Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

OpenD, a D language fork that is open to your contributions

dpldocs.info

Post reply on HN
Print view View original

91–100 of 322 posts

Re: OpenD, a D language fork that is open to your contributions

#91

Post by WalterBright »

dataflow wrote: post #50
TillE wrote: post #9

Earlier quoted context omitted.

I think the fundamental problem with D is much simpler than that, it's the technical context. D is a better C++, but it's not better enough to escape the gravitational pull of C++. Rust is sufficiently compelling, with its focus on memory safety.

I keep saying this and having the exact same argument with everyone every single time I mention it (and I'm sure here we'll go again), but D's fundamental problem was its GC. Its mere presence colors the code that uses it, and once your dependencies use the GC, you have no recourse. It's impossible to abstract away. I fail to see how such a language could ever substitute for C++.

GC makes CTFE (Compile Time Function Execution) simple and very easy. It's quite normal for projects to use GC for CTFE and other management methods for runtime.

Re: OpenD, a D language fork that is open to your contributions

#92

Post by teleforce »

countWSS wrote: post #84

My suggestions: completely get rid of GC(use refcounts or borrowed pointers), add macros(see FreeBasic extensions of C macros), migrate everything complex/optional out of stdlib to specific packages(Rust ecosystem is centered on packages not stdlib), and you will have a solid competitor to C++ and perhaps Rust. Rust of course has better macros and type system, but D looks simpler and more approachable for rapid proto…

Your suggestions are moving backward, and GC by default and no macro made D intuitive and Pythonic that are big plus in any modern programming language construct. The non GC is not really needed unless you're working on OS control primitives but again D give you alternative unlike Go. Every modern languages should avoid macro like a plaque otherwise you will sooner or later create a ghetto inside your community not unlike Ruby on Rails. C++ and Rust are complex languages and it's evident by their long compilation time, while Go and D are much simpler hence faster compilation time. D get many things right, and C++ and later languages like Nim have been copying D uniques features left and right since its introduction.

Re: OpenD, a D language fork that is open to your contributions

#93

Post by countWSS »

teleforce wrote: post #92
countWSS wrote: post #84

My suggestions: completely get rid of GC(use refcounts or borrowed pointers), add macros(see FreeBasic extensions of C macros), migrate everything complex/optional out of stdlib to specific packages(Rust ecosystem is centered on packages not stdlib), and you will have a solid competitor to C++ and perhaps Rust. Rust of course has better macros and type system, but D looks simpler and more approachable for rapid proto…

Your suggestions are moving backward, and GC by default and no macro made D intuitive and Pythonic that are big plus in any modern programming language construct. The non GC is not really needed unless you're working on OS control primitives but again D give you alternative unlike Go. Every modern languages should avoid macro like a plaque otherwise you will sooner or later create a ghetto inside your community not u…

Then, D will compete with GC languages that outclass it in ecosystem diversity and metaprogramming.(i.e. it will remain same niche language it was for decades). Having a huge runtime with GC doesn't seem appealing or efficient.

Re: OpenD, a D language fork that is open to your contributions

#94

Post by fsckboy »

rerdavies wrote: post #48
pstuart wrote: post #21

Earlier quoted context omitted.

Let's call it E.

E, F, G, and H are already taken. It's hard to say whether I is available, or un-googleable. J through W are also taken. X++ is taken, but X seems to be available. Y and Z are taken. A would be an excellent choice were it not completely un-googleable. Finally, B is a famous precursor to C.

>A would be an excellent choice were it not completely un-googleable.

APL is named after the book A Programming Language

Re: OpenD, a D language fork that is open to your contributions

#95

Post by crq-yml »

Kranar wrote: post #7

D is such a sad case of how a good technical product can fail due to poor leadership and decision making. A language is more than just a compiler, it's also an ecosystem and a community. D's community is actually very welcoming but man, watch some of the DConf videos and Q&As on Youtube and it's cringe seeing key leaders talking down to people, or dismissing people's concerns, or just have this ego about themselves l…

D has some great developers working on it, past and present.

The thing that moved me away from it was a lack of purpose, which isn't exactly how I would have put it at the time, but in hindsight seems obvious.

The way in which I'm certain that this is true is that if you had the leads try to create a Venn diagram of what the project does and what is central to it vs periphery, they would fail to agree. And if they fail to do that, they will fail to coordinate on the features. And that creates the stagnation and difficulty with contributions.

As well, the reason why that failure leads to a retreat into technical digression - which I know happens out of personal experience - is because that deflects questions by adding scope. It promises an escape from the confrontation: "if I just accumulate more work, it will all come together someday". The problem is that adding scope doesn't solve the contradictions that made you retreat in the first place, it can actually deepen them by creating sunk costs. It's a bad adaptation to the challenge.

If we compare it with, for example, the Niklaus Wirth approach, he would have the Venn diagram completely clear in his mind before committing to any real implementation. And therefore the language would do exactly what he wanted it to do, and it wouldn't need to be scoped into a 20-year project. Thus, he made many languages during his life, mostly of a similar flavor, but with a clear intent of adding something specific that he hadn't covered before.

Re: OpenD, a D language fork that is open to your contributions

#97

Post by ithkuil »

fsckboy wrote: post #94
rerdavies wrote: post #48

Earlier quoted context omitted.

E, F, G, and H are already taken. It's hard to say whether I is available, or un-googleable. J through W are also taken. X++ is taken, but X seems to be available. Y and Z are taken. A would be an excellent choice were it not completely un-googleable. Finally, B is a famous precursor to C.

> A would be an excellent choice were it not completely un-googleable. APL is named after the book A Programming Language

The ultimate language will be TPL (the programming language)

Re: OpenD, a D language fork that is open to your contributions

#99

Post by ziotom78 »

foresto wrote: post #90
Kranar wrote: post #7

D is such a sad case of how a good technical product can fail due to poor leadership and decision making. A language is more than just a compiler, it's also an ecosystem and a community. D's community is actually very welcoming but man, watch some of the DConf videos and Q&As on Youtube and it's cringe seeing key leaders talking down to people, or dismissing people's concerns, or just have this ego about themselves l…

If you're put off by key leaders talking down to people, I think you'll want to spend some time in Nim's forum and issue tracker before considering it as an alternative. (YMMV, of course, but I was put off by what I found.)

Agreed. Nim could be perfect, were it not for its BDFL.

Re: OpenD, a D language fork that is open to your contributions

#100

Post by blarg1 »

rerdavies wrote: post #48
pstuart wrote: post #21

Earlier quoted context omitted.

Let's call it E.

E, F, G, and H are already taken. It's hard to say whether I is available, or un-googleable. J through W are also taken. X++ is taken, but X seems to be available. Y and Z are taken. A would be an excellent choice were it not completely un-googleable. Finally, B is a famous precursor to C.

D++
Post reply on HN