http://www.jelovic.com/articles/why_people_arent_using_haskell.htm

I came across the link above and I thought of D while reading it.  Basically, the author just ruminates over why Haskell hasn't taken off more and how it seems (he feels) the community has a lot to do with that reason.

I'm not drawing any one to one links here or anything like that... just food for thought.  The D community is just incredible IMO in their dedication to the D language.  Quite frankly, it's part of what kept and increased my interest in D.


-Kramer

P.S. Haskell is a pretty cool language.  I'm not proposing this, but if it were possible to have a partial functional language in D, wow!  That's just day-dreaming though.  D's sweet as it is. :)

On Thu, 09 Feb 2006 21:08:32 +0000, Kramer wrote:

> http://www.jelovic.com/articles/why_people_arent_using_haskell.htm
> 
> I came across the link above and I thought of D while reading it.  Basically, the author just ruminates over why Haskell hasn't taken off more and how it seems (he feels) the community has a lot to do with that reason.
> 
> I'm not drawing any one to one links here or anything like that... just food for thought.  The D community is just incredible IMO in their dedication to the D language.  Quite frankly, it's part of what kept and increased my interest in D.
> 
> 
> -Kramer
> 
> P.S. Haskell is a pretty cool language.  I'm not proposing this, but if it were possible to have a partial functional language in D, wow!  That's just day-dreaming though.  D's sweet as it is. :)

Isn't it a little early for that for the moment ? I've discovered D some weeks ago, and I found it pretty interesting at first, before I ran into the first problems:

- Compiler output is terrible.
- There are still many bugs as one can read in the different news groups.
- There is not really a good library one can use with it.
- An advantage of D when looking at the languages which places it
could take is probably pre-, postconditions and assert. The problem is
that these are not really verbose enough (as many people tend to find -
see D.gnu)

Somehow, it would probably bring D forward a lot if one could trust that the problem is the code one wrote and not the compiler. I don't say other compilers are perfect, but ... it seems there are still a lot of problems here.

So before complaining about not as many people using D as one would like, it would maybe be good to fix these problems and when it's done, I think that many people will, after being introduced to it, surely use it more.

I tried to tell some of my colleagues to try it out, but after a while, I stopped, because I saw that D isn't ready for real use.

I hope I haven't offended anybody and if I did, sorry.

Comments inline...

In article <pan.2006.02.09.21.49.38.264161@gmail.com>, Yves Jacoby says...
>
>On Thu, 09 Feb 2006 21:08:32 +0000, Kramer wrote:
>
>> http://www.jelovic.com/articles/why_people_arent_using_haskell.htm
>> 
>> I came across the link above and I thought of D while reading it.  Basically, the author just ruminates over why Haskell hasn't taken off more and how it seems (he feels) the community has a lot to do with that reason.
>> 
>> I'm not drawing any one to one links here or anything like that... just food for thought.  The D community is just incredible IMO in their dedication to the D language.  Quite frankly, it's part of what kept and increased my interest in D.
>> 
>> 
>> -Kramer
>> 
>> P.S. Haskell is a pretty cool language.  I'm not proposing this, but if it were possible to have a partial functional language in D, wow!  That's just day-dreaming though.  D's sweet as it is. :)
>
>Isn't it a little early for that for the moment ? I've discovered D some weeks ago, and I found it pretty interesting at first, before I ran into the first problems:

Agreed that it's too early.  That's why I said I was just day-dreaming. <g>

>
>- Compiler output is terrible.

Anything specific here?

>- There are still many bugs as one can read in the different news groups.
>- There is not really a good library one can use with it.
>- An advantage of D when looking at the languages which places it
>could take is probably pre-, postconditions and assert. The problem is
>that these are not really verbose enough (as many people tend to find -
>see D.gnu)
>
>Somehow, it would probably bring D forward a lot if one could trust that the problem is the code one wrote and not the compiler. I don't say other compilers are perfect, but ... it seems there are still a lot of problems here.
>
>So before complaining about not as many people using D as one would like, it would maybe be good to fix these problems and when it's done, I think that many people will, after being introduced to it, surely use it more.

I don't think what I said should be taken as complaining; just helpful nudges. Also, the point of the link and what I tried to impart in my message was related to the community (which is an incredible online community dedicated to seeing D flourish).  Walter works on what he can affect and the community does the same. At some point there's a critical mass where the community starts to make things better in tandem with Walter.

>
>I tried to tell some of my colleagues to try it out, but after a while, I stopped, because I saw that D isn't ready for real use.
>
>I hope I haven't offended anybody and if I did, sorry.

No offense. :)  Thanks for the thoughts.

-Kramer

Yves Jacoby wrote:
> On Thu, 09 Feb 2006 21:08:32 +0000, Kramer wrote:
> 
>> http://www.jelovic.com/articles/why_people_arent_using_haskell.htm
>>
>> [SNIP]
>>
>> -Kramer
>>
> 
> Isn't it a little early for that for the moment ? I've discovered D some weeks ago, and I found it pretty interesting at first, before I ran into [SNIP]
> 
> I hope I haven't offended anybody and if I did, sorry.

Yes and Yes.

I think both of you are right. The world is ready for D, but D isn't quite ready for the world*. Once D is ready, we have to make sure D doesn't go the way of Haskell. I see two problems that need fixing:

1) Online presence - (See the thread "Time to release 1.0")
2) Organization - This is a big one. There are TWELVE gui libraries. If
the efforts that went into those were concentrated there could have been
2 or 3 really awesome(finished) libraries*. Someone needed to organize
the effort. This goes for all D community efforts, not just the GUI ones.

How do we make sure that #1 and #2 happens? Suggestions?


*D could be ready for use at my job very very soon. We write in-house simulators for graphics chips. That doesn't require anything but the standard libraries and a good IDE.

Kramer wrote:
> P.S. Haskell is a pretty cool language.  I'm not proposing this, but if it were
> possible to have a partial functional language in D, wow!  That's just
> day-dreaming though.  D's sweet as it is. :)

Yep, Haskell and functional programming in general is definitely very cool. I think FP is probably going to have a much greater role in the future where the pressure for provable program correctness and safety keeps increasing.

But I'd say the ideological difference between functional and imperative programming is too great to merge the best of both worlds into D (I tend not to like imperative languages with 5% of FP added), and even if it would be possible to do it elegantly (like in Scala), it's too late now.

No one forbids us to use both paradigms though. It would be great if in the future D would be *the* imperative language which we could use in addition to FP.

-- 
Niko Korhonen
SW Developer

On Thu, 09 Feb 2006 23:05:53 -0800, nick wrote:

> Yves Jacoby wrote:
>> On Thu, 09 Feb 2006 21:08:32 +0000, Kramer wrote:
>> 
>>> http://www.jelovic.com/articles/why_people_arent_using_haskell.htm
>>>
>>> [SNIP]
>>>
>>> -Kramer
>>>
>> 
>> Isn't it a little early for that for the moment ? I've discovered D some weeks ago, and I found it pretty interesting at first, before I ran into [SNIP]
>> 
>> I hope I haven't offended anybody and if I did, sorry.
> 
> Yes and Yes.
> 
> I think both of you are right. The world is ready for D, but D isn't quite ready for the world*. Once D is ready, we have to make sure D doesn't go the way of Haskell. I see two problems that need fixing:
> 
> 1) Online presence - (See the thread "Time to release 1.0")
> 2) Organization - This is a big one. There are TWELVE gui libraries. If
> the efforts that went into those were concentrated there could have been
> 2 or 3 really awesome(finished) libraries*. Someone needed to organize
> the effort. This goes for all D community efforts, not just the GUI ones.
> 
> How do we make sure that #1 and #2 happens? Suggestions?
> 
> 
> *D could be ready for use at my job very very soon. We write in-house simulators for graphics chips. That doesn't require anything but the standard libraries and a good IDE.


I don't know how the D community is organized internally for the moment, but as a Gentoo user and sometimes patch submitter, seeing how good the Gentoo community works, I would say a community based on the same schemes would be good.

Decide on some responsible people for different parts of the organization
and have a lot of communication. For the moment, as I see it with D, 1
person is carrying all the effort for the compiler on windows,
one person is making it work on linux and many people are working on many
different libraries.

Another thing that could surely give D an advantage is to be portable, but this would have to be decided at the top, such that windows and linux are supported from the root on. This should from my point of view be the main criteria to choose a library and go on developing it effectively.

On Thu, 09 Feb 2006 22:11:55 +0000, Kramer wrote:

> Comments inline...
> 
> In article <pan.2006.02.09.21.49.38.264161@gmail.com>, Yves Jacoby says...
>>
>>On Thu, 09 Feb 2006 21:08:32 +0000, Kramer wrote:
>>
>>> http://www.jelovic.com/articles/why_people_arent_using_haskell.htm
>>> 
>>> I came across the link above and I thought of D while reading it.  Basically, the author just ruminates over why Haskell hasn't taken off more and how it seems (he feels) the community has a lot to do with that reason.
>>> 
>>> I'm not drawing any one to one links here or anything like that... just food for thought.  The D community is just incredible IMO in their dedication to the D language.  Quite frankly, it's part of what kept and increased my interest in D.
>>> 
>>> 
>>> -Kramer
>>> 
>>> P.S. Haskell is a pretty cool language.  I'm not proposing this, but if it were possible to have a partial functional language in D, wow!  That's just day-dreaming though.  D's sweet as it is. :)
>>
>>Isn't it a little early for that for the moment ? I've discovered D some weeks ago, and I found it pretty interesting at first, before I ran into the first problems:
> 
> Agreed that it's too early.  That's why I said I was just day-dreaming. <g>
> 
>>
>>- Compiler output is terrible.
> 
> Anything specific here?
> 
>>- There are still many bugs as one can read in the different news groups.
>>- There is not really a good library one can use with it.
>>- An advantage of D when looking at the languages which places it
>>could take is probably pre-, postconditions and assert. The problem is
>>that these are not really verbose enough (as many people tend to find -
>>see D.gnu)
>>
>>Somehow, it would probably bring D forward a lot if one could trust that the problem is the code one wrote and not the compiler. I don't say other compilers are perfect, but ... it seems there are still a lot of problems here.
>>
>>So before complaining about not as many people using D as one would like, it would maybe be good to fix these problems and when it's done, I think that many people will, after being introduced to it, surely use it more.
> 
> I don't think what I said should be taken as complaining; just helpful nudges. Also, the point of the link and what I tried to impart in my message was related to the community (which is an incredible online community dedicated to seeing D flourish).  Walter works on what he can affect and the community does the same. At some point there's a critical mass where the community starts to make things better in tandem with Walter.
> 

I wasn't saying complaining answering to your post in particular, there are some other posts I read, where people seem to say that there are not enough supporters for D.

>>
>>I tried to tell some of my colleagues to try it out, but after a while, I stopped, because I saw that D isn't ready for real use.
>>
>>I hope I haven't offended anybody and if I did, sorry.
> 
> No offense. :)  Thanks for the thoughts.
> 
> -Kramer

Another thing I would really find helpful for this community would be a how-to, telling people how to start if they want to patch the compiler and especially how-to submit the patches in an ordered manner. I think this would surely save some time to developers, if they get the patches delivered. Choosing to include it or not is then left over to the developer, who could than judge of the quality of the patch.

Yves

In article <pan.2006.02.09.21.49.38.264161@gmail.com>, Yves Jacoby says...
>
>- Compiler output is terrible.

Some specifics?

FWIW, I agree that there are certainly some areas (e.g.: floating point), but
generally DMD holds its own, and blows dynamic systems (Java, .NET) away for a
lot of things, at least from what I've seen.

Thanks,

- Dave

In article <dshe0f$aai$1@digitaldaemon.com>, nick says...
>2) Organization - This is a big one. There are TWELVE gui libraries. If the efforts that went into those were concentrated there could have been 2 or 3 really awesome(finished) libraries*. Someone needed to organize the effort. This goes for all D community efforts, not just the GUI ones.

Since some of the GUI libraries don't seem to have anyone working on them (and even most active projects are still one-man operations anyway), the sum of the hours available to pool to community library projects may not be as great as you think. ;) I'm pretty sure no one is working on dig or unDig these days. If L8night is still developing DFC, he hasn't mentioned it in a long time. Also, at least one library is just for GTK+ fans. Another is just for MacOS fans. Several seem to be limited to Windows. They don't share all of the same goals, so it makes since that there are several projects.

But I still agree with your idea that we can probably accomplish more by working together than by everyone doing their own thing. I think Walter made a good move in this direction by proposing to make DWT the "standard" library.

Also, we should probably start editing our GUI library list (http://www.prowiki.org/wiki4d/wiki.cgi?AvailableGuiLibraries) to reflect which libraries have the best potential (or at least add some brief lists of positives and negatives for each). That way people can quickly find what they're looking for without their eyes glazing over.

jcc7

In article <pan.2006.02.10.11.55.04.160245@gmail.com>, Yves Jacoby says...
>I don't know how the D community is organized internally for the moment, but as a Gentoo user and sometimes patch submitter, seeing how good the Gentoo community works, I would say a community based on the same schemes would be good.

We have a pretty simple organization. Walter is in charge of the D programming language. Everyone else is invited to submit suggestions, bugs, patches, and new modules for the library, but Walter is the only who makes the changes in the language and the DMD compiler.

>Decide on some responsible people for different parts of the organization
>and have a lot of communication. For the moment, as I see it with D, 1
>person is carrying all the effort for the compiler on windows,
>one person is making it work on linux and many people are working on many
>different libraries.

This has been suggested several times in the past (and for D to be successful, it'll have to happen eventually), but until Walter himself asks for help, there's only so much that the rest of us can do. You're right that someone is helping out with gdc (and someone else with gdcmac). And there's also the dsource.org projects. And everyone who helps with Wiki4D and the other wikis benefits the D community. But Walter hasn't let go of the "official" stuff, and I can mostly understand why: D is still pretty young and could easily be destroyed by letting a committee get ahold of it too early.

jcc7