D looks like a nifty language.

I maybe would have left out a few things: templates (doesn't add much value), and certainly pointer arithmetic. I'd rather do something like java's jni. So that it guarantees that D sources don't show up with pointer-based, complicated looking stuff and scare people off.

I'm not going to go any deeper into this at this point. The fact that the compiler is no free software, means that no one can improve the compiler; especially in its platform-specific optimalizations. I mean, nobody knows *everything* about *all* platforms. What about support for StrongArm? I'm sure people would port the thing, if they were free to do so. Conclusion: it's too much a one-company, or even one-man show to my taste. Without the source, it's also a risky proposition to make yourself dependent on something that may not be maintained any longer, at some point in the future.

On Wed, 21 Apr 2004 02:44:27 +0000, erik wrote:

> D looks like a nifty language.
> 
> I maybe would have left out a few things: templates (doesn't add much value), and certainly pointer arithmetic. I'd rather do something like java's jni. So that it guarantees that D sources don't show up with pointer-based, complicated looking stuff and scare people off.
> 
> I'm not going to go any deeper into this at this point. The fact that the compiler is no free software, means that no one can improve the compiler; especially in its platform-specific optimalizations. I mean, nobody knows *everything* about *all* platforms. What about support for StrongArm? I'm sure people would port the thing, if they were free to do so. Conclusion: it's too much a one-company, or even one-man show to my taste. Without the source, it's also a risky proposition to make yourself dependent on something that may not be maintained any longer, at some point in the future.

There is a D front end for GCC. It's just not yet in the main GCC release, as it's only a few months (weeks?) old.

Mike Swieton
__
Flight by machines heavier than air is unpractical and insignificant, if not
utterly impossible.
	- Simon Newcomb; 1902 - 18 months before Kitty Hawk

"erik" <erik_member@pathlink.com> wrote in message news:c64n6b$1qmt$1@digitaldaemon.com...
> D looks like a nifty language.
>
> I maybe would have left out a few things: templates (doesn't add much
value),
> and certainly pointer arithmetic. I'd rather do something like java's jni.
So
> that it guarantees that D sources don't show up with pointer-based,
complicated
> looking stuff and scare people off.
>
> I'm not going to go any deeper into this at this point. The fact that the compiler is no free software, means that no one can improve the compiler; especially in its platform-specific optimalizations. I mean, nobody knows *everything* about *all* platforms. What about support for StrongArm? I'm
sure
> people would port the thing, if they were free to do so. Conclusion: it's
too
> much a one-company, or even one-man show to my taste. Without the source,
it's
> also a risky proposition to make yourself dependent on something that may
not be
> maintained any longer, at some point in the future.
>

D is a language... not a compiler: Which, as far as I know, is free for use by anyone. DMD the compiler is only one possible implementation of this language, so any engineer out there willing to write his own compiler is free to do so. I'm sure Microsoft, Borland, and MetroWorks doesn't hand you the source to their compiler. They implement and you use. Don't know but I don't think Bjarne Stroustrup gave up the C++ compiler either, merely the Language and interested parties created their own compilers. I also might be wrong but I don't think Walter had the source code when he sat down to implement Zortech C++.

Point is, the Language is free, so implement a free compiler if you doubt the future of this one as much as you claim to.

Andrew

erik wrote:
> D looks like a nifty language. 
> 
> I maybe would have left out a few things: templates (doesn't add much value),
> and certainly pointer arithmetic. I'd rather do something like java's jni. So
> that it guarantees that D sources don't show up with pointer-based, complicated
> looking stuff and scare people off.

No value in templates?

template index(T:T[]) {
    int index(T target, T[] arr) {
        foreach (int idx, T item; arr)
            if (item == target)
                return idx;
        return -1;
    }
}

Hmm.. now I can find the index of a value in absolutely any non-object array, without having to write a seperate function for each array type or play around with void*/void[]...  And yes I could modify/overload the template to make it object-compatable...  This is just one tiny little example of exactly why templates are indeed valuable.

Do away with pointer math?

Wouldn't be good for our C compatability.  And there are some generics and other things which are accomplished nicely using pointers.  Plus you can avoid them entirely if you want.  Instead of pointer arguments to functions, use inout, for intance.

> I'm not going to go any deeper into this at this point. The fact that the
> compiler is no free software, means that no one can improve the compiler;
> especially in its platform-specific optimalizations. I mean, nobody knows
> *everything* about *all* platforms. What about support for StrongArm? I'm sure
> people would port the thing, if they were free to do so. Conclusion: it's too
> much a one-company, or even one-man show to my taste. Without the source, it's
> also a risky proposition to make yourself dependent on something that may not be
> maintained any longer, at some point in the future.

Its free for any non-commercial use (which is a good thing, since I'm using the heck out of it).  And the front-end is open-sourced, and as someone already said there is a GCC port of it ("GDC") which has proven itself nicely -- although I haven't yet personally had the pleasure. Walter started this train moving, that's why he has the position he has, but nothing's to stop anyone from walking off and writing their own implementation... in fact, I bet he'd love it.

-C. Sauls
-Invironz

Let's have a vote here! This particular trout is:

a) trolling
b) just found BASIC and figures goto is the height of cool
c) self-consumed grad-student who ... oh never mind.

"There, but for the grace of God, goes God"; Winston Churchill


"erik" <erik_member@pathlink.com> wrote in message news:c64n6b$1qmt$1@digitaldaemon.com...
> D looks like a nifty language.
>
> I maybe would have left out a few things: templates (doesn't add much
value),
> and certainly pointer arithmetic. I'd rather do something like java's jni.
So
> that it guarantees that D sources don't show up with pointer-based,
complicated
> looking stuff and scare people off.
>
> I'm not going to go any deeper into this at this point. The fact that the compiler is no free software, means that no one can improve the compiler; especially in its platform-specific optimalizations. I mean, nobody knows *everything* about *all* platforms. What about support for StrongArm? I'm
sure
> people would port the thing, if they were free to do so. Conclusion: it's
too
> much a one-company, or even one-man show to my taste. Without the source,
it's
> also a risky proposition to make yourself dependent on something that may
not be
> maintained any longer, at some point in the future.
>
>

> D looks like a nifty language.

It's getting there.

> I maybe would have left out a few things: templates (doesn't add much
value),

Bizarre point of view.

> and certainly pointer arithmetic. I'd rather do something like java's jni.
So
> that it guarantees that D sources don't show up with pointer-based,
complicated
> looking stuff and scare people off.

As something of an expert in mapping C/C++ to other languages, I can say that JNI stinks. But, then, so does Java ...

;/

> I'm not going to go any deeper into this at this point. The fact that the compiler is no free software, means that no one can improve the compiler; especially in its platform-specific optimalizations. I mean, nobody knows *everything* about *all* platforms. What about support for StrongArm? I'm
sure
> people would port the thing, if they were free to do so. Conclusion: it's
too
> much a one-company, or even one-man show to my taste. Without the source,
it's
> also a risky proposition to make yourself dependent on something that may
not be
> maintained any longer, at some point in the future.

The front end is open-source. Currently the only Win32 back-end is that of DMC++, but already there exist back-ends on Linux. So, basically, your point is moot.

erik wrote:
> D looks like a nifty language. 
> 
> I maybe would have left out a few things: templates (doesn't add much value),

The usefulness of templates becomes more apparent as you get used to thinking with them.

> and certainly pointer arithmetic. I'd rather do something like java's jni. So
> that it guarantees that D sources don't show up with pointer-based, complicated
> looking stuff and scare people off.

I think that D does the Right Thing by offering every reason not to need things like goto and pointer arithmetic, but not being afraid to offer them.  I like to think that I'm smart enough to know when the Wrong Way is the Right Way. :)

> I'm not going to go any deeper into this at this point. The fact that the
> compiler is no free software, means that no one can improve the compiler;

As others have said, the source to the frontend is included in the distribution.  David Friedman has already had notable success connecting it to GCC.

 -- andy

Andy Friesen wrote:

> erik wrote:
> 
>> D looks like a nifty language.
>> I maybe would have left out a few things: templates (doesn't add much value),
> 
> 
> The usefulness of templates becomes more apparent as you get used to thinking with them.

You know?  That's true... I find that they are practically useless if you don't know how use them.  Go figure!  ;-)

Kris wrote:

>Let's have a vote here! This particular trout is:
>
>a) trolling
>b) just found BASIC and figures goto is the height of cool
>c) self-consumed grad-student who ... oh never mind.
>
>"There, but for the grace of God, goes God"; Winston Churchill
>  
>
It's not nice to suggest people are trolls, not the D way <g>.

-- 
-Anderson: http://badmama.com.au/~anderson/

J Anderson wrote:
> not the D way <g>.

And thus begins the first steps toward there one day being published, "The Tao of D"...

-C. Sauls
-Invironz