Maybe D is right about GC after all !

Dan Partelly i at i.com
Sat Dec 23 08:09:19 UTC 2017


On Friday, 22 December 2017 at 17:21:12 UTC, Joakim wrote:
> You seem not to have read the original link and Walter's first 
> post: why would we expend all that effort on a small, shrinking 
> niche of the language market that is populated by a couple 
> thousand diehard programmers who won't look at any new language?


  Empiricall Id say that a lot of those programmers are interested 
in both old and new languages, from Lisp to D. But nevermind 
that. Those people are not idiots really. They would use any 
language which gets the job done, and they would use as high 
level as possible to get the job done. They look and then they go 
back from practical reasons. And a 10% market of C programmers is 
not really small niche. D could have got a lot of those, if it 
made just one more good decision regarding the "how" of GC and 
libs. The IOT is upon us, automation is upon us, embeded devices 
slowly become ubiquitous in cosumer market, while a lot of 
languages shoot themselves in the foot from beeing a part of this 
market with poor decisions.


>  Given that GC is already here and addresses a much bigger 
> market, there's no point in Osborning all that's available 
> _today_ by forever dreaming for a fully no-GC stdlib that's not 
> going to be here anytime soon.

Sure. This is why I said I lament the choiche done. Today the 
vast std lib of D should have been not GC dependent. So IMO, D 
made a good decision having GC (they did a lot of great decsions 
btw, D is phenomenal good IMO) and D made a very bad decision 
having a std lib dependant on GC, convenience language features 
dependant by gc (vectors , hashes and the like belong to the 
library ), and core language features dependant on GC (classes / 
exceptions whatever)

So there you have it. Regarding GC, in my opinion D made both a 
good decision and a bad one.


More information about the Digitalmars-d mailing list