www.digitalmars.com         C & C++   DMDScript  

digitalmars.D.announce - Re: Interview with InformIT part 2/3

reply bearophile <bearophileHUGS lycos.com> writes:
Walter Bright:

 The other problem with a pinned/notpinned object is the object itself cannot 
 control who or how someone is pointing to it.



The type system may tell apart three kinds of pointers/references:
1) hand-managed pointers, to GC memory or C heap memory;
2) GC-managed pointers to pinned memory;
3) GC-managed pointers to unpinned memory.

But this is a long story, I have already discussed this topic a bit, there are
problems with pointers on the stack:
http://www.digitalmars.com/webnews/newsgroups.php?art_group=digitalmars.D&article_id=108544

I am not yet able to design a thing so complex alone, so sorry for the noise
:-) I need to learn more and improve, first.

Bye,
bearophile
Aug 18 2010
parent reply Walter Bright <newshound2 digitalmars.com> writes:
bearophile wrote:

 Walter Bright:

 The other problem with a pinned/notpinned object is the object itself
 cannot control who or how someone is pointing to it.



 The type system may tell apart three kinds of pointers/references: 1)
 hand-managed pointers, to GC memory or C heap memory; 2) GC-managed pointers
 to pinned memory; 3) GC-managed pointers to unpinned memory.
 
 But this is a long story, I have already discussed this topic a bit, there
 are problems with pointers on the stack: 
 http://www.digitalmars.com/webnews/newsgroups.php?art_group=digitalmars.D&article_id=108544
 
 
 I am not yet able to design a thing so complex alone, so sorry for the noise
 :-) I need to learn more and improve, first.



Microsoft's managed C++ on .net comes with multiple pointer types - managed and 
unmanaged pointers - as far as I know, this was a technical success yet a 
massive failure with users.

I have plenty of experience with multiple pointer types coming from my 16 bit 
compiler work. I prefer to run screaming from that path.
Aug 18 2010
parent reply bearophile <bearophileHUGS lycos.com> writes:
Walter Bright:

 Microsoft's managed C++ on .net comes with multiple pointer types - managed
and 
 unmanaged pointers - as far as I know, this was a technical success yet a 
 massive failure with users.



How do you define failure? Maybe for D2 multiple pointer types are a failure as
you say, but in my opinion "managed C++" is not a language, it's not designed
to write complete programs, it's designed to build bridges between C# (dotnet)
and C++ (and C, etc). I know people that use managed C++ professionally, no one
of them likes to use it, but it seems they will keep using it. So I don't think
managed C++ is a failure.

Bye,
bearophile
Aug 18 2010
parent reply Walter Bright <newshound2 digitalmars.com> writes:
bearophile wrote:

 Walter Bright:

 Microsoft's managed C++ on .net comes with multiple pointer types - managed
 and unmanaged pointers - as far as I know, this was a technical success yet
 a massive failure with users.



 How do you define failure?



Nobody wanted to use it.



 Maybe for D2 multiple pointer types are a failure
 as you say, but in my opinion "managed C++" is not a language,



It meets every definition of one.



 it's not designed to write complete programs,



Yes, it is. It was intended to be a big deal. It fell way short of that with
users.



 it's designed to build bridges between
 C# (dotnet) and C++ (and C, etc). I know people that use managed C++
 professionally, no one of them likes to use it, but it seems they will keep
 using it. So I don't think managed C++ is a failure.



Please revisit the "no one of them likes to use it".

Being forced to use something doesn't make that thing a success.
Aug 18 2010
parent "Nick Sabalausky" <a a.a> writes:
"Walter Bright" <newshound2 digitalmars.com> wrote in message 
news:i4hvjh$91i$1 digitalmars.com...

 Being forced to use something doesn't make that thing a success.



Unfortunately, I can think of a lot of counterexamples (any monopoly or 
oligopoly, for instance). But I agree in spirit :)
Aug 18 2010