Is D really community driven?

Ilya Minkov minkov at in.tum.de
Sat Jul 28 17:58:08 PDT 2007


Hello.

Bruno Medeiros wrote:
> D, the language itself at least, is Inner Circle driven. (I say this
> facetiously but seriously)
> Meaning that in the great majority of situations Walter only considers
> the opinion of those of the Inner Circle, even when there is large
> community consensus.

This is certainly NOT TRUE.

I was here back when this newsgroup did not even exist, and the volume of
communication was not that large. One could thoroughly read each and every
post within a reasonable amount of time. In fact, one could even read all
of the history, all of the communication from day 1, this was quite a
bearable task. Something i did back then.

Back then, Walter read everything, and took his time to think about each and
every proposition. One would always get a clear answer, yes, no, and why
exactly no. Thus, if a proposition came which would disrupt the language in
some manner, one would know it pretty quick and the discussions went down.

What changed? D gained lots of public attention. Walter still managed to
read most of the newsgroup, but he started putting off thinking about
propositions. He simply didn't have enough time. Still, it has always been
possible to convince Walter.

What happened later? The volume of communication grew and grew, and i did in
fact quit because i couldn't follow. I frankly don't know how Walter
manages to get through all that, but i believe he certainly cannot give the
same attention to every person. People who have been able to articulate
their opinion properly are likely to get more attention, because it always
means a lot of work to Walter

So, if ever possible. when there is a request, following information has to
be supplied:

- specific usage cases. what gets better? also: what gets worse? As
complete, as detailed as only possible. This is a lot of work to supply. It
may seem obvious, but it has been shown that this needs to be done anyway.
- interaction of the change with ALL possible features of the language. What
improves, what breaks? What features seem like might be affected, but in
fact aren't?
- comparison with other languages. It must include not only the feature
itself, but also the features that might be affected. This requieres
reference of other sources - e.g. circumventions of some problem in
libraries for other languages would be a pretty heavy hint that this also
needs to be adressed in D.

Every piece of information on this list either has to be supplied
explicitly, or Walter has to work it out himself, and in fact he usually
has to. Never forget this!

Someone with a lot of D experience is obviously better in position to supply
use cases. The questions on interaction requiere experience AND in-depth
understanding for both D and a number of other languages.

> I think the cases where Walter has actually changed the language upon
> the opinions of the D community (aka the Outer Circle? :P ), was, in a
> rough sense, when there was no parallel to C++ whatsoever, and thus
> there was no influence or opinion from the Inner Circle or the C++
> community.

Convincing Walter on core issues IS possible. It has been done. However,
even the most respected and knowledgeable members of community have to put
a great effort into that.

However, all the core issues have been discussed to death, 2 or 3 years ago
or even more. With most things, we have tossed around every possibilty and
decided that it just wouldn't work at some point. Nontheless, the same
requests pop up again and again, perhaps in some other form, and there even
seems to be some "consensus" among those who request. But it just wouldn't
work.

Perhaps extra work is needed on documenting such things. I am not going to
search through the newsgroup archives, but if you care to put together some
of the more important requests, i might want to put up a wiki page with a
thorough discussion.

I must say, i'm glad that this is a Designer language, not a
community-driven language -- it is surprisingly both consistent and
powerful for a language that is based upon C. Nontheless, i must say,
Walter is completely free of arrogance. He has been fair to EVERYONE,
within his possibilities.

-eye/photoAllergics

(uh, sorry, my english is getting rusty. :)



More information about the Digitalmars-d mailing list