On Saturday, 24 June 2017 at 18:17:42 UTC, Laeeth Isharc wrote:
> what is the barest minimum we need to enable that?

The phobos devs or the dmd packager just have to grab code and docs they are interested in. It's open source; the author has already made it available. Drop it in the default import and lib paths and run.

It doesn't have to be a whole production.


Alternatively, we could beef up the code.dlang.org website to feature the most useful libraries and do some kind of automatic curation and evaluation coupled with nice presentation and search there.

(or both)

I have some ideas for code.dlang.org and emailed them during the google summer of code process last time, and have even considered doing it myself, but I don't really have time for it right now...

But even basic stuff like taking the download count (available on the existing dub rest api) and using it to weigh in some ranking algorithm might help, and showing more info about subpackages and documentation would definitely be nice, and encourage users to add tutorials to the site and get them searchable.

Reading all the above posts on how to use vibe.d for a two line web server was very interesting.

I still think that it's more simple to do it with Go and with D, but I may be wrong, as I'm currently using Go for web development, and not D, for the reasons above.

And I see that not many people on this forum don't think that D could improve its popularity if :

1/ it was promoted as :

   a/ a great scripting language for scripters and learners looking for an easy to learn and to use typescript-like scripting language (which it is)

   b/ a solid general purpose programming language for the software industry, already equipped for web and desktop application development.

2/ existing web-related D code was refactored to provide the few Go-like functions just enough to serve static files, process an URL and send back pages/data to the browser, so we can "natively" use D for web application development.

3/ any existing GUI library was pre-installed (i.e. imports renamed to "std.*") so that we can also "natively" use D for desktop application development.

I may be wrong, but I think that refactoring an existing web library, renaming the imports of an existing GUI library, and making a few "marketing" changes to the D web site to make it more beginner-friendly would not require much man-power, while this may have the potential to make a HUGE impact on how D is actually PERCEIVED.

And I clearly see that many software development experts on this forum don't see any interest in those suggestions, which I perfectly understand, as we can get the job done with dub, and all the required information can be found on the existing website anyway.

But for less-skilled developers and scripters who didn't graduated from the University (like me), these tiny and seemingly insignificant details are often those which will actually make the difference between something they will use, because it seems really easy to learn and use, and something which doesn't seem good for them, because it looks a bit too complicated to learn or use.

For instance, in my case, I must say that I actually wasn't convinced at all by the current website, which is why I have tried several other general purpose scripting/programming languages (Python, Ruby, Node.js, Nim, etc) before finally trying D, when I was actually lacking other options.

Now I'm happy to have made that last effort, as D has now become my favorite programming language, but maybe other people have had the same experience as me, and didn't try it at all eventually...

Am 25.06.2017 um 01:01 schrieb Wulfklaue:
> On Saturday, 24 June 2017 at 22:16:11 UTC, Wulfklaue wrote:
>
>> Attempt 1:
>> Attempt 2:
>> Attempt 3:
>> Attempt 4:
>> Attempt 5:
>> =========
>
> Almost forgot to mention, this was done under Windows ( where the inline
> does not work ). It works under Linux but that is a different issue.

With inline, do you mean the embedded package recipe? Works for me on Windows if you add the colon.

> I may be wrong, but I think that refactoring an existing web library, renaming the imports of an existing GUI library, and making a few "marketing" changes to the D web site to make it more beginner-friendly would not require much man-power, while this may have the potential to make a HUGE impact on how D is actually PERCEIVED.

Python language maintainers have a similar approach, and it pays off...

If you look at the PYPL page (http://pypl.github.io/PYPL.html), you will read this :

"Worldwide, Java is the most popular language, Python grew the most in the last 5 years (8.7%) and PHP lost the most (-5.0%)"

Here is what Dlang's official blog says about Python :

"On the author’s MacBook Pro, the Python version takes 12.6 seconds, the D program takes 3.2 seconds. This makes sense as the D program is compiled to native code. But suppose we run the Python program with PyPy, a just-in-time Python compiler? This gives a result of 2.4 seconds, actually beating the D program, with no changes to the Python code."

(https://dlang.org/blog/2017/05/24/faster-command-line-tools-in-d/)

So actually it can even run faster than D, without any significant optimization...

For those still not convinced, I suggest they compare the popularity of D, Kotlin, Golang and Python on Google trends :

https://trends.google.com/trends/explore?q=dlang,kotlin,golang,python
https://trends.google.com/trends/explore?q=dlang%20language,kotlin%20language,golang%20language,python%20language

The green line kilometers high above the ground this Python, whose I praised the website.

The red line is Kotlin. While available since only a few months, it is already competing with Go's orange line in terms of popularity.

Btw their first example is a "Hello World", in different versions.

And the completely flat blue line below all the three other curves, which desperately looks like the electrocardiogram of a dead heart, it is D's current popularity (or relative unpopularity should I say) in Google searches.

QED...

Anyway, it seems that not many people on this forum see any interest in improving D's popularity on the web with just a few cheap cosmetic changes to better align it with contenders like Python or Go, so I will bothered people with unwanted advices.

Instead, may I suggest that we dig the grave deeper by putting even more complicated examples on D's code roulette so that we make sure that only the elite of programming could remain interested in using this language ?

Whoops, sorry, this is already done at the moment actually. Ok, then I said nothing... LOL

On Sunday, 25 June 2017 at 07:34:57 UTC, Sönke Ludwig wrote:
> With inline, do you mean the embedded package recipe? Works for me on Windows if you add the colon.

Strange, colon was there. Did not work on Windows. Same error as without the colon. This is why i mentioned that it worked under linux ( windows linux subsystem ) but not on pure Windows.

On Monday, 26 June 2017 at 07:19:21 UTC, Wulfklaue wrote:
> On Sunday, 25 June 2017 at 07:34:57 UTC, Sönke Ludwig wrote:
>> With inline, do you mean the embedded package recipe? Works for me on Windows if you add the colon.
>
> Strange, colon was there. Did not work on Windows. Same error as without the colon. This is why i mentioned that it worked under linux ( windows linux subsystem ) but not on pure Windows.

LOL

QED... ;)

On Monday, 26 June 2017 at 05:10:34 UTC, Ecstatic Coder wrote:
> Anyway, it seems that not many people on this forum see any interest in improving D's popularity on the web with just a few cheap cosmetic changes to better align it with contenders like Python or Go, so I will bothered people with unwanted advices.

Remove Python from your link:

https://trends.google.com/trends/explore?q=dlang,kotlin,golang

D does not even register beyond 1, even without python disrupting the whole scale line.

> Whoops, sorry, this is already done at the moment actually. Ok, then I said nothing... LOL

I know how you feel but trust me that trying to fight this seems to be a fight you will not win.

If there is something very clear about the D community, its comes mostly down to this:

* If you want something done, do it yourself
* If you want something done, pay for it
* If its a massive obvious flaw, maybe somebody will fix it

D with some help can become a much more general accepted programming language that is not just a C++ replacement language. The D language as i stated before is not the problem. Its mostly what is around the language, is where D falls short.

But fixing those issues tend to be:

* Repetitive work, nobody like doing that
* None glorious work.

In other words, not a big project where somebody can put there own stamp upon it. This seems to be one of the major issues. A lot of brilliant people but very few that want to do the grunt work.

Frankly, there are moments that i wanted to do some of the grunt work but the same issue as everybody else. No time to spend hours and hours into fixing things ( and some of the issues that needs to be fixed can be weeks or months of work unfortunately ).

This is why i personally think the only way to "solve" these issues, is if D hires people that work full time on these issues. And that in return needs money.

And then we come back to the same sticky point. Why are very few people donating? Maybe its just from my perspective but when i donate, i want to feel that my money will be used good. And because there is no real public accounting on where the money that D gets is being spend, ... maybe there is but again, it feels like you need to dig to find it, like with most D things.

---

Saturday i spend some time with Kotlin. And frankly, i liked it more then i expected.

Install jetbrains community edition, install JVM. And your ready to rock.

* Front end writing / Javascript generation. Works out of the box.

* Java generation / writing. Fast, works perfectly.

* Great editor support! I just love the whole "rust" style mutable recognition build in. The extra type hinting. The ability to alter the coding formatting. Etc ... and all in the familiar editor then i am used to from my php days.

* Kotlin Native (llvm). A bit more work to get going but its only at 0.3 version so one gives it some slack. But they did impressive work in the last two months from the initial release. Hell, they even beat Swift by offering Windows support already *lol*

---

While i am sure there are javascript converting modules for D. Again ... no focus on it, 3th party, no clear information etc ...

When it comes down to editor support, sorry its mediocre at best. Here are work i am fighting the workspace-d dfmt simply because i have some long shell lines that i want on a single line. They are > 120 characters but dfmt keeps breaking my line formatting. Frustrating because it looks bad and because dfmt is compiled into workspace-d, no documentation how to override dfmt. Again, it feel like a lot of modules slapped together and they work but no steps beyond this.

...

I am at work so i do not have a lot of time to write all this. D simply feels incomplete compared to its competitors. Not the language, that is great and really feature complete. But everything around it feels fragmented, incomplete, in a constant state of unfinished. That is just my opinion.

Am 26.06.2017 um 09:19 schrieb Wulfklaue:
> On Sunday, 25 June 2017 at 07:34:57 UTC, Sönke Ludwig wrote:
>> With inline, do you mean the embedded package recipe? Works for me on
>> Windows if you add the colon.
>
> Strange, colon was there. Did not work on Windows. Same error as without
> the colon. This is why i mentioned that it worked under linux ( windows
> linux subsystem ) but not on pure Windows.

Can you post the error message? I didn't use the Linux sub system.

On Monday, 26 June 2017 at 08:29:54 UTC, Sönke Ludwig wrote:
> Am 26.06.2017 um 09:19 schrieb Wulfklaue:
> Can you post the error message? I didn't use the Linux sub system.

From what i remember on my home system, the issue was it tried to store the file in Windows temporary directory.

Just tried it on my my work PC ( with visual studio code portable ) and got a slightly different error:

> C:\Users\xxx\AppData\Local\Temp\VSCodePortableTemp\.dub\build\vibe-d:utils-0.7.31: The directory name is invalid.

I assume the error is similar to the one i had in the weekend because its trying to use the temp folder again. Just different temp path because of the portable editor ;). Note: At home its a full VSC installation, not this portable version.

Thanks for having taken the time to write this very enlighting post.

I think this is a very pragmatic analysis of the main roots of the problem, which clearly explains why D remains much less popular than Go, Python, Kotlin, etc despite the language itself is at least as modern and convenient (and even more IMHO).

I don't think the situation will change any time soon, so I feel exactly in the same mood as you.

Hence my suggession to simply make a few "cosmetic" changes to change D's perception on the web, to compensate for the lack of money and manpower.

But they won't probably be applied.

And I know that for the "do it yourself" I'm probably part of the problem...