Feedback on Átila's Vision for D

Ola Fosheim Grøstad ola.fosheim.grostad at gmail.com
Thu Oct 17 19:00:47 UTC 2019


On Thursday, 17 October 2019 at 17:23:29 UTC, welkam wrote:
> to tell me how I should spend my free time. People do not react 
> well when they are treated almost like slaves by a master.

Sure, but it isn't really about resources. It is about process. 
If you after version 1.0 have one experimental branch and don't 
let features into the stable branch until they have been  
harnessed and used heavily then you avoid many issues. D has 
moved more like an experiment than an engineered artefact.  From 
that there will be more noise...


> time. Second I was talking about how other people coming here 
> go "why the F#$k its so negative here.

When people slam the door when they are leaving it is because 
they've hit roadblocks that they think should not have been let 
into any stable branch... most likely.  They expected something 
carefully engineered and felt they walked into something that 
wasn't.


> Could you elaborate. What Manu usually wants is quick fixes 
> fast and that is the wrong way of building a programming 
> language. Before any change is committed to the language it 
> should be thoroughly studies otherwise you will get a mess.

I'm not here to evaluate anyone, but I am totally convinced that 
D would not have started moving in the right direction without 
constant pressure from people like Manu, that's all I have to 
say. If something ought to have happened five years ago then it 
is reasonable to push harder. Maybe it works?

If there is no push, nothing happens in the direction of real 
time programming.  That's been obvious for at least a decade.


> In D's  case 99% of all problems are due to lack of resources so

Not resources. Process... Expanding the feature set... Rather 
than narrowing down the focus.

Like, there is no point in planning for a borrow checker if you 
don't have people with the right know how and have plenty on the 
plate.  It's a great idea, but probably too late.  That's not a 
dismissive statement, that's just realism.

So why was it put on the table?


> If dichotomy isn't false then answer weather D is high or low 
> level language? The dichotomy exist for other languages because 
> of their design limitations not because of some lows of nature.

Low level.  With some high level runtime features.  There is very 
little room for high level optimization in D.

> All languages that became popular did so because of some big 
> backer. It would be mistake to conclude that language X became 
> popular because of feature Y so we should also have feature Y 
> and we too become popular.

As I stated elsewhere in this thread, languages become popular 
because they have a use case where they are superior (even if the 
language isn't). You don't need a big backer (and you don't need 
a great language). But you need some adoption before somebody 
implements something that enables that "superior use case", and 
probably also the right set of features to make it easy to do so.

So the only way you can plan for traction is to understand what 
use case you want to be superior for and design the language with 
that in mind.

Otherwise popularity becomes a random event (someone accidentally 
chose language X to build a great framework).


> D is still on exponential growth curve.

I wouldn't know, and I don't think it matters all that much.  I 
was talking about the language/compiler.

The C++ people have realized that Rust is coming to take a slice 
of their cake and they appear to be pushing new features out the 
door at a high pace.  I don't think that would have happened 
without Rust.  I think it was a wakeup call.



More information about the Digitalmars-d mailing list