which there aren't necessarily any well-known, established implementations, again as in my linked list example--collaboratively, in a wiki-esque fashion, but not tied to any specific open-source project.

So, as a conclusion of sorts, I was kind of envisioning a situation like the following: I find myself faced with a new problem. Maybe it isn't common enough to be something that is addressed in a framework like .NET. But it's common enough that some developers here and there are independently working on it. If a website exists like what I'm imagining, maybe at some point one of those developers working on the problem could post an idea on that website, and over time others might discover it and suggest improvements/modifications, and given enough time and participation, a pretty darn good implementation might result from all this collaboration. And from there, eventually, something like this implementation might be considered fairly "standard", just like a linked list implementation, or a quicksort implementation, or, I don't know, some well-known pseudo-random number generator.

Does this make any more sense to anyone now? I feel quite confident that what I'm talking about is not absurd, but hey, if that's what people think, then maybe it is.

Open source projects are very popular. Some of these are libraries suited for specific purposes, the best of which include some very well-written code.

However, if you're interested in contributing to an open source project, finding a project that is well-suited to your skills can be quite a task. At the same time, if you're interesting in using an open source project in your own work, finding a project that is well-suited to your needs can also be difficult, especially when, for example, open-source library X has a lot of functionality you could use, as does library Y, and these two libraries' capabilities overlap so that integrating both into your code could be messy.

We've all seen questions, here on Stack Overflow and elsewhere on the web, posted by one developer: "How would I implement this idea?" and answered by others, often accompanied by a plethora of example code. Sometimes these answers link to an open source project/library that provides functionality similar to what the poster is asking about.

My question is: are there any well-known websites or other sources that are open in nature and provide "best-known implementations" for common (or even not-so-common) programming problems, but not associated with any particular open source project?

As a generic example, suppose I have a need for some algorithm that does X. I post a question on SO or some other site requesting ideas, asking for suggestions on how best to implement it. One person points me to project P1, which contains some code that performs something very similar to this algorithm. Another person points me to project P2. Someone else writes some sample code and says, "maybe you could do it like this."

It seems to me, if there are all these different versions of this idea floating around out in the world, it would make sense for there to be a site, somewhat in the vein of Wikipedia, where a quasi-"official" implementation ("official" is not the right word; I'm just having trouble thinking of a better one right now) could be published and modified as improvements are developed/discovered.

I feel like I have stumbled across a few different sites like this in the past, but I'm interested to know if anyone else has found any resources like what I'm describing.

The very idea is absurd. It means that there's one, single opinion on "best-known implementations" with no changes based on other people having better ideas.

It implies a that best practices are static and can be accumulated into a single repository.

If they could be collected, then Google would have them and would simply charge for access.

Interestingly, they don't have all the best practices. Interestingly, they have to expend mountains of computing power looking for more information. Then people (like you) have to read and think and judge and decide.

The read-think-judge-decide is really hard to eliminate. Unless, of course, you want someone to think for you. In which case, there are many companies who have a single solution that requires less thinking. Call Microsoft or Oracle or IBM. They have solutions that are all in one place, unified best practices, no reading, no thinking, no judging, no deciding required.

Open -- by definition -- means it's impossible to have a single authoritative source.

Here is something, maybe not the best implementations. But a book called Design Patterns contains what is considered by many programmers some of the best patterns to follow!