[AccessD] OT: Why did Mac switch to Intel revisted

Josh McFarlane darsant at gmail.com
Fri Aug 19 13:40:34 CDT 2005


On 8/19/05, Shamil Salakhetdinov <shamil at users.mns.ru> wrote:
> Josh,
> 
> Do you mean VS.NET 2005 Beta 2 when you talk about Edit & Continue feature?
> You guess they will never be able to make it stable? (they managed to do
> that with VB6 and even in MS Access 2003 (although both VB6 and MS Access
> 2003 still may crash unexpectedly in complicated cases) - they will probably
> make it working stable somewhere in not that far future? No any hopes you
> think?

Well, I'm using it mainly in reference to VC++, and having tried it on
VS.NET 2002, 2003, and now 2005, they all have inherient crash issues
even when doing simple things (for example, the crash I described can
even happen when you just begin to type on a line while the program is
at a breakpoint.
 
> By using reflection I mean injecting specific attributes BEFORE Just-in-time
> compilation, just as one of thr possible ways to do that, in fact it will
> not be needed probably because attributes will be inline with the source
> code...

Ah, I misunderstood you there.

> Did you try any of obfuscators or other methods to protect compiled code
> from reverse engineering? - I did - and I found they are good enough even
> the cheap ones, and expensive ones are very good.

I've tried a few different ones, and while some parts of the code
would be obfuscatied, core components such as algorithms were still
semi-reproducable in some test cases. Granted, this was a little bit
ago and they may have made improvements to the obfuscators now.

> I have heard they made good progress with ATL in VS.NET 2005. But I can't
> judge here because I do use ATL a few.
> I didn't know that much about prioritizing threads to logical processors and
> HyperThreading.

Basically the way the multicore processors work at this moment (at
least to the current library of MS code) is that they are just treated
as hyper-threading processors. If you search the MSDN docs you should
be able to find easy ways to build this implementation in.


> "Good software design policies" if I got that correctly is a very broad term
> Do you mean they can restrict also the quantity of attributes and methods of
> a class? If yes - could you please post an example of such software design
> ploicy here or give a reference to it, which also shows that it worked in a
> real life project?

Well, the classic example I like to use is the garbage collector.
Normally in C++ you handle allocation / deallocation of objects on the
heap yourself. In managed C++ with .NET you can use the garbage
collector so you don't have to worry about memory leaks from heap
allocated objects, restricting you from a common easy error, but at
the cost of speed and efficiency.

I wasn't trying to promote the .NET as a restrictive language. I was
trying to state that MS and others can make framework seem more
flexible and easier to maintain, but that the same effect can also be
achieved by good coding practices and software design policy.

In the example of the garbage collector: It is unnecessary to use the
collector if you manage your resources of your classes and code
properly so that you do not leak any resources.

Microsoft seems to be taking the approach towards creating an
easier-to-use interface (.NET) and it seems that the more default
customizable interfaces are being left in the dust because of it.

-- 
Josh McFarlane
"Peace cannot be kept by force. It can only be achieved by understanding."
-Albert Einstein



More information about the AccessD mailing list