November 24
On Wednesday, 24 November 2021 at 21:41:10 UTC, forkit wrote:
> On Wednesday, 24 November 2021 at 17:14:56 UTC, IGotD- wrote:
>> ..
>> I'd say lets not finish it and start on D3 instead.
>
> No. D2 needs to first come to a design halt. It seems far from that at the moment.

Agreed.  I think we *need* a stable D2 before we can hope for widespread adoption.  This could mean either finishing the incomplete things or backing them out, but it's got to go one way or the other on all significant outstanding things.  Making a list of these things and deciding if they are "in" or "out" of the upcoming LTS release should be job #1.  And it must be made clear that there will, indeed, be an LTS release upcoming.  Without that clearly stated no one should use D for anything other than hobby programming.  To do so would be irresponsible.

>
> D3 is inevitable, of course. The only question is what design decisions will 'force' that move to the next version of D?
>
> I'd be more interested to hear in detail, about that aspect of moving D forward, rather just the call to dump D2 and create a D3.
>
> No fluffy 'vision'.. I wanna see the details.

Well, I think first having a general vision is a prerequisite before details can be worked out...  No one person can work out all the details, and multiple people need to adopt some shared vision in order to move things forward.  Lack of shared vision is one of our biggest problems.
November 24

On Wednesday, 24 November 2021 at 21:37:58 UTC, Dukc wrote:

>

The assumption here appears to be that since the people we're trying to attract are not already using D, people already using D can't know what would buy them in, but a word from an outside complainer is more reliable. People know their own motivations best, right?

I think this overstating the issue. Not-yet-enthusiatic-but-interested people may be using D, but not for anything that they will commit to long term.

You need more of those to be enthusiastic in order to get frameworks built. You cannot make them enthusiastic without addressing the issues they have.

>

Given that, It does not sound a very good idea to design the language around what everyone is lobbying for on the forums.

Early C++ compilers were buggy. Not a good example. Just a fun anecdote. (Even g++ was quite buggy for a very long time.)

As a designer you should ask yourself why N people independently point out weakness X, and then ask yourself what the total impact of X is.

>

I'd much rather concentrate on specific bug reports, questions and improvement proposals. With them the designer can at least trust they show something that really matters, not just made-up excuses for some unacknowledged bias.

You should of course polish your product!

However, it is generally negative to add a long series of "improvements" to scratch some enthusiastic individual's itch and become slightly better at something you already have covered. You gain more from extending into areas that you do not cover well.

What would happen if the most enthusiastic C# or TypeScript users were given the freedom to add new cool features? The languages would eventually fail to appeal to any other group than the most enthusiastic ones.

November 25

On Wednesday, 24 November 2021 at 21:55:46 UTC, Greg Strong wrote:

>

Lack of shared vision is one of our biggest problems.

D's vision can be summed up as better, faster, safer and more elegant.
As long as we stick to that goal, you can attract your target user.

November 25

On Thursday, 25 November 2021 at 00:56:07 UTC, zjh wrote:

>

On Wednesday, 24 November 2021 at 21:55:46 UTC, Greg Strong wrote:

>

Lack of shared vision is one of our biggest problems.

D's vision can be summed up as better, faster, safer and more elegant.
As long as we stick to that goal, you can attract your target user.

better than what? (than shooting yourself in the foot?)
faster than what? (a speeding train?)
safer than what? (no safety?)
more elegant than what? (the inelegance of C and C++?)

They just fluffy adjectives with no real meaning.

November 25

On Thursday, 25 November 2021 at 00:56:07 UTC, zjh wrote:

>

On Wednesday, 24 November 2021 at 21:55:46 UTC, Greg Strong wrote:

>

Lack of shared vision is one of our biggest problems.

D's vision can be summed up as better, faster, safer and more elegant.
As long as we stick to that goal, you can attract your target user.

The vision, as I see it (at the moment), can be summed up like this:

D++

:-(

November 25

On Thursday, 25 November 2021 at 01:19:50 UTC, forkit wrote:

>

The vision, as I see it (at the moment), can be summed up like this:
D++
:-(

How is D++ bad?
I like C++. It's easy to use.
It would be nice if we had D++.

November 25

On Thursday, 25 November 2021 at 01:07:20 UTC, forkit wrote:

>

better than what? (than shooting yourself in the foot?)

D just needs to be faster, safer, more elegant, better than D its self.

November 25

On Thursday, 25 November 2021 at 01:49:37 UTC, zjh wrote:

>

On Thursday, 25 November 2021 at 01:19:50 UTC, forkit wrote:

>

The vision, as I see it (at the moment), can be summed up like this:
D++
:-(

How is D++ bad?
I like C++. It's easy to use.
It would be nice if we had D++.

My concern is that a D++ will for 'compatability reasons' bring with it, many of the flaws in C++ - just as C++ bought with it many of the flaws in C.

This is where I think, Go got it right (although they made some design decisions that I simply refuse to accept).

What we really need is a simpler compiler that we can trust (through formal verfication and proofs). That is how one implements @safe and @trusted ;-)

D is already a complex beast (just to learn - nevermind implementing a compiler for it)!

And D++...well... whoohoo....hold on tight!

November 25

On Thursday, 25 November 2021 at 02:14:34 UTC, forkit wrote:

>

And D++...well... whoohoo....hold on tight!

C++, so complicated, but so what? Complexity is left to the compiler.
As a user, I enjoy it. That's right. Therefore,C++ has a large number of users.
D++ can also be used in this way, things are complicated, you have to provide more things. complexity is one of the essence of language. And go now also in the complex road, there is no basic template, which can abstract well, go now, has betrayed his simple concept. It feels like cheating.

November 25

On Thursday, 25 November 2021 at 02:14:34 UTC, forkit wrote:

>

What we really need is a simpler compiler that we can trust (through formal verfication and proofs). That is how one implements @safe and @trusted ;-)

That is an ideal, not a reality.

4 5 6 7 8 9 10 11 12 13 14 15 16