Just to clarify Ant's concern:

As far as I'm concerned, what I'm doing is *a* DTL, not *the* DTL. In other words, there's no agreement between me and Walter that this is automatically *the* (container) template library for D. If someone else is to write one that's better, then presumably that will become the standard.

Yes, Walter's making compiler/language changes as a result of my work on (my) DTL, but these changes are things that are required generally, and may be part of any DTL.

As well as the practical reasons I've stated in other parts of this thread, and in other threads, the reason I'm working on this alone at the moment is that when I'm concocting things such as this, where there's a large amount of "unknown" involved, I tend to "feel" before I understand it; the same's been the case for many of the techiques I've invented in C++/STLSoft (and which you can read about in Sept/Oct in "Imperfect C++" <G>). As such, there's little point trying to share something, especially when it's not in person but over a NG forum, when you can't even explain what you mean or how it will work, is there?

As soon as the language support is there, and my fundamental designs are in working code, I will release it asap. I am very much hoping that the handful of containers that I will have written will serve as templates - pardon the pun - for others, and that people will at that stage get involved, and write singly-linked lists, multimaps and all the others that have been mentioned. I have neither the time nor the inclination nor (quite) the ego/selfishness to be the author of every aspect of DTL.

As I've said many times over the last few months, nothing's a done deal. I've not done a blood deal with big-W. If you want to write your own DTL and offer that up to everyone, go for it.

Cheers


-- 
Matthew Wilson

Contributing editor, C/C++ Users Journal
    (www.synesis.com.au/articles.html#columns)
STLSoft moderator
    (http://www.stlsoft.org)

"An Englishman by birth, a Yorkshireman by the grace of God" -- Michael Gibbs

-------------------------------------------------------------------------------

Matthew wrote:

>As such, there's little point trying to
>share something, especially when it's not in person but over a NG forum, when you
>can't even explain what you mean or how it will work, is there?
>  
>
I think if DTL was designed by the NG, it would be years before anything was done.  Most of the time would be spent arguing about one aspect or the other.  It is good your making a start on something that will hopefully form a platform for future DTL's.

Will you eventually put DTL on something like dsource, so other smart people (not like myself <g>) can improve your version of DTL?

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

> Matthew wrote:
>
> >As such, there's little point trying to
> >share something, especially when it's not in person but over a NG forum, when
you
> >can't even explain what you mean or how it will work, is there?
> >
> >
> I think if DTL was designed by the NG, it would be years before anything was done.  Most of the time would be spent arguing about one aspect or the other.  It is good your making a start on something that will hopefully form a platform for future DTL's.

Thanks. That's pretty much my opinion.

> Will you eventually put DTL on something like dsource, so other smart people (not like myself <g>) can improve your version of DTL?

I'm not sure. If Walter likes it, and no-one else writes one, I assume it'll go into Phobos. I don't know - haven't really thought that far. Certainly I'm open to that idea.

Matthew wrote:

> As well as the practical reasons I've stated in other parts of this thread, and
> in other threads, the reason I'm working on this alone at the moment is that when
> I'm concocting things such as this, where there's a large amount of "unknown"
> involved, I tend to "feel" before I understand it; the same's been the case for

I didn't mean to clutter up your work, by any means! (sorry, I just get find it fun to poke at the name I use online..)

> many of the techiques I've invented in C++/STLSoft (and which you can read about
> in Sept/Oct in "Imperfect C++" <G>). As such, there's little point trying to
> share something, especially when it's not in person but over a NG forum, when you
> can't even explain what you mean or how it will work, is there?
> 
> As soon as the language support is there, and my fundamental designs are in
> working code, I will release it asap. I am very much hoping that the handful of
> containers that I will have written will serve as templates - pardon the pun -
> for others, and that people will at that stage get involved, and write
> singly-linked lists, multimaps and all the others that have been mentioned. I
> have neither the time nor the inclination nor (quite) the ego/selfishness to be
> the author of every aspect of DTL.

Personally, having read many of your posts on this forum I think you are a qualified person to write the DTL.  While I and others may not always agree with you, that is the nature of us human beings.

As has been said here before, people tend to argue because they want the best or most perfect implementation.  This makes sense, and you can't blame anyone for this - but, developing to please everyone is always going to cause problems.

This is why in collaborative projects, different people get assigned different tasks.  The problem is, if you start like this, you get several different parts that don't interact well.  On the other hand, if you do this after the framework is built... everyone understands how to get it to work together, and logically along side everything else.

So, in my opinion, it is best to have the thing worked out by one such as yourself because it is publicly developed at all.  One person can make something that is "the same" and "logical" much better than a couple of people - and once it gets to the point everyone can work on it, it may even change significantly... but, the framework, the ideas, the consistency will be there.

And, consistency is one of the most important things, in my humble opinion, to have when you are programming - more important than an exactly "perfect" DTL.

Just my opinions though.

-[Unknown]

> As far as I'm concerned, what I'm doing is *a* DTL, not *the* DTL. In
other
> words, there's no agreement between me and Walter that this is
automatically
> *the* (container) template library for D. If someone else is to write one
that's
> better, then presumably that will become the standard.

Go for it!!! D needs every good piece of code. It deserves to succeed.

On Mon, 26 Apr 2004 06:42:43 +1000, Matthew wrote:

> Just to clarify Ant's concern:

thanks :)
> 
> As far as I'm concerned, what I'm doing is *a* DTL, not *the* DTL.

Dont' get me worng. We thank you for it!

I just happen to believe that if show what you have
people that will be using it might contribute.

I said before that this kind of volunteer projects
need to be started by one person. If the project
has merit others will join.

Ant
PS seems that pan's "Articles/cancel" really works...