I like D from what I've seen of it, but I am concerned that D might never make it to the mainstream. My concerns are not related to language features and libraries for this that and the other, which I think is where most other concerns have been. My concern is that AFAIK there is no champion for D other than DigitalMars and Walter Bright. I wonder what would have happened to C++ if it had stayed forever in the confines of AT&T and the development was done by Stroustrup only? I think java was also headed this way for a while but an open spec meant other implementations emerged after a while and Sun themselves made the JDE and JRE available for a wide number of platforms. Disaster averted.

I think the section 'Future' in the D language manual says it all when all it talks about is language features. Where's the talk about releasing it, what the license might be, if the specs will be open, what part the GCC family might play, etc etc? BrightD seems to be a bit static since its formation in 2002.

Currently D seems to be available as a ready-to-run environment for few platforms (I might be wrong about this but that's how it looks from where I'm standing). People seem more intent on perfecting it for the platforms available right now. Why not expand the scope before it is perfect? That's what other successful languages tend to do.

C++ got a headstart because of compatibility with C. I think D needs to have some sort of a headstart also. Perhaps this might be done by defining an official POSIX wrapper for D? Then there would be no need to drop to C or C++ when using D since all could be done in D. I think this is one of the things holding Eiffel back, although there is a POSIX wrapper defined now at long last. A POSIX wrapper also means that OS resources can be handled in a OS-independent way, e.g sockets can be done with BSD sockets on POSIX platforms and with WinSock for that other widely deployed OS.

Regards,

Andrew Marlow

> I like D from what I've seen of it, but I am concerned that D might never make
> it to the mainstream. My concerns are not related to language features and
> libraries for this that and the other, which I think is where most other
> concerns have been. My concern is that AFAIK there is no champion for D other
> than DigitalMars and Walter Bright. I wonder what would have happened to C++ if
> it had stayed forever in the confines of AT&T and the development was done by
> Stroustrup only? I think java was also headed this way for a while but an open
> spec meant other implementations emerged after a while and Sun themselves made
> the JDE and JRE available for a wide number of platforms. Disaster averted.

Just What We Need - Yet another person ignorantly prophecises us failure. Walter Bright was one of the people who made C++ popular, and D would have to become popular the same way. Walter has many times expressed that he wants further D compilers to appear without any restrictions. The source code for the front-end (language-dependant) parts of the compiler are open and available under GPL and under the extremely liberal PERL Artistic license.

For now, there have been following working alternative compilers:
* GDC from David Friedman, a GCC based compiler
(currently a few months not updated)
* D.NET from Deja Augustine, which generates CLI Assembly, currently in development.
* DLI from Burton Radons, a compiler with non-optimizing x86 backend, which experimented on adding language features - currently long dead.

> I think the section 'Future' in the D language manual says it all when all it
> talks about is language features. Where's the talk about releasing it, what the
> license might be, if the specs will be open, what part the GCC family might
> play, etc etc? BrightD seems to be a bit static since its formation in 2002.

Walter said many times he would like a D compiler to enter the GNU Compiler Collection. Specs are open. Walter has never been specific on the dates, so it looks like the final will be released when it feels done - there are open principal issues and the progress is not smooth and predictable.

-eye

Andrew Marlow wrote:
>
> I like D from what I've seen of it, but I am concerned that D might never make it to the mainstream. My concerns are not related to language features and libraries for this that and the other, which I think is where most other concerns have been. My concern is that AFAIK there is no champion for D other than DigitalMars and Walter Bright.

A single dedicated developer can in a way accomplish more in consistency and quality than a large company. Think for example TeX Don Knuth.

If you can define "mainstream" the community can think how to put D into it
and how to provide the trust that's needed to make D spread (even more quickly).

Yesterday I visited DMOZ
  http://dmoz.org/Computers/Programming/Languages/
and happily found a well-developed page for D (46 links) and
obviously catching up. D is already "top 50" among 136 languages.

-- 
Helmut Leitner    leitner@hls.via.at
Graz, Austria   www.hls-software.com

P.S. this is the ordered list
  <http://www.wikiservice.at/wiki4d/wiki.cgi?DmozLanguageStatistics/Aug04>

-- 
Helmut Leitner    leitner@hls.via.at
Graz, Austria   www.hls-software.com

> If you can define "mainstream" the community can think how to put D into
it
> and how to provide the trust that's needed to make D spread (even more
quickly).
>
> Yesterday I visited DMOZ
>   http://dmoz.org/Computers/Programming/Languages/
> and happily found a well-developed page for D (46 links) and
> obviously catching up. D is already "top 50" among 136 languages.
>
> --
> Helmut Leitner    leitner@hls.via.at
> Graz, Austria   www.hls-software.com


D is the 8th hit on google for programming language. Even before C/C++ at 11.

Andrew Marlow wrote:

> I like D from what I've seen of it, but I am concerned that D might never make it to the mainstream. My concerns are not related to language features and libraries for this that and the other, which I think is where most other concerns have been. My concern is that AFAIK there is no champion for D other than DigitalMars and Walter Bright. I wonder what would have happened to C++ if it had stayed forever in the confines of AT&T and the development was done by Stroustrup only? I think java was also headed this way for a while but an open spec meant other implementations emerged after a while and Sun themselves made the JDE and JRE available for a wide number of platforms. Disaster averted.

It's true D doesn't have a champion as big as AT&T or Sun and it probably won't for a while. Things take time. c'est la vie.

> I think the section 'Future' in the D language manual says it all when all it talks about is language features. Where's the talk about releasing it, what the license might be, if the specs will be open, what part the GCC family might play, etc etc? BrightD seems to be a bit static since its formation in 2002.

Check out the digitalmars.gnu.D group. It doesn't get too much traffic but it is alive. For the license and spec: check out the distribution.

> Currently D seems to be available as a ready-to-run environment for few platforms (I might be wrong about this but that's how it looks from where I'm standing). People seem more intent on perfecting it for the platforms available right now. Why not expand the scope before it is perfect? That's what other successful languages tend to do.

Again, the gcc port will interest you

> C++ got a headstart because of compatibility with C. I think D needs to have some sort of a headstart also. Perhaps this might be done by defining an official POSIX wrapper for D? Then there would be no need to drop to C or C++ when using D since all could be done in D. I think this is one of the things holding Eiffel back, although there is a POSIX wrapper defined now at long last. A POSIX wrapper also means that OS resources can be handled in a OS-independent way, e.g sockets can be done with BSD sockets on POSIX platforms and with WinSock for that other widely deployed OS.

Sounds interesting. How hard would it be? If it means writing something like cygwin then that's too hard. If it mean writing a dozen or so functions then that's more reasonable. I bet most people would continue to do platform-specific work, though. Windows is after all the biggest platform out there so why take all those Windows developers and make them learn POSIX?

> Regards,
> 
> Andrew Marlow

In article <cfnftd$fgu$1@digitaldaemon.com>, Andrew Marlow says...
>
>I like D from what I've seen of it, but I am concerned that D might never make it to the mainstream. My concerns are not related to language features and libraries for this that and the other, which I think is where most other concerns have been. My concern is that AFAIK there is no champion for D other than DigitalMars and Walter Bright. I wonder what would have happened to C++ if it had stayed forever in the confines of AT&T and the development was done by Stroustrup only? I think java was also headed this way for a while but an open spec meant other implementations emerged after a while and Sun themselves made the JDE and JRE available for a wide number of platforms. Disaster averted.
>

Wacky idea, but maybe try to get some interest from the educational community if that isn't already being done?

One big complaint I've seen is that Java is too far away from the metal to teach some important concepts, but instructors like to standardize on tools so they live with it.

Borland was (is still?) successful at selling licenses for Delphi because it has a shorter learning curve than C++ but still allowed lower-level concepts to be taught. I think probably JBuilder is in that space now.

Heck, considering D is providing platform standardized inline assembler, I think even some of that could be thrown in to build a pretty comprehensive curriculam around the language: assembler concepts, systems, common C library, OOP, C/C++/Java 'feel', generics, GC or manual memory management, contract programming - it has it all.

The price is certainly right.

Big drawback: The newgroups would start to get littered with homework help requests <g>

- Dave

On Mon, 16 Aug 2004 03:07:08 +0000, Dave wrote:
> Big drawback: The newgroups would start to get littered with homework help requests <g>

digitalmars.D.homework

Ben Hinkle wrote:
> Andrew Marlow wrote:

>>handled in a OS-independent way, e.g sockets can be done with BSD sockets
>>on POSIX platforms and with WinSock for that other widely deployed OS.
> 
> 
> Sounds interesting. How hard would it be? If it means writing something like
> cygwin then that's too hard. If it mean writing a dozen or so functions
> then that's more reasonable. I bet most people would continue to do
> platform-specific work, though. Windows is after all the biggest platform
> out there so why take all those Windows developers and make them learn
> POSIX?
> 


POSIX unlocks a lot of doors Win32 doesn't go to (embedded programming).
A D.POSIX interface would not be anything like Cygwin, which actually IMPLEMENTS an incomplete rendition of the POSIX subsystem on top of Win32 (a more standards compliant native POSIX subsystem on Windows comes from...Microsoft (SFU - Interix)).

It would be "just" a wrapping affair around the standard C POSIX interface, ideally kept identical so that POSIX documentation applies; the catch is there is a lot of (tedious) work because the POSIX interface is large, with plenty of macros.

I'm not offering :), just making the point.