On Monday, 26 November 2018 at 18:01:08 UTC, welkam wrote:
> On Monday, 26 November 2018 at 10:19:14 UTC, Chris wrote:
>> I may well be that Walter and other core devs really feel that they are making great progress when porting DMD to D and stuff like that.
>
> Actually D had postponed porting from C++ to D for a long time. Other languages like nim and jai from beginning had their implementation done in itself. Secondly translating was done with a tool and my guess didnt took much of work for DMD backend. Well compared to what would it take to make android app development easy. Thirdly compiler being in D means more people from community can participate in compiler development. I started twiddling with compiler because it was mostly in D and would not touched it if it was in C++.

Fair enough, but again it's a very narrow focus, and people interested in it can "twiddle around". For deployment we have LDC anyway. A language needs a broader focus, i.e. infrastructure, tools and, yes, an IDE.

>> Indeed, their own projects might be emotionally rewarding and trigger feelings of euphoria.
>
> Its clear you havent read single whitepaper on behavioral psychology or neuroscience.

If you have worked with people (and you know a bit about yourself) you don't need to read all that to know how people feel and what motivates them.

>> A project like D cannot survive if it's only driven by personal preferences.
>
> The good thing is that its not only driven by personal preference. Now we have d foundation and companies who sponsor work on libraries like Symmetry Investments. This trend will only increase but not at the speed you or I want.

Again, often a very narrow focus. Bits and pieces here and there, special interest.

>> In my own job I sometimes work on interesting and emotionally rewarding stuff, but I also have to do the head wrecking and boring stuff that may not even be related to writing code - boring but necessary.
>
> You compare open source with for profit company again. We all do boring but necessary stuff in paid jobs. Your not exception. What matters here is that you expect other people who worked 8h of boring stuff to go home and work on more boring stuff for free. These kind of people are rare and you are not one of them yourself.

Well, I used to contribute a bit. Bug reports, the odd PR for dlang and other simple stuff. I donated money twice (I think, or was it three times?). I was seriously thinking about contributing more, but then

1) I had to spend time fixing my own code due to dmd updates
2) I saw the whole PR / priorities culture (or lack thereof)
3) I thought, if my own code breaks randomly or is no longer "state-of-the-art" due to a new paradigm being introduced, what happens if I make a contribution? Will I have to re-write the code forever and ever and ever like poor Sisyphus?
4) regarding stuff like ARM and JNI, I would have helped to improve it, my own experience / setup would have trickled into the project. However, it took too long until it became a real option, and other technologies emerged in the meantime, technologies that where closing the gap ignored by D.

Yes, I would have worked on boring stuff too. People contribute for various reasons, e.g. self-interest, for the glory, gratitude (give something back) or because they have / share a vision or an idea. Usually it's a mix of all of them. (Except for gratitude, everything else is more or less self-interest).

Now you may discard all of my points above as nonsense, because maybe they are not mentioned in papers about behavioral neuroscience and psychology (I've actually read quite a bit about behavioral biology), but those were my thoughts.

>> It's not just a question of the - by now famous - $1,600 a month. I've seen other open source projects thrive because of a different community culture.
>
> Stop being vague and start naming open source projects that are as big scope as D and thrive without corporate sponsorship. We might learn something.

There's loads, look at Linux distros and libraries that are still used everywhere.

>> it takes ages to get a review / accepted, then why bother?
>
> You would be surprised what a little bit of money can change and Nicholas already doing good work
> https://www.flipcause.com/secure/cause_pdetails/NDUwNTY=
>

Nicholas is a legend. Fair play to him. I wish him luck and hope he'll succeed!

> Oh and about fixing autodecode https://youtu.be/Lo6Q2vB9AAg?t=4044

On Tuesday, 27 November 2018 at 08:52:07 UTC, Chris wrote:
>
> Fair enough, but again it's a very narrow focus, and people interested in it can "twiddle around". For deployment we have LDC anyway. A language needs a broader focus, i.e. infrastructure, tools and, yes, an IDE.

Lets say A, B and C needs to be done but core team can only work on one thing. If they work on A people complain that B and C is not worked on. If they worked on B people complain that A and C is not worked on and if they worked on C people would complain that A and B is not being worked on.

In principle D could be best language for many things and people want their use case to be prioritized. But D doesnt have enough resources to work on all of that. Hack we dont have people working on std.io library, some std libraries are sub-par. If you write your project from scratch and it doesn't depend on external libraries then D is awesome but for everything else its just meh.

And about tools. These kinds of captain obvious advises are not welcomed. First they imply astronomical incompetence of everyone who works on D and second they just waste time. If you asked on forum if better tooling was useful for language growth you will get almost everyone agreeing. Its not that they just agree with words but they agree with actions
https://dlang.org/blog/2018/07/13/funding-code-d/

From my personal experience tooling is noticeably better than last time I tired D.

>
> If you have worked with people (and you know a bit about yourself) you don't need to read all that to know how people feel and what motivates them.

Until you dive deep and find out that the last person you can trust is yourself. I found that most people project their thinking and feeling onto others and call it understanding. If we narrowed scope to what motivates people to work for free then one single reason might be that people work on things that they think are worth wile. Dont know how to better express that. English is not my native tongue.


>> What matters here is that you expect other people who worked 8h of boring stuff to go home and work on more boring stuff for free. These kind of people are rare and you are not one of them yourself.
>
> Well, I used to contribute a bit. Bug reports, the odd PR for dlang and other simple stuff. I donated money twice (I think, or was it three times?). I was seriously thinking about contributing more

So you havent worked on boring stuff after 8h of potential boring stuff. Thats what I tried to convey.

> but then
>
> 1) I had to spend time fixing my own code due to dmd updates

From what I can tell DMD improved on this. I use to hear about new updates breaking peoples code and now most of what I hear is that updates used to break code in the past. My guess unit tests improved this and they came mostly because compiler was converted to D.

>
> Now you may discard all of my points above as nonsense, because maybe they are not mentioned in papers about behavioral neuroscience and psychology (I've actually read quite a bit about behavioral biology), but those were my thoughts.

My intention with that post was not to discredit your points. It was response to Walters question. I have seen many game devs listening too much on surface problems and not looking at underlying core issue. Then they release patch that doesnt make game better. What I want people to take form your posts is that D needs more people working on more stuff. We dont need people pointing stuff that is already know issue. It just waste people`s times. What we need is concrete action plan that is grounded in reality and doesn't have big flaws.
>
> There's loads, look at Linux distros and libraries that are still used everywhere.
>

Like Ubuntu, red hat and suse? Red hat annual revenue is $2.9 billion. That with a B. Now I cant find it but I remember reading that pacman was maintained by single person and some people were angry that for long time pacman didnt check download validity or something. If my examples were not what you thought than stop being vague and be more concrete.

P.s. we might be the only one left talking in this thread. We could talk about hentai and no one would be wiser :D

On Wednesday, 28 November 2018 at 20:09:20 UTC, welkam wrote:
> P.s. we might be the only one left talking in this thread. We could talk about hentai and no one would be wiser :D

I heard that!

:p

On Wednesday, 28 November 2018 at 20:09:20 UTC, welkam wrote:
>
> P.s. we might be the only one left talking in this thread. We could talk about hentai and no one would be wiser :D

Back in the conversation just for this.

On Wednesday, 28 November 2018 at 20:09:20 UTC, welkam wrote:
> On Tuesday, 27 November 2018 at 08:52:07 UTC, Chris wrote:
>>

>
> Lets say A, B and C needs to be done but core team can only work on one thing. If they work on A people complain that B and C is not worked on. If they worked on B people complain that A and C is not worked on and if they worked on C people would complain that A and B is not being worked on.
>
> In principle D could be best language for many things and people want their use case to be prioritized. But D doesnt have enough resources to work on all of that. Hack we dont have people working on std.io library, some std libraries are sub-par. If you write your project from scratch and it doesn't depend on external libraries then D is awesome but for everything else its just meh.
>
> And about tools. These kinds of captain obvious advises are not welcomed. First they imply astronomical incompetence of everyone who works on D and second they just waste time. If you asked on forum if better tooling was useful for language growth you will get almost everyone agreeing. Its not that they just agree with words but they agree with actions
> https://dlang.org/blog/2018/07/13/funding-code-d/
>
> From my personal experience tooling is noticeably better than last time I tired D.

My suggestion is one year of rehab and detox for D (open a DTox fork if you like):

- feature freeze
- fix old bugs
- do a general clean up
- develop sound and stable tooling
- improve integration with existing technologies (Java / GraalVM / Android / iOS)
- release a stable version of D with LTS - i.e. establish a long term contract between users and creators.

The fist three bullet points above are normal in software development. And the rest is common sense.

Unfortunately, D is all about features and "showing C++" and whatnot. It sometimes seems to me like the boys want to play, whatever strikes their fancy at a given moment.


>
> So you havent worked on boring stuff after 8h of potential boring stuff. Thats what I tried to convey.

Cos it wasn't worth it.

>
>> but then
>>
>> 1) I had to spend time fixing my own code due to dmd updates
>
> From what I can tell DMD improved on this. I use to hear about new updates breaking peoples code and now most of what I hear is that updates used to break code in the past. My guess unit tests improved this and they came mostly because compiler was converted to D.

The problem is that you never know when it will hit you (again) as long as the core devs and community are trigger happy when it comes to new features. Your code must go to make room for feature X.


>
> Like Ubuntu, red hat and suse? Red hat annual revenue is $2.9 billion. That with a B. Now I cant find it but I remember reading that pacman was maintained by single person and some people were angry that for long time pacman didnt check download validity or something. If my examples were not what you thought than stop being vague and be more concrete.

Linux: ArchLinux, Manjaro. Especially Manjaro started as a one man project.

SoX: to this day used in audio software all over the world, by professionals.

Especially in the audio world there are many successful projects / libs.

> P.s. we might be the only one left talking in this thread. We could talk about hentai and no one would be wiser :D

We've been found out!