Its adherents rarely surprise me anymore.

Read my post on LTU and snk_kid's response:
http://lambda-the-ultimate.org/node/1583#comment-19266

I am not trying to single him out, but the responses are pretty common in my (arguably anectdotal) experience.

C++ is getting X in next version...
Using boost/somelib we can do X...
GC/Feature N will be great for C++ but for language Y its slow...

Sigh.

Its funny when people bring up libraries to make their points.
Templates in C++ are slow to compile and bloated in my exposure to them.
And even with all those libraries can you finally type:
string s = "Hello" + " World";

After using D for a while now, i almost _despise_ C++.
I think I would rather just use (python & C) instead of going back.

Walter you have your work cut out for you.
Like most things in IT, the human/political issues outweigh the technical ones.

Sorry for the rant.
-DavidM

David Medlock wrote:
> Its adherents rarely surprise me anymore.
> 
> Read my post on LTU and snk_kid's response:
> http://lambda-the-ultimate.org/node/1583#comment-19266
> 
> I am not trying to single him out, but the responses are pretty common in my (arguably anectdotal) experience.
> 
> C++ is getting X in next version...
> Using boost/somelib we can do X...
> GC/Feature N will be great for C++ but for language Y its slow...
> 
> Sigh.
> 
> Its funny when people bring up libraries to make their points.

Not for C++.  For better or worse, a primary design goal of C++ was keep as much as possible out of the language definition by supporting user defined types that are indistinguishable from built-in types.  So a valid comparison to C++ must necessarily include what can be done in library code.

That said, I don't agree with many of snk_kid's observations.  D *does* support implicit template instantiation, for example.  The compiler implementation is just incomplete.  As for concepts, I do think they would be useful in D but I think they're less necessary than in C++ because D has static if and static assert.  Also, you can inherit from template type parameters as far as I know.

As for boost::function--it's a horrible bit of code and is far more awkward to use than delegates (particularly anonymous delegates) but it works pretty well aside from all that.  However, snk_kid doesn't address the idea that inner functions (ie. delegates) have access to local stack data, or that actually defining a new boost::function is nothing at all like defining a plain old function--it's obviously a class-based hack.

Meta logic is much the same... it's possible in C++ but is far more awkward.  And C++ obviously doesn't support strings or decimal numbers as template value parameters.  Not to mention aliases.

I think one problem is that some people just look at the spec to learn about the language, and the spec is quite sparse in places.  I don't think it mentions implicit template instantiation, and mention of some other features is hard to find.

> Walter you have your work cut out for you.
> Like most things in IT, the human/political issues outweigh the technical ones.

Always.  But as long as D remains in development and isn't supplanted by some magic new language with all of D's features *and* a My Little Pony (tm) then it will have users.  Particularly with GDC available for the non-Windows folks.


Sean

"David Medlock" <noone@nowhere.com> wrote in message news:e819r5$h8i$1@digitaldaemon.com...

> Read my post on LTU and snk_kid's response:

snk_kid is just another C++ fanboy.  I'm dealing with him on gamedev.net too.

Jarrett Billingsley wrote:
> "David Medlock" <noone@nowhere.com> wrote in message news:e819r5$h8i$1@digitaldaemon.com...
> 
>> Read my post on LTU and snk_kid's response:
> 
> snk_kid is just another C++ fanboy.  I'm dealing with him on gamedev.net too. 
> 
> 

link?

Sean Kelly wrote:
> David Medlock wrote:
>> Its adherents rarely surprise me anymore.
>>
>> Read my post on LTU and snk_kid's response:
>> http://lambda-the-ultimate.org/node/1583#comment-19266
>>
>> I am not trying to single him out, but the responses are pretty common in my (arguably anectdotal) experience.

> I think one problem is that some people just look at the spec to learn about the language, and the spec is quite sparse in places.  I don't think it mentions implicit template instantiation, and mention of some other features is hard to find.

On the "Comparison with other languages" page, it explicitly says that D doesn't have implicit function template instantiation. AND it's been removed from the 'future directions' page.
When I first looked at D, the absence of IFTI was the #1 turn-off. It's only because I saw it in the 'future directions' page that I gave the language a chance.

>> Walter you have your work cut out for you.
>> Like most things in IT, the human/political issues outweigh the technical ones.

True, but I'm going to stick up for snk_kid. Based on the spec, he's quite justified in thinking D doesn't have IFTI and probably never will; and the implications for library design are huge. He's justified in thinking that D template libraries will never be as powerful as C++.
When you read the spec, you should get the impression that D templates are far more powerful than C++ templates.

Sean Kelly wrote:
> For better or worse, a primary design goal of C++ was keep as much as possible out of the language definition by supporting user defined types that are indistinguishable from built-in types.  So a valid comparison to C++ must necessarily include what can be done in library code.

For a language with minimal features as a primary design goal, C++ turned out to be, by far, the most complex core language ever to implement!

clayasaurus wrote:
> Jarrett Billingsley wrote:
>> "David Medlock" <noone@nowhere.com> wrote in message news:e819r5$h8i$1@digitaldaemon.com...
>>
>>> Read my post on LTU and snk_kid's response:
>>
>> snk_kid is just another C++ fanboy.  I'm dealing with him on gamedev.net too.
>>
> 
> link?

Shame on you, Clay, for not keeping up with my blog!

http://www.gamedev.net/community/forums/topic.asp?topic_id=400729

Don Clugston wrote:
> Sean Kelly wrote:
> 
 >
> True, but I'm going to stick up for snk_kid. Based on the spec, he's quite justified in thinking D doesn't have IFTI and probably never will; and the implications for library design are huge. He's justified in thinking that D template libraries will never be as powerful as C++.
> When you read the spec, you should get the impression that D templates are far more powerful than C++ templates.


I would agree with you, but look at the first sentence of my post(above his).

Mike Parker wrote:
> clayasaurus wrote:
> 
>> Jarrett Billingsley wrote:
>>
>>> "David Medlock" <noone@nowhere.com> wrote in message news:e819r5$h8i$1@digitaldaemon.com...
>>>
>>>> Read my post on LTU and snk_kid's response:
>>>
>>>
>>> snk_kid is just another C++ fanboy.  I'm dealing with him on gamedev.net too.
>>>
>>
>> link?
> 
> 
> Shame on you, Clay, for not keeping up with my blog!
> 
> http://www.gamedev.net/community/forums/topic.asp?topic_id=400729

Interesting, in both places snk_kid keeps harping on his claim that "D is a failed attempt to be truely better than C++ in *all aspects* which it isn't."

Where did he get that impression????

I have never seen anyone else make the claim that D is trying to be better than anything in *all* respects. Walter doesn't claim this. And considering it is provable that it can't be done, It would be stupid for anyone to claim that.

Also, in places he argues that D will not take over because it can't do something that C++ programmers do. This is only a problem if you try to wright C++ programs in D. The question shouldn't be: "Can the language generate this construct?", it should be: "How easy is it to solve this problem in this language?".

D isn't a better C++. No programming language can be a better C++. D is trying to be a better *Programming Language*.

Don Clugston wrote:
> On the "Comparison with other languages" page, it explicitly says that D doesn't have implicit function template instantiation.

Fixed (How did I miss that!)