A Brief History Of C# Style

re: A Brief History Of C# Style

Tom — Sun, 25 May 2008 15:19:32 GMT

I am grateful that Microsoft has released this tool, however, I am not grateful that, in it's default configuration, it promulgates overuse of "this." I haven't been able to find a good justification for this, I haven't come across any C# books advocating this style, and I personally believe that adding "this." everywhere clutters the code and is unnecessary.

After all, we all by now know that we are writing classes, and that classes have fields, properties, and methods that are in scope without derefencing an object reference or pointer. By insisting that "this." prefix all these references, it is a step backwards towards C style (pre classes).

re: A Brief History Of C# Style

Johnnylightbulb — Sun, 25 May 2008 18:09:20 GMT

Thanks for the background - I wasn't aware this was an 'on-going battle' within Microsoft. Btw the tool is really awesome and will definitely become part of our process.

re: A Brief History Of C# Style

Jeremy Gray — Sun, 25 May 2008 20:44:53 GMT

Thanks for the brief recap, which confirms what I've often heard out of softies over the years. Sadly, it also confirms the exact moment I stopped paying attention to many MSFT coding guidelines, that being the moment when the "this." craze took a single character discriminator and turned it into a five character source of density-increasing noise that offers zero increased utility over a single underscore. For lack of a better way of phrasing it on this muggy morning: epic fail.

re: A Brief History Of C# Style

W. Kevin Hazzard — Sun, 25 May 2008 22:38:02 GMT

It's interesting that ReSharper discourages superfluous use of the this reference while StyleCop encourages it. The MSIL generated by the C# compiler is the same in both cases so one can't argue that it's much more than a style choice. Personally, I like using the this reference to make it perfectly clear to new readers of the code. Then again, the m_ prefix has the same effect. And it's arguable that the s_ prefix for static members fields used by some C++ shops is even more descriptive. All in all, having a standard and enforcing it is what's important.

re: A Brief History Of C# Style

Claude Baillargeon — Mon, 26 May 2008 03:56:09 GMT

Could we expect a source analysis for VB .Net ?

re: A Brief History Of C# Style

Patrick Smacchia — Mon, 26 May 2008 16:48:14 GMT

> All in all, having a standard and enforcing it is what's important.

... and that is why it is important to be able to write seamlessly custom rules. This 'this.' debate underline the fact that not everybody wish to have the same rule.

With NDepend, enforcing the m_ and s_ prefix is as easy as writing these 2 conventions:

WARN IF Count > 0 IN SELECT FIELDS WHERE !NameLike "^s_" AND IsStatic

WARN IF Count > 0 IN SELECT FIELDS WHERE !NameLike "^m_" AND !IsStatic

re: A Brief History Of C# Style

James — Tue, 27 May 2008 03:43:48 GMT

In my own code, I'm pretty big on using "m_" and "s_". I tried to use "this" for a while, but I would usually forget to type it. Now that I have this tool to play with, I think I'll give it another go.

What is the value of "this"? (or _ or m_ or s_?)

Dave Cameron — Wed, 28 May 2008 06:17:57 GMT

Are "this", "_", "m_" and "s_" only about explicitly repeating the scope of the variable?

I realize it's a preference thing, but I'm curious about why people prefer it.

re: A Brief History Of C# Style

andrei — Wed, 28 May 2008 09:57:30 GMT

"I realize it's a preference thing, but I'm curious about why people prefer it."

It takes about 2 seconds to determine the scope of a variable used in a method without the 'm_'. With it, it's instantaneous.

re: A Brief History Of C# Style

David Regan — Wed, 28 May 2008 12:42:06 GMT

I've used stylecop on projects when working for Microsoft and its value is clear however I personally found it a little too picky. For example, the enforced ordering of constructors, properties, methods in a source file as determined by their accessibility is difficult to get right without considerable thought.

Because the tool just tells you that your source is wrong rather than fixing the problem means that developing code is slowed down and some of the joy of cutting code is taken away.

Consequently, on projects that I've led outside of Microsoft I've preferred the approach of using Resharper amd defining a common project style used by all the projects. The advantages are that reshaper's warning are very well integerated into VStudio and most importantly the Ctrl-Alt-F key stroke rewrites the source according to the style.

The disadvantages are: the style are not as rigourous (the aforementioned ordering) and the build doesn't fail for the warnings.

re: A Brief History Of C# Style

Tom — Wed, 28 May 2008 16:25:58 GMT

I find it ironic that Microsoft, on the one hand discourages the use of prefixes to denote class members (e.g., '_', "m_", etc.) but then advocates "this." as a prefix. How is "this." any different than "m_"?

re: A Brief History Of C# Style

Ricky — Thu, 29 May 2008 00:59:28 GMT

Why StyleCop is suggesting to put using statement under the namespace block?

Is that just a style to write or it has some differences?

If the difference is just to make using statements available for that namespace and class with-in only then I am not convinced to use this style because the another suggestion says that we should have only one class in a file.

So what is the catch here?

Microsoft Source Analysis for C# + ReSharper = Real-time Source Analysis

Howard van Rooijen's Blog — Thu, 29 May 2008 15:57:34 GMT

I've worked on a couple of projects for Microsoft UK - one of the many great aspects of those projects

Interesting Finds: 2008.05.31

gOODiDEA.NET — Sat, 31 May 2008 09:25:19 GMT

.NET A Brief History Of C# Style Writing your own rules for Microsoft Source Analysis for C# Checking

re: A Brief History Of C# Style

Andrew — Sat, 31 May 2008 17:32:10 GMT

I tend to use the underscore (_name) to indicate that its a backing variable for a property, while a private or local variable would just start with lower case (name).

Its meant to communicate the fact that you should not use _name to except through the property.

Source Analysis for T-SQL perchance?

SSIS Junkie — Tue, 03 Jun 2008 00:54:39 GMT

My learned colleague Howard van Rooijen recently blogged about the new release of Microsoft Source Analysis

re: A Brief History Of C# Style

RichB — Fri, 06 Jun 2008 11:14:30 GMT

> the CLR team began writing the first version

> of the .Net Framework. Most of this code was

> written in C#. This was the first production

> C# code written anywhere in the world

Peter Hallam (one of the original designers of C#) says that WinForms was the original consumer of the C# language, not the BCL. See:

http://wm.microsoft.com/ms/msdn/visualcsharp/peter_hallam_2006_11/PeterHallam01.wmv

Therefore, the .Net BCL was either written after WinForms (unlikely as WinForms was ported from WFC December 1999), or was written in the C-dialect SMC and later ported to C# much like ASP.Net was originally SMC.

On an unrelated note, could you give us an idea of which teams within Microsoft are using StyleCop? My understanding is that it is only the MCS teams.

re: A Brief History Of C# Style

Chuck Snyder — Fri, 06 Jun 2008 18:05:55 GMT

I'll ask too, any chance for this to come out for VB????

re: A Brief History Of C# Style

John Saunders — Tue, 10 Jun 2008 14:57:43 GMT

As several comments here have indicated, style is a matter of preference. While I agree that a given Development organization needs to have a single set of style preferences enforced, I'm afraid I can't agree on having the Microsoft style preferences enforced on my organization.

The lack of configurability makes this tool a non-starter for us.

John

re: A Brief History Of C# Style

chocho — Fri, 20 Jun 2008 11:52:10 GMT

Since it is agreed that a prefix to distinguish scope is needed why not save yourself typing and reading and use the shortest?

"this." is too much trouble. "m_" underscode will lead to carpal syndrome. just use "m"! and I'm not a C++ programmer :)

re: A Brief History Of C# Style

Kishore — Tue, 01 Jul 2008 16:39:43 GMT

Where can I find StyleCop rules? If I had one, I can give it to my developers so that they can code accordingly. I know there will be slippages and there comes StyleCop handy.

Thanks for the wonderful tool.

Kishore

re: Why use "m_"

Bryan — Tue, 01 Jul 2008 18:58:10 GMT

> Are "this", "_", "m_" and "s_" only about explicitly repeating the scope of the variable?

Because the compiler won't give any messages on FuncA:

class A

{

private int a = 0;

public A()

{

a = 3;

}

public int Value

{

get { return a; }

set { a = value; }

}

public int Func(int a)

{

return a*a;

}

re: A Brief History Of C# Style

santilajara — Wed, 02 Jul 2008 13:58:22 GMT

It will be good to know to which extend this tool in its style rules has official support in Microsoft, and if will see them widely adopted in the source code shown by Microsoft in the future. If this is not the case, I think maybe it is not worth investing to adopt rules that break what we use to see in the code found in internet inside and outside Microsoft.

re: A Brief History Of C# Style

alwin — Sun, 06 Jul 2008 00:43:53 GMT

"I tend to use the underscore (_name) to indicate that its a backing variable for a property, while a private or local variable would just start with lower case (name).

Its meant to communicate the fact that you should not use _name to except through the property."

Haha i do exactly the same! And here i was thinking i had an original idea...

I used to do this. all the time, just for the intellisense. But since i got resharper i don't bother anymore.

re: A Brief History Of C# Style

Dennis C — Fri, 11 Jul 2008 14:34:41 GMT

This tool is SO cool.. but most of our projects are web projects. Any ideas on making it work for those? Or suggestions for an alternate tool....??

re: A Brief History Of C# Style

jasonall — Fri, 11 Jul 2008 16:04:24 GMT

Hi Dennis C. The tool should already support web projects. There were a couple of problems around this which will be fixed in 4.3. Perhaps you are hitting that?

re: A Brief History Of C# Style

cdr — Fri, 11 Jul 2008 19:15:55 GMT

When can we expect better documentation of the rules? The lack of documentation is the #1 problem with the project right now for me.

C# style conventions: more religious wars coming!

LA.NET [EN] — Wed, 16 Jul 2008 01:24:50 GMT

Yes, it's one of those posts where you either agree or disagree completely :) Anyway, after having

C# style conventions: more religious wars coming!

ASPInsiders — Wed, 16 Jul 2008 02:13:08 GMT

Yes, it's one of those posts where you either agree or disagree completely :) Anyway, after having

re: A Brief History Of C# Style

Christian Gross — Sat, 19 Jul 2008 22:59:12 GMT

Ah yes, lets start the debate of where to put the curly brackets and where to put spaces...

Lo and behold, you know I could use a tool like resharper and automatically reformat the code to how I like to see it... WOW...

I use resharper and have to say its rules are actually very very good. They are both logical and not bothersome. They worry about things that need to be worried about. Maybe Microsoft should learn from Reshaper a bit?

re: A Brief History Of C# Style

martinmarv — Wed, 06 Aug 2008 18:17:54 GMT

It sounds like this could be a useful tool, but more useful would be a set of industry-wide guidelines on coding-style. The difficulty is how to decide who wins on debates like m_, s_ and "this.".

There could take the form of an open forum, with discussions (like the one above) on contentious issues, and voting for the top options, with everyone agreeing to go for the democratically chosen choice, or at least having the democratic choice being a named-standard (e.g. The StyleCop Way). People can learn it, and either follow it, or know where they are differing.

Just my 2 cents!

re: A Brief History Of C# Style

Kit — Thu, 18 Sep 2008 20:01:47 GMT

Why is each developer looking at the same text? Something that lets me see it my way, and lets you see it your way is what we really need. Semantics, of course, should be preserved, and these should be subject to more important static analysis tools.

re: A Brief History Of C# Style

Jesper — Sun, 28 Sep 2008 01:46:49 GMT

For at least 30 years a leading underscore has been discouraged in user code, for the simple reason that it was reserved for use by the C language implementors. I shake my head every time I see user code (in any language) where variables or (God forbid!) function names begin with an underscore. I guess you children are too young to have read Kernighan & Richie...

One rule I've tried to follow is that identifiers should be readable in "plain English". There are no words that I know of in the English language that begin with "m_" - this is a "wort", on par with using Hungarian notation in object oriented code.

The simplest solution (to me) is to use a trailing underscore on private identifiers. However, I have to admit that I also don't know of any words in the English language that ends in an underscore, but I can live with that...

The only style I generally follow is consistent - and exactly the style I prefer and use, including many "standard" exceptions to the rules!

re: A Brief History Of C# Style

jasonall — Tue, 14 Oct 2008 14:40:16 GMT

Please use the StyleCop Discussion Forum for further comments as it provides a better mechanism for tracking topics and follow-ups.