October 29

On Sunday, 29 October 2023 at 16:46:02 UTC, Imperatorn wrote:

>

On Sunday, 29 October 2023 at 16:39:28 UTC, monkyyy wrote:

>

On Sunday, 29 October 2023 at 12:54:33 UTC, Imperatorn wrote:

>

The plan is between 50 thousand (realistic) to 100 thousand (optimistic) devices each year.

However, we are becoming unsure if D is really an option for us

I'd probably consider hiring compiler devs who are doing something interesting but grumpy with the management and making a d subset compiler that you control if you actually of a size large enough to justify it

Maybe, but it would be best if we didn't have to and could stay mainline D as long as possible. I want to hear other peoples experiences with this for comparison.

In my experience, only 1 bug report of mine has ever been addressed; and the drama this week was patches written 6 years ago.

>

Can we count on that if we find an issue with D that it will be taken care of?

so the answer is just no; you should not expect d to change meaningfully ever. If you want a foundation to base a lot of people's livelihood on you should day 1 plan to build that foundation in parallel with whatever product you making so you have the talent on hand and meaningful control over the outcomes.

October 29

On Sunday, 29 October 2023 at 12:54:33 UTC, Imperatorn wrote:

>

Is the recommended course of action to fork dmd, druntime and phobos and have a completely parallel version of D, a bit like

That is the only future of D unfortunately, that someone with a little vision of the future continues with D.

Don't wait for something to happen with the D foundation because you will probably retire before that.

There are other languages but not as nice as D in this particular niche which is really sad.

October 29

On Sunday, 29 October 2023 at 17:20:36 UTC, IGotD- wrote:

>

On Sunday, 29 October 2023 at 12:54:33 UTC, Imperatorn wrote:

>

Is the recommended course of action to fork dmd, druntime and phobos and have a completely parallel version of D, a bit like

That is the only future of D unfortunately, that someone with a little vision of the future continues with D.

Don't wait for something to happen with the D foundation because you will probably retire before that.

There are other languages but not as nice as D in this particular niche which is really sad.

Choosing a programming ecosystem for a business is a bit like a marriage, and if peoples' livelihoods depend on it then the decision should be taken seriously and objectively, looking at it from many angles. Having fun programming in D is important but there are many other factors.

Even businesses the size of Meta will think twice before choosing a language if if it means having to fork the language and maintain it. When smaller businesses get sucked into an area which is not in their core competence then it often leads to major problems and downstream regrets after the initial shine wears off.

D is great but a business decision to use it should be based on sound principles, not fanboi-isms. For me, if you need to fork D to use it for a business, that's a red flag.

October 29

On Sunday, 29 October 2023 at 17:20:36 UTC, IGotD- wrote:

>

On Sunday, 29 October 2023 at 12:54:33 UTC, Imperatorn wrote:

>

Is the recommended course of action to fork dmd, druntime and phobos and have a completely parallel version of D, a bit like

That is the only future of D unfortunately, that someone with a little vision of the future continues with D.

Don't wait for something to happen with the D foundation because you will probably retire before that.

There are other languages but not as nice as D in this particular niche which is really sad.

Can you explain your standpoint? Why do you think so?

Thanks

October 29

On Sunday, 29 October 2023 at 17:05:10 UTC, monkyyy wrote:

>

On Sunday, 29 October 2023 at 16:46:02 UTC, Imperatorn wrote:

>

[...]

In my experience, only 1 bug report of mine has ever been addressed; and the drama this week was patches written 6 years ago.

>

[...]

so the answer is just no; you should not expect d to change meaningfully ever. If you want a foundation to base a lot of people's livelihood on you should day 1 plan to build that foundation in parallel with whatever product you making so you have the talent on hand and meaningful control over the outcomes.

Hmm. I will take your points into consideration. Thanks

I'm making a list of pros and cons.

October 29

On Sunday, 29 October 2023 at 17:29:51 UTC, Abdulhaq wrote:

>

D is great but a business decision to use it should be based on sound principles, not fanboi-isms. For me, if you need to fork D to use it for a business, that's a red flag.

Why? I would rather say that forking a programming language in order to tailor it for ones need is a serious commitment.

October 29

On Sunday, 29 October 2023 at 18:00:13 UTC, IGotD- wrote:

>

On Sunday, 29 October 2023 at 17:29:51 UTC, Abdulhaq wrote:

>

D is great but a business decision to use it should be based on sound principles, not fanboi-isms. For me, if you need to fork D to use it for a business, that's a red flag.

Why? I would rather say that forking a programming language in order to tailor it for ones need is a serious commitment.

Yes that's a fair comment. What I mean is that for most businesses, the choice between D and e.g. C++ is best understood in terms of risk /reward.

C++ is the low risk option, for a typical business, because it will still be around in 20 years, it will be well maintained, you will always be able to find a pool of developers who can maintain your code, it will link and compile with thousands of industry standard libraries, frameworks, protocols, hardware platforms.

So D is higher risk in that you may find yourself having to spend time (and much money) coding up your own libraries and hardware support. In 10 years time you might struggle to find D developers. However, the reward with D is that you can achieve the required functionality, over the next few years, much more quickly (i.e. more cheaply) than with C++.

So you choose D if you want to code up a well understood set of algorithms/functionalities on known platforms and you want to do it quickly/cheaply, and you know that D can provide all that you need.

You choose C++ for long term support, a huge breadth of platforms, some of which are future platforms you don't yet know about, and general maintenance ability over the years.

To finally answer your question, maintaining a compiler/ecosystem is a significant cost overhead the reduces the rewards of D to such an extent that it is normally wiser to choose C++.

October 29
On 10/29/2023 5:54 AM, Imperatorn wrote:
> However, we are becoming unsure if D is really an option for us given the response we got trying to making almost the smallest change imaginable to phobos (changing a single word).

What is that change?

October 29

On Sunday, 29 October 2023 at 18:12:46 UTC, Abdulhaq wrote:

>

C++ is the low risk option, for a typical business

That seems dubious, theres also talk of c++ dying, c++ has super long compile times and maybe you use a feature that steps onto a landmine of those hours compiling. And as hard as making a compiler with your subset of d is, making a compiler for c++ seems all that much worse.

October 30
I'm pretty certain it was: https://github.com/dlang/phobos/pull/8832