I'm a C++ programmer, and I've been keeping an eye on this new language since last October. I am hesistant to start any D pet projects because of the lack of popular libraries similar to STL and Boost. This lack is understandable, considering that D is still young and evolving. But we shouldn't wait too long to put a system in place to manage the developement of libraries, because things will get too dispersed and chaotic. I'd rather not rely on libraries developed by a single person, but instead on libraries that have been peer-reviewed and put through their paces by many people, on many systems.

I'm sure that a lot of effort is wasted because people are trying to implement their own libraries independently of each other. Would it be feasable to set up some kind of library repository web site similar to Boost? Here are some things I can think of that would be cool features of that site:

- Style guide and rules (e.g. mandatory unit tests) so that everyone codes their libraries in a consistent manner that is easy to understand by everybody.
- List of feature requests, ranked by popularity.
- Configuration management (CVS), bug tracking
- System for rationale management, where people can register design alternatives, and their pros and cons. This could be implemented with a simple newsgroup, as long as everybody uses the same convention.
- Libraries/modules would be peer-reviewed before being officially accepted.
- Anyone could contribute to the development of any library/module, as long as his/her work is peer-reviewed (i.e. nobody "owns" a library/module).
- Libraries/modules are designed to be as orthogonal as possible, like the way it is with Deimos.

I think having this kind of democratic system for building up standard modules and libraries would be the optimal way to approach this. Or maybe I'm just dreaming in colours, hehe.

If I could, I'd try to setup such a site myself, but I don't know much about web design other than your typical HTML and CGI in C/C++ (does anyone use CGI anymore?). It's too bad, since I'm currently unemployed and have OODLES of free time on my hands. Maybe I should start learning these dozens of new web design technologies... yech! :-)

Emile Cormier

Emile Cormier wrote:
> I'm a C++ programmer, and I've been keeping an eye on this new language since last October. I am hesistant to start any D pet projects because of the lack of popular libraries similar to STL and Boost. This lack is understandable, considering that D is still young and evolving. But we shouldn't wait too long to put a system in place to manage the developement of libraries, because things will get too dispersed and chaotic. I'd rather not rely on libraries developed by a single person, but instead on libraries that have been peer-reviewed and put through their paces by many people, on many systems.

You're right that most of the projects thus far have been one-person operations. I suspect that's largely do to the newness of D. I feel like I need to work on D stuff by myself first, so that I can actually learn D.

At least we have Wiki4D where we can add our own links: http://www.wikiservice.at/d/wiki.cgi?AllLibraries.

Similar ideas regarding community repositories have been posed in the past, most recently by Brad Anderson (http://www.digitalmars.com/drn-bin/wwwnews?D/24165). I think it's a great idea.

> I'm sure that a lot of effort is wasted because people are trying to implement their own libraries independently of each other. Would it be feasable to set up some kind of library repository web site similar to Boost? Here are some things I can think of that would be cool features of that site:

There is some duplication of effort, but hopefully the better libraries will win out in the long run. "Survival of the fittest" is the goal.

> - Style guide and rules (e.g. mandatory unit tests) so that everyone codes their libraries in a consistent manner that is easy to understand by everybody.
> - List of feature requests, ranked by popularity.
> - Configuration management (CVS), bug tracking
> - System for rationale management, where people can register design alternatives, and their pros and cons. This could be implemented with a simple newsgroup, as long as everybody uses the same convention.
> - Libraries/modules would be peer-reviewed before being officially accepted.
> - Anyone could contribute to the development of any library/module, as long as his/her work is peer-reviewed (i.e. nobody "owns" a library/module).
> - Libraries/modules are designed to be as orthogonal as possible, like the way it is with Deimos.

All interesting ideas.

> 
> I think having this kind of democratic system for building up standard modules and libraries would be the optimal way to approach this. Or maybe I'm just dreaming in colours, hehe.

D isn't exactly a democracy, but Walter will listen to reason if we all gang up on him. ;)

> If I could, I'd try to setup such a site myself, but I don't know much about web design other than your typical HTML and CGI in C/C++ (does anyone use CGI anymore?). It's too bad, since I'm currently unemployed and have OODLES of free time on my hands. Maybe I should start learning these dozens of new web design technologies... yech! :-)

I've been learning PHP and ASP recently. I prefer PHP because I find ASP to be too proprietary.

> Emile Cormier


-- 
Justin
http://jcc_7.tripod.com/d/

J C Calvarese schrieb:

>> I think having this kind of democratic system for building up standard modules and libraries would be the optimal way to approach this. Or maybe I'm just dreaming in colours, hehe.
> 
> D isn't exactly a democracy, but Walter will listen to reason if we all gang up on him. ;)

Actually no problem. That is, the standard library isn't irreplacable. We could set up a development of a replacement standard library, and Walter may decide what gets incorporated into standard and compiler distro.

>> If I could, I'd try to setup such a site myself, but I don't know much about web design other than your typical HTML and CGI in C/C++ (does anyone use CGI anymore?). It's too bad, since I'm currently unemployed and have OODLES of free time on my hands. Maybe I should start learning these dozens of new web design technologies... yech! :-)

Free time is a precious resource most of us lack. ;) If you need some design that is simple and loads fast, you can simply borrow one from here: http://www.oswd.org/ . I did some fancy design stuff a few years ago which even sized correctly and worked well in Netscape 4 (!!!!) but i'd think it's plain wrong for a serious website and i don't have time for such things any longer.

If i recall correctly Benji Smith said he had *almost* setup something like a central project repository. His newsgroup e-mail is dlanguage@xxagg.com

> I've been learning PHP and ASP recently. I prefer PHP because I find ASP to be too proprietary.

-eye

"Emile Cormier" <ecorm@nbnet-dot-nb.ca> wrote in message news:c38bep$2o6h$4@digitaldaemon.com...
> I'm a C++ programmer, and I've been keeping an eye on this new language since last October.

Where did you hear about it?

> I am hesistant to start any D pet projects because
> of the lack of popular libraries similar to STL and Boost.

I'm currently working on the DTL, and have got several basic implementations done. I expect to make more progress once Walter's back from SDWest, and we can make some changes to the compiler.

> This lack is
> understandable, considering that D is still young and evolving. But we
> shouldn't wait too long to put a system in place to manage the
> developement of libraries, because things will get too dispersed and
> chaotic. I'd rather not rely on libraries developed by a single person,
> but instead on libraries that have been peer-reviewed and put through
> their paces by many people, on many systems.

The group's still small enough that if you want to contribute something, and ask for review, you'll get enough volunteers.

> I'm sure that a lot of effort is wasted because people are trying to implement their own libraries independently of each other. Would it be feasable to set up some kind of library repository web site similar to Boost? Here are some things I can think of that would be cool features of that site:
>
> - Style guide and rules (e.g. mandatory unit tests) so that everyone
> codes their libraries in a consistent manner that is easy to understand
> by everybody.
> - List of feature requests, ranked by popularity.

This would be good

> - Configuration management (CVS), bug tracking
> - System for rationale management, where people can register design
> alternatives, and their pros and cons. This could be implemented with a
> simple newsgroup, as long as everybody uses the same convention.
> - Libraries/modules would be peer-reviewed before being officially
accepted.
> - Anyone could contribute to the development of any library/module, as long as his/her work is peer-reviewed (i.e. nobody "owns" a
library/module).
> - Libraries/modules are designed to be as orthogonal as possible, like the way it is with Deimos.
>
> I think having this kind of democratic system for building up standard modules and libraries would be the optimal way to approach this. Or maybe I'm just dreaming in colours, hehe.
>
> If I could, I'd try to setup such a site myself, but I don't know much about web design other than your typical HTML and CGI in C/C++ (does anyone use CGI anymore?). It's too bad, since I'm currently unemployed and have OODLES of free time on my hands. Maybe I should start learning these dozens of new web design technologies... yech! :-)

Looks to me like you've answered your own question.

I'm sorry for not having followed this newsgroup religiously since I first knew about D. Up until now, I've only been "dipping" into the newsgroup on occasion, mostly to find out if D will have generic programming features as powerful as C++. From your comments, it seems that a few of the questions I raised have been covered before.

Other comments embedded below.

Matthew wrote:
> "Emile Cormier" <ecorm@nbnet-dot-nb.ca> wrote in message
> news:c38bep$2o6h$4@digitaldaemon.com...
> 
>>I'm a C++ programmer, and I've been keeping an eye on this new language
>>since last October.
> 
> 
> Where did you hear about it?
> 

I saw someone post about the D language in a game development website. I don't remember which one though.

> 
>>I am hesistant to start any D pet projects because
>>of the lack of popular libraries similar to STL and Boost.
> 
> 
> I'm currently working on the DTL, and have got several basic implementations
> done. I expect to make more progress once Walter's back from SDWest, and we
> can make some changes to the compiler.
> 

Is this DTL a port of the STL? If so, then it would certainly help attract people from the C++ flock who are used to the STL. On the other hand, should we be copying blindly from the STL? Maybe there are some design flaws we should avoid. Then again, I don't remember noticing any flaws while studying the STL. I only remember being impressed by it's sheer elegance. :-) Maybe some STL wizards (such as you) could point out some design flaws. Maybe some design decisions behind the STL don't apply to the D language. This is why I'd like to see some kind of centralized rationale management system, so that issues like this can be examined before library interfaces are "etched into stone".

> 
>>This lack is
>>understandable, considering that D is still young and evolving. But we
>>shouldn't wait too long to put a system in place to manage the
>>developement of libraries, because things will get too dispersed and
>>chaotic. I'd rather not rely on libraries developed by a single person,
>>but instead on libraries that have been peer-reviewed and put through
>>their paces by many people, on many systems.
> 
> 
> The group's still small enough that if you want to contribute something, and
> ask for review, you'll get enough volunteers.
> 
> 
>>I'm sure that a lot of effort is wasted because people are trying to
>>implement their own libraries independently of each other. Would it be
>>feasable to set up some kind of library repository web site similar to
>>Boost? Here are some things I can think of that would be cool features
>>of that site:
>>
>>- Style guide and rules (e.g. mandatory unit tests) so that everyone
>>codes their libraries in a consistent manner that is easy to understand
>>by everybody.
>>- List of feature requests, ranked by popularity.
> 
> 
> This would be good
> 
> 
>>- Configuration management (CVS), bug tracking
>>- System for rationale management, where people can register design
>>alternatives, and their pros and cons. This could be implemented with a
>>simple newsgroup, as long as everybody uses the same convention.
>>- Libraries/modules would be peer-reviewed before being officially
> 
> accepted.
> 
>>- Anyone could contribute to the development of any library/module, as
>>long as his/her work is peer-reviewed (i.e. nobody "owns" a
> 
> library/module).
> 
>>- Libraries/modules are designed to be as orthogonal as possible, like
>>the way it is with Deimos.
>>
>>I think having this kind of democratic system for building up standard
>>modules and libraries would be the optimal way to approach this. Or
>>maybe I'm just dreaming in colours, hehe.
>>
>>If I could, I'd try to setup such a site myself, but I don't know much
>>about web design other than your typical HTML and CGI in C/C++ (does
>>anyone use CGI anymore?). It's too bad, since I'm currently unemployed
>>and have OODLES of free time on my hands. Maybe I should start learning
>>these dozens of new web design technologies... yech! :-)
> 
> 
> Looks to me like you've answered your own question.
> 
> 

It seems that other people are already getting a repository site together, so I won't duplicate the effort by building my own from scratch. :-) What seems to be missing so far is a system for rationale management. I've done some googling, and it seems that there still isn't any specialized software available for that purpose. I'm surprised that the open source community hasn't yet used rationale management to get more people involved in the decision-making process. If Walter had such a system in place, I bet that he would find it much easier to collect ideas/suggestions from the community and weigh them in a systematic manner according to his design criteria. Decisions would then become much less gut wrenching.

Maybe the best way for me to contribute is to look more into rationale management, and to perhaps start developing a system for it that is usable for large open source projects. From what I can tell by web searches, it seems that rationale management hasn't yet been adopted by the mainstream development community. For those who are interested, this  subject is covered in chapter 8 of "Object Oriented Software Engineering" by Bruegge and Dutoit. I haven't found any websites that discusses the subject in a general manner, but I haven't dug too deep either.

I have found a freely downloadable groupware tool called Compendium for mapping discussions in a manner that would allow rationale management:

http://compendiuminstitute.org

Users can collaborate over the Internet via a Java interface if they have the Compendium client. It is not possible for users to collaborate via a web interface, but it is possible to generate an HTML file with a the discussion on it. However, they plan to release Compendium as open source in a few months, so it might be possible to create some kind of web interface for it.

If any of you are well-endowed in the arcane knowledge of law, could you check out the licence agreement to see if there would be a problem for us using it for an open source project?

http://www.compendiuminstitute.org/download/license.htm



Emile Cormier wrote:
> I'm a C++ programmer, and I've been keeping an eye on this new language since last October. I am hesistant to start any D pet projects because of the lack of popular libraries similar to STL and Boost. This lack is understandable, considering that D is still young and evolving. But we shouldn't wait too long to put a system in place to manage the developement of libraries, because things will get too dispersed and chaotic. I'd rather not rely on libraries developed by a single person, but instead on libraries that have been peer-reviewed and put through their paces by many people, on many systems.
> 
> I'm sure that a lot of effort is wasted because people are trying to implement their own libraries independently of each other. Would it be feasable to set up some kind of library repository web site similar to Boost? Here are some things I can think of that would be cool features of that site:
> 
> - Style guide and rules (e.g. mandatory unit tests) so that everyone codes their libraries in a consistent manner that is easy to understand by everybody.
> - List of feature requests, ranked by popularity.
> - Configuration management (CVS), bug tracking
> - System for rationale management, where people can register design alternatives, and their pros and cons. This could be implemented with a simple newsgroup, as long as everybody uses the same convention.
> - Libraries/modules would be peer-reviewed before being officially accepted.
> - Anyone could contribute to the development of any library/module, as long as his/her work is peer-reviewed (i.e. nobody "owns" a library/module).
> - Libraries/modules are designed to be as orthogonal as possible, like the way it is with Deimos.
> 
> I think having this kind of democratic system for building up standard modules and libraries would be the optimal way to approach this. Or maybe I'm just dreaming in colours, hehe.
> 
> If I could, I'd try to setup such a site myself, but I don't know much about web design other than your typical HTML and CGI in C/C++ (does anyone use CGI anymore?). It's too bad, since I'm currently unemployed and have OODLES of free time on my hands. Maybe I should start learning these dozens of new web design technologies... yech! :-)
> 
> Emile Cormier

"Emile Cormier" <ecorm@nbnet-dot-nb.ca> wrote in message news:c39u92$2d70$1@digitaldaemon.com...
> I'm sorry for not having followed this newsgroup religiously since I first knew about D. Up until now, I've only been "dipping" into the newsgroup on occasion, mostly to find out if D will have generic programming features as powerful as C++.

We don't, yet, but it is very likely that it will be superior by the end of year.

> From your comments, it seems
> that a few of the questions I raised have been covered before.
>
> Other comments embedded below.
>
> Matthew wrote:
> > "Emile Cormier" <ecorm@nbnet-dot-nb.ca> wrote in message news:c38bep$2o6h$4@digitaldaemon.com...
> >
> >>I'm a C++ programmer, and I've been keeping an eye on this new language since last October.
> >
> >
> > Where did you hear about it?
> >
>
> I saw someone post about the D language in a game development website. I don't remember which one though.
>
> >
> >>I am hesistant to start any D pet projects because
> >>of the lack of popular libraries similar to STL and Boost.
> >
> >
> > I'm currently working on the DTL, and have got several basic
implementations
> > done. I expect to make more progress once Walter's back from SDWest, and
we
> > can make some changes to the compiler.
> >
>
> Is this DTL a port of the STL?

No, though there is some influence

> If so, then it would certainly help
> attract people from the C++ flock who are used to the STL. On the other
> hand, should we be copying blindly from the STL?

I think other people are trying to do this, but it's not the approach we're taking. It will more closely map D's notions of slices, in what I call Ranges. I have been working on the Range concept for C++ for some time with a couple of other people, some of which is featured in "Imperfect C++" (my new book, due out later this year), and many of the same aspects of the concept translate to D.

> Maybe there are some
> design flaws we should avoid. Then again, I don't remember noticing any
> flaws while studying the STL. I only remember being impressed by it's
> sheer elegance. :-) Maybe some STL wizards (such as you) could point out
> some design flaws.

Hmmm. That's a big question with a long answer. I'm taking the fifth on that, for the moment. No doubt when we release DTL 0.1 to the ng there will be vast amounts of debate on these issues then.

> Maybe some design decisions behind the STL don't
> apply to the D language.

That is my belief.

> This is why I'd like to see some kind of
> centralized rationale management system, so that issues like this can be
> examined before library interfaces are "etched into stone".

That's what the NG is for.

There's nothing etched in stone yet, since we expect there to be a lot of debate before the DTL takes its final form.

> >
> >>This lack is
> >>understandable, considering that D is still young and evolving. But we
> >>shouldn't wait too long to put a system in place to manage the
> >>developement of libraries, because things will get too dispersed and
> >>chaotic. I'd rather not rely on libraries developed by a single person,
> >>but instead on libraries that have been peer-reviewed and put through
> >>their paces by many people, on many systems.
> >
> >
> > The group's still small enough that if you want to contribute something,
and
> > ask for review, you'll get enough volunteers.
> >
> >
> >>I'm sure that a lot of effort is wasted because people are trying to implement their own libraries independently of each other. Would it be feasable to set up some kind of library repository web site similar to Boost? Here are some things I can think of that would be cool features of that site:
> >>
> >>- Style guide and rules (e.g. mandatory unit tests) so that everyone
> >>codes their libraries in a consistent manner that is easy to understand
> >>by everybody.
> >>- List of feature requests, ranked by popularity.
> >
> >
> > This would be good
> >
> >
> >>- Configuration management (CVS), bug tracking
> >>- System for rationale management, where people can register design
> >>alternatives, and their pros and cons. This could be implemented with a
> >>simple newsgroup, as long as everybody uses the same convention.
> >>- Libraries/modules would be peer-reviewed before being officially
> >
> > accepted.
> >
> >>- Anyone could contribute to the development of any library/module, as long as his/her work is peer-reviewed (i.e. nobody "owns" a
> >
> > library/module).
> >
> >>- Libraries/modules are designed to be as orthogonal as possible, like the way it is with Deimos.
> >>
> >>I think having this kind of democratic system for building up standard modules and libraries would be the optimal way to approach this. Or maybe I'm just dreaming in colours, hehe.
> >>
> >>If I could, I'd try to setup such a site myself, but I don't know much about web design other than your typical HTML and CGI in C/C++ (does anyone use CGI anymore?). It's too bad, since I'm currently unemployed and have OODLES of free time on my hands. Maybe I should start learning these dozens of new web design technologies... yech! :-)
> >
> >
> > Looks to me like you've answered your own question.
> >
> >
>
> It seems that other people are already getting a repository site together, so I won't duplicate the effort by building my own from scratch. :-)

There's been a lot of talk, but I know of no site as yet. Maybe you should write a collaborative site to manage the efforts to make collaborative sites? :-)

> What seems to be missing so far is a system for rationale
> management. I've done some googling, and it seems that there still isn't
> any specialized software available for that purpose. I'm surprised that
> the open source community hasn't yet used rationale management to get
> more people involved in the decision-making process. If Walter had such
> a system in place, I bet that he would find it much easier to collect
> ideas/suggestions from the community and weigh them in a systematic
> manner according to his design criteria. Decisions would then become
> much less gut wrenching.
>
> Maybe the best way for me to contribute is to look more into rationale
> management, and to perhaps start developing a system for it that is
> usable for large open source projects. From what I can tell by web
> searches, it seems that rationale management hasn't yet been adopted by
> the mainstream development community. For those who are interested, this
>   subject is covered in chapter 8 of "Object Oriented Software
> Engineering" by Bruegge and Dutoit. I haven't found any websites that
> discusses the subject in a general manner, but I haven't dug too deep
> either.

Java? Java!?

Seriously, if it's not web-based, it's not going to fly. ;/


"Emile Cormier" <ecorm@nbnet-dot-nb.ca> wrote in message news:c3a5ai$2qnn$1@digitaldaemon.com...
> I have found a freely downloadable groupware tool called Compendium for mapping discussions in a manner that would allow rationale management:
>
> http://compendiuminstitute.org
>
> Users can collaborate over the Internet via a Java interface if they have the Compendium client. It is not possible for users to collaborate via a web interface, but it is possible to generate an HTML file with a the discussion on it. However, they plan to release Compendium as open source in a few months, so it might be possible to create some kind of web interface for it.
>
> If any of you are well-endowed in the arcane knowledge of law, could you check out the licence agreement to see if there would be a problem for us using it for an open source project?
>
> http://www.compendiuminstitute.org/download/license.htm
>
>
>
> Emile Cormier wrote:
> > I'm a C++ programmer, and I've been keeping an eye on this new language since last October. I am hesistant to start any D pet projects because of the lack of popular libraries similar to STL and Boost. This lack is understandable, considering that D is still young and evolving. But we shouldn't wait too long to put a system in place to manage the developement of libraries, because things will get too dispersed and chaotic. I'd rather not rely on libraries developed by a single person, but instead on libraries that have been peer-reviewed and put through their paces by many people, on many systems.
> >
> > I'm sure that a lot of effort is wasted because people are trying to implement their own libraries independently of each other. Would it be feasable to set up some kind of library repository web site similar to Boost? Here are some things I can think of that would be cool features of that site:
> >
> > - Style guide and rules (e.g. mandatory unit tests) so that everyone
> > codes their libraries in a consistent manner that is easy to understand
> > by everybody.
> > - List of feature requests, ranked by popularity.
> > - Configuration management (CVS), bug tracking
> > - System for rationale management, where people can register design
> > alternatives, and their pros and cons. This could be implemented with a
> > simple newsgroup, as long as everybody uses the same convention.
> > - Libraries/modules would be peer-reviewed before being officially
> > accepted.
> > - Anyone could contribute to the development of any library/module, as
> > long as his/her work is peer-reviewed (i.e. nobody "owns" a
> > library/module).
> > - Libraries/modules are designed to be as orthogonal as possible, like
> > the way it is with Deimos.
> >
> > I think having this kind of democratic system for building up standard modules and libraries would be the optimal way to approach this. Or maybe I'm just dreaming in colours, hehe.
> >
> > If I could, I'd try to setup such a site myself, but I don't know much about web design other than your typical HTML and CGI in C/C++ (does anyone use CGI anymore?). It's too bad, since I'm currently unemployed and have OODLES of free time on my hands. Maybe I should start learning these dozens of new web design technologies... yech! :-)
> >
> > Emile Cormier

> It seems that other people are already getting a repository site together, so I won't duplicate the effort by building my own from scratch. :-) What seems to be missing so far is a system for rationale management. I've done some googling, and it seems that there still isn't any specialized software available for that purpose. I'm surprised that the open source community hasn't yet used rationale management to get more people involved in the decision-making process. If Walter had such a system in place, I bet that he would find it much easier to collect ideas/suggestions from the community and weigh them in a systematic manner according to his design criteria. Decisions would then become much less gut wrenching.


I dont think anyone is currently working on a repository site , or if they were its long since died.  Several have been proposed but none ever got off the ground.  I truly hope you will start ( and complete! ) one, im sure many of us will lend time if you need.  I've never heard of rationale management Ill have to look it up ;).

C

On Wed, 17 Mar 2004 12:27:09 -0400, Emile Cormier <ecorm@nbnet-dot-nb.ca> wrote:

> I'm sorry for not having followed this newsgroup religiously since I first knew about D. Up until now, I've only been "dipping" into the newsgroup on occasion, mostly to find out if D will have generic programming features as powerful as C++. From your comments, it seems that a few of the questions I raised have been covered before.
>
> Other comments embedded below.
>
> Matthew wrote:
>> "Emile Cormier" <ecorm@nbnet-dot-nb.ca> wrote in message
>> news:c38bep$2o6h$4@digitaldaemon.com...
>>
>>> I'm a C++ programmer, and I've been keeping an eye on this new language
>>> since last October.
>>
>>
>> Where did you hear about it?
>>
>
> I saw someone post about the D language in a game development website. I don't remember which one though.
>
>>
>>> I am hesistant to start any D pet projects because
>>> of the lack of popular libraries similar to STL and Boost.
>>
>>
>> I'm currently working on the DTL, and have got several basic implementations
>> done. I expect to make more progress once Walter's back from SDWest, and we
>> can make some changes to the compiler.
>>
>
> Is this DTL a port of the STL? If so, then it would certainly help attract people from the C++ flock who are used to the STL. On the other hand, should we be copying blindly from the STL? Maybe there are some design flaws we should avoid. Then again, I don't remember noticing any flaws while studying the STL. I only remember being impressed by it's sheer elegance. :-) Maybe some STL wizards (such as you) could point out some design flaws. Maybe some design decisions behind the STL don't apply to the D language. This is why I'd like to see some kind of centralized rationale management system, so that issues like this can be examined before library interfaces are "etched into stone".
>
>>
>>> This lack is
>>> understandable, considering that D is still young and evolving. But we
>>> shouldn't wait too long to put a system in place to manage the
>>> developement of libraries, because things will get too dispersed and
>>> chaotic. I'd rather not rely on libraries developed by a single person,
>>> but instead on libraries that have been peer-reviewed and put through
>>> their paces by many people, on many systems.
>>
>>
>> The group's still small enough that if you want to contribute something, and
>> ask for review, you'll get enough volunteers.
>>
>>
>>> I'm sure that a lot of effort is wasted because people are trying to
>>> implement their own libraries independently of each other. Would it be
>>> feasable to set up some kind of library repository web site similar to
>>> Boost? Here are some things I can think of that would be cool features
>>> of that site:
>>>
>>> - Style guide and rules (e.g. mandatory unit tests) so that everyone
>>> codes their libraries in a consistent manner that is easy to understand
>>> by everybody.
>>> - List of feature requests, ranked by popularity.
>>
>>
>> This would be good
>>
>>
>>> - Configuration management (CVS), bug tracking
>>> - System for rationale management, where people can register design
>>> alternatives, and their pros and cons. This could be implemented with a
>>> simple newsgroup, as long as everybody uses the same convention.
>>> - Libraries/modules would be peer-reviewed before being officially
>>
>> accepted.
>>
>>> - Anyone could contribute to the development of any library/module, as
>>> long as his/her work is peer-reviewed (i.e. nobody "owns" a
>>
>> library/module).
>>
>>> - Libraries/modules are designed to be as orthogonal as possible, like
>>> the way it is with Deimos.
>>>
>>> I think having this kind of democratic system for building up standard
>>> modules and libraries would be the optimal way to approach this. Or
>>> maybe I'm just dreaming in colours, hehe.
>>>
>>> If I could, I'd try to setup such a site myself, but I don't know much
>>> about web design other than your typical HTML and CGI in C/C++ (does
>>> anyone use CGI anymore?). It's too bad, since I'm currently unemployed
>>> and have OODLES of free time on my hands. Maybe I should start learning
>>> these dozens of new web design technologies... yech! :-)
>>
>>
>> Looks to me like you've answered your own question.
>>
>>
>
> It seems that other people are already getting a repository site together, so I won't duplicate the effort by building my own from scratch. :-) What seems to be missing so far is a system for rationale management. I've done some googling, and it seems that there still isn't any specialized software available for that purpose. I'm surprised that the open source community hasn't yet used rationale management to get more people involved in the decision-making process. If Walter had such a system in place, I bet that he would find it much easier to collect ideas/suggestions from the community and weigh them in a systematic manner according to his design criteria. Decisions would then become much less gut wrenching.
>
> Maybe the best way for me to contribute is to look more into rationale management, and to perhaps start developing a system for it that is usable for large open source projects. From what I can tell by web searches, it seems that rationale management hasn't yet been adopted by the mainstream development community. For those who are interested, this   subject is covered in chapter 8 of "Object Oriented Software Engineering" by Bruegge and Dutoit. I haven't found any websites that discusses the subject in a general manner, but I haven't dug too deep either.



-- 
Using M2, Opera's revolutionary e-mail client: http://www.opera.com/m2/

I would donate webspace and bandwidth to this cause.

I own Deksesuma.com, a soon to be shareware game and application project. I get 1.5 gig of space and 150 gig bandwidth. That should be plenty alone for a while. I don't think my site alone would use that much in a month.

If you want, contact me at this address (webmaster@deksesuma.com) and I'll set you up with a ftp account to upload content.

I also do some PHP, so I could help out with that also :).




Matthew wrote:
> Java? Java!?
> 
> Seriously, if it's not web-based, it's not going to fly. ;/
> 
> 
> "Emile Cormier" <ecorm@nbnet-dot-nb.ca> wrote in message
> news:c3a5ai$2qnn$1@digitaldaemon.com...
> 
>>I have found a freely downloadable groupware tool called Compendium for
>>mapping discussions in a manner that would allow rationale management:
>>
>>http://compendiuminstitute.org
>>
>>Users can collaborate over the Internet via a Java interface if they
>>have the Compendium client. It is not possible for users to collaborate
>>via a web interface, but it is possible to generate an HTML file with a
>>the discussion on it. However, they plan to release Compendium as open
>>source in a few months, so it might be possible to create some kind of
>>web interface for it.
>>
>>If any of you are well-endowed in the arcane knowledge of law, could you
>>check out the licence agreement to see if there would be a problem for
>>us using it for an open source project?
>>
>>http://www.compendiuminstitute.org/download/license.htm
>>
>>
>>
>>Emile Cormier wrote:
>>
>>>I'm a C++ programmer, and I've been keeping an eye on this new language
>>>since last October. I am hesistant to start any D pet projects because
>>>of the lack of popular libraries similar to STL and Boost. This lack is
>>>understandable, considering that D is still young and evolving. But we
>>>shouldn't wait too long to put a system in place to manage the
>>>developement of libraries, because things will get too dispersed and
>>>chaotic. I'd rather not rely on libraries developed by a single person,
>>>but instead on libraries that have been peer-reviewed and put through
>>>their paces by many people, on many systems.
>>>
>>>I'm sure that a lot of effort is wasted because people are trying to
>>>implement their own libraries independently of each other. Would it be
>>>feasable to set up some kind of library repository web site similar to
>>>Boost? Here are some things I can think of that would be cool features
>>>of that site:
>>>
>>>- Style guide and rules (e.g. mandatory unit tests) so that everyone
>>>codes their libraries in a consistent manner that is easy to understand
>>>by everybody.
>>>- List of feature requests, ranked by popularity.
>>>- Configuration management (CVS), bug tracking
>>>- System for rationale management, where people can register design
>>>alternatives, and their pros and cons. This could be implemented with a
>>>simple newsgroup, as long as everybody uses the same convention.
>>>- Libraries/modules would be peer-reviewed before being officially
>>>accepted.
>>>- Anyone could contribute to the development of any library/module, as
>>>long as his/her work is peer-reviewed (i.e. nobody "owns" a
>>>library/module).
>>>- Libraries/modules are designed to be as orthogonal as possible, like
>>>the way it is with Deimos.
>>>
>>>I think having this kind of democratic system for building up standard
>>>modules and libraries would be the optimal way to approach this. Or
>>>maybe I'm just dreaming in colours, hehe.
>>>
>>>If I could, I'd try to setup such a site myself, but I don't know much
>>>about web design other than your typical HTML and CGI in C/C++ (does
>>>anyone use CGI anymore?). It's too bad, since I'm currently unemployed
>>>and have OODLES of free time on my hands. Maybe I should start learning
>>>these dozens of new web design technologies... yech! :-)
>>>
>>>Emile Cormier
> 
> 
>