June 25, 2022

On Saturday, 25 June 2022 at 07:37:51 UTC, Ola Fosheim Grøstad wrote:

>

Indeed, but you should pick a target group. If there are 5 different use scenarios for D and 5% is missing for each scenario then you have a 95% solution for each, but maybe a 75% solution in total.

Yes. So, finally, the question is "positioning". Who is your "target user"?
What kind of people are you going to attract to join 'd'?
I have been proposing to attract talents in c++.
Because 'c++' talents care about 'speed/memory' most, and they are also willing to write libraries. This is an excellent group! Why not attract them?
Listen carefully to their opinions. What they are not satisfied with is the direction of D's efforts!
Through them to help 'd' writing the library and increasing the peripheral ecology of 'd'!
Our opponent is 'rust' , so we must compete with with rust in who attract most C++ talents!
This is where D should go.Talents is the most important! Meeting their needs is more important than meeting the needs of ordinary users!

June 25, 2022

On Saturday, 25 June 2022 at 07:37:51 UTC, Ola Fosheim Grøstad wrote:

>

So it is better to tell 4 out of 5 NO and give special treatment to 1/5.

Just pick the most promising scenario.

This is also very important. One '100%' is better than three '95%'!
Now, since this is the case, we should list all the existing problems and sort them according to the important procedures. The most important problems should be solved first, and then the second important ones. One by one.
Then d users can boldly boast all over the world
If the foundation is not qualified, marketing will only be counterproductive!

June 25, 2022
On Friday, 24 June 2022 at 06:31:37 UTC, rikki cattermole wrote:
>
> One thing that we could do, is pay someone about once a year to make an album of music (royalty free so it can be used for live streamers!) to commemorate the year released around DConf. Somebody like Popskyy.
>
> While I don't think any other programming language does this, the OpenBSD community does[0].
>
> [0] https://www.openbsd.org/lyrics.html

You know, I used OpenBSD extensively, for more than a decade.

I cannot recall ever paying attention to their marketing ploys.

A secure operating system, by design, was sufficient for me.
(although, I readily acknowledge, not everyone considers it to be 'secure').

btw, back in the 90s', when I was learning C++, my friend and I walked into the software store one day (yeah, no online back then), and the product we both decided on, was the product with the biggest box (i.e. the most books).

At that particular time, it was Borland C++ that had the biggest box, with the most books. So we settled on that.

Of course I never read them all.

If we were in the 90's still, I'd say that maybe D needs a bigger box.

But other than a bigger box, I'm not so sure... perhaps a product that delivered on what it claims it is, would probably do just fine ;-)

June 25, 2022

On Saturday, 25 June 2022 at 08:13:56 UTC, zjh wrote:

>

This is also very important. One '100%' is better than three '95%'!

D must establish a list of 'excellent articles'. Even if people don't use d, they can access these 'excellent articles' to learn!
Here are two Chinese versions of comparison to rust: 1And2
here, this is the Chinese versionList of posts for 'd beginners'.
See the link at the beginning of the article for the English version.
It would be great if could provide a 'link' to it somewhere in the 'd' official website.

June 25, 2022
On Friday, 24 June 2022 at 06:31:37 UTC, rikki cattermole wrote:
>
> One thing that we could do, is pay someone about once a year to make an album of music (royalty free so it can be used for live streamers!) to commemorate the year released around DConf. Somebody like Popskyy.
>
> While I don't think any other programming language does this, the OpenBSD community does[0].
>
> [0] https://www.openbsd.org/lyrics.html

You know, I used OpenBSD extensively, for more than a decade.

I cannot recall ever paying attention to their marketing ploys.

Actually being a secure operating system, as it claimed, was sufficient for me.
(although, I readily acknowledge, not everyone considers it to be 'secure').

btw, back in the 90s', when I was learning C++, my friend and I walked into the software store one day (yeah, no online back then), and the product we both decided on, was the product with the biggest box (i.e. the most books).

At that particular time, it was Borland C++ that had the biggest box, with the most books. So we settled on that.

Of course I never read them all.

If we were in the 90's still, I'd say that maybe D needs a bigger box.

But other than a bigger box, I'm not so sure... perhaps a product that delivered on what it claims it is, would probably do just fine ;-)


June 25, 2022
On Saturday, 25 June 2022 at 03:54:03 UTC, forkit wrote:
> At least it all resulted in something practical. That is, something you can test for yourself write now, in your code, if you choose to.

We can both agree that the "ends" was something relatively good, and we clearly disagree on the "means" by which it was partly achieved; regardless, we can move on.

I hope the experimental feature flag is positive, and perhaps it ends up in the language proper.

Jordan
June 25, 2022
On Friday, 24 June 2022 at 19:51:29 UTC, Walter Bright wrote:
> On 6/23/2022 10:30 PM, forkit wrote:
>> Just wear something under it. It'll make it a little less uncomfortable.
>
> What I mean by wearing a hairshirt is insistence on negging D. There are many ways to be honest about D's capabilities without describing them in negative terms.
>
> For example:
>
> Hairshirt - D's gc is terrible compared with Java's.
>
> Honesty - Since Java is heavily dependent on the gc, it is worth it to insert write gates into the generated code to accurately track usage of gc allocated memory. Usage of D's gc is optional, and D's semantics are much less reliant on the gc. Hence, in the interest of maximum performance for most code, D does not insert write gates in the generated code.

I don't think D's semantics are much less reliant on the GC. Have you tried writing CTFE without the GC?

D has a worse GC than java, that's a fact. You can do manual memory management in D, even though it isn't really supported by its own feature set. Using something like C++ and Java together is the better option as it provides the best of both worlds. Like what Android did. D went a direction that provides a worse GC without providing the tools needed for the alternative it claims to support.


June 25, 2022
On Saturday, 25 June 2022 at 01:24:43 UTC, Timon Gehr wrote:
> On 25.06.22 00:02, Dukc wrote:
>> Anyone who wants D to become popular
>
> D is already popular.

Yes, good correction. Should be "more popular than already" or something like that.
June 25, 2022

On Saturday, 25 June 2022 at 12:56:37 UTC, mee6 wrote:

>

D has a worse GC than java, that's a fact. You can do manual memory management in D, even though it isn't really supported by its own feature set. Using something like C++ and Java together is the better option as it provides the best of both worlds. Like what Android did. D went a direction that provides a worse GC without providing the tools needed for the alternative it claims to support.

IIRC Atila Neves has expressed some positivity towards having some kind of actor model with local GC per actor in the forums. That in combination with ARC for shared objects could be an interesting direction.

Applications have to be given control over scheduling though, and actors should be written to have short life times so that you can bypass garbage collection as long as you have low memory usage. (The GC heap then becomes an area-collector in practice, so everything is freed wholesale when the actor dies.).

LLVM also makes it possible to do 100% precise collection, but the current blocker is the C union. It would also help to ban destructors on GC-allocated objects.

So the main issue isn't that there are no options, but that there are no signs of future movement towards something better beyond those very sparse positive bleeps from Atila.

June 25, 2022
On 6/25/2022 1:05 AM, forkit wrote:
> At that particular time, it was Borland C++ that had the biggest box, with the most books. So we settled on that.


Borland had stellar marketing. They were adept at branding mundane features as incredible new technology.