Reply to: C# 4.0

[threaded]

Did it invoke Godwins law?

[DimPrawn]

Amazing I post a silly throw-away post about C# and it develops into an all out flame war.

Geeks are great aren't they? Call their mothers a whore and they don't bat an eyelid, but diss Java or C++, and feck, they are ready to punch your lights out.

[bogeyman]

I despair for the clients of the previous two posters.

Yes - me too. I despair all the way to the bank.

[threaded]

That was what I loved when OO became "hot": do it our way and your code will be reusable. Step 1: throw away all your code.

Yeah, that attitude shows a true understanding of the power that lies within OO. I despair for the clients of the previous two posters.

[expat]

That's what's so great about it: one can go on site and produce all sorts of evidence that the previous design was cack and has to be re-written to make it reusable...

That was what I loved when OO became "hot": do it our way and your code will be reusable. Step 1: throw away all your code.

[threaded]

Suggest we have a separate forum dedicated to this topic.

No need, me and AtW are quite happy to share this one thank you.

[bogeyman]

Threaded, you really are a cock.

Suggest we have a separate forum dedicated to this topic.

[Churchill]

That's what's so great about it: one can go on site and produce all sorts of evidence that the previous design was cack and has to be re-written to make it reusable...

Threaded, you really are a cock.

[threaded]

Object orientation and reusability are in the mind of the beholder.

That's what's so great about it: one can go on site and produce all sorts of evidence that the previous design was cack and has to be re-written to make it reusable...

[VectraMan]

If you want ASM, use ILASM and you can write code for the CLR. I've written a workflow to native compiler for .Net before believe it or not. Quite simple!

Always wondered if you could do that.

Pointers are NOT needed - the CLR dynamically assembles the intermediate language to x86 / x86-64 machine code in the most efficient way based on how it is used. Sometimes pointers make things slower (!).

Sounds like you have a little too much faith in the optimiser. A custom solution will always be better than a generic one if you're going for out an out performance.

I was looking into C++ .NET, and it seems they've now fudged the language to use ^ as a managed pointer type.

[expat]

Whenever anyone asks my thoughts on C++ I just answer that that Barney Sodastream guy is a Dane from Aarhus. It is enough of an answer I feel.

C is a programming language, one that "wears well" as K&R put it. C++ adds things that it is really the programmer's job to add. Object orientation and reusability are in the mind of the beholder.

Humbug!

[threaded]

<Fr Dougal>

And why would they do that then Ted?

Is that like when you're standing at a bus stop, or washing your Porche, or something?

Or do they do it to see what a big feckin' eejot you are?

</Fr Dougal>

It's 'cause I'm on some committee to do wi' one of these OO language things, innit.

[mcquiggd]

Well, jobsOverseas is not declared and Agenda() will never run as its private and no other functions in class use it. The funniest thing however is that loop is going to run much longer than you probably think...

Oh dear AtW... please chill out.... read it as a light hearted comment, if id been serious I would have written it in VS.Net then posted....

[AtW]

In the end it doesn't matter - it may end up being converted to pointers when it's assembled by the JIT.

It does matter - direct access to memory allows to avoid range checks on arrays, also it allows to manipulate bytes/words/dwords directly rather than use BitConverter calls that are not cheap.

[AtW]

There are pointers in C# (!). See language reference.

Well, I can't really see where I said that C# has not got pointers??!

As i said usage of pointers can double performance - direct access to memory bypasses range checking, which is fairly costly. Naturally their usage is only worth at real hotspots in the code.