Phobos incubator project ?

simendsjo simen.endsjo at pandavre.com
Sun Aug 15 16:55:12 PDT 2010


On 15.08.2010 23:24, BLS wrote:
> But somehow I am not able to figure out what your opinion is regarding
> an incubator project :)

Well.. I think it's a good idea, but I don't really think I have an 
opinion. I have no idea how patches are dealt with, and I'm new to D, 
phobos.

If you're really, really bored, read on for a long rant:

(serious rant ahead..)
..
..
..
(still bored?...)



If patches isn't attended to it's better to have a a "scrapple" project 
like tango.
Just require the boost, or compatible, license and let the phobos core 
developers pick and refactor as they like.

I haven't gotten the idea it's a PhD mentality now, people get's picked 
up as core developers from good contributions (from bugzilla?, but maybe 
they're all PhD's :) ).

Can't you just set up a project at dsource (phobos_incubus or 
something), put up a lot of code you've done, and announce it? I'm sure 
a lot will follow.

I'm "new" to D (have coded _one_ nontrivial project in d1 between 07-09, 
but no d2 experience) and from what I see, several people have a lot of 
_almost_ the same templates in their libraries (mostly type/traits stuff).
Nobody has taken the time to go through these, generalized them, and put 
them to phobos.

Hopefully, with an initiative like this, people with put a lot of 
"almost" generalized code from their libraries into the public so 
std.traits++ can grow.

It's easy to get into the "give me what I want" mode, but I don't pay 
any money, and as far as I know, there are nobody's actually making any 
money of D.
No companies are supporting D either financially or with success stories 
(outside digitalmars that is a _one_ man army(!) ).
There are very few contractors - and none of these are making much money 
of D.
D still have very few contributors, some maintaining the gnu frontend 
(d1), some lcc (d1) and walter+co(d1/d2).
I still dream of the day when D has a D frontend, backend and is 
implemented as SaaS so every IDE has good support for it :)

But... The fact that d2 is not stable yet (despite of the fact that 
Andreis book should end this) is a huge problem. When will actually D be 
stable??? I used it a couple of years ago, AFTER it has been made 
stable, and it was a really pain in the ass!... I'm actually for 
breaking compability between versions in an evolving language, but in 
the case on D1 I had to update _ALL_ libraries I was working on... With 
_every_ compiler release!

But; disregard the last paragraph! Things have changes, and this might 
not be any problem anymore! I've just started testing D2, and it seems 
like a really

Then again... Many people have been spoiled by Eclipse, Visual Studio 
and their kind, so it's difficult to switch to "notepad" mode.
I myself cant say how much I love C#, VS and CodeRush... I know.. 
Blasphemy :)

And for the record.. I implemented a "larger" (30k incl. 
comment/whitespace/scope) program in D1, but it went to hell! I had 
really high hopes, but in the business world, there are no DB drivers, 
there are no disconnected DataSet's, poor wsdl support etc. etc. etc.

As a side note: dsource is a great idea, but it has detoriated 
(rusted... I don't know how to spell it :) ) so it's pretty useless as a 
lookup resource for libraries.
Probably 50-70% of the projects is abandoned (or is written for an 
ancient dmd version).
Two new columns and the site would have been ten times as useful;
"last time updated" and "d version(s)".

And as another side note...: If you had a contribute button on you're 
site and that was distributed among those involved in D and phobos... 
Would it help the progress? As I see now, there are large bugfix 
releases every month and large changes to phobos too, so I wont complain..

So... How can the little man help?


More information about the Digitalmars-d mailing list