Worst Phobos documentation evar!

[ketmar via Digitalmars-d]

On Tue, 30 Dec 2014 15:39:18 -0800
Andrei Alexandrescu via Digitalmars-d <digitalmars-d at puremagic.com>
wrote:

> I'm choosing to reply to this because it is a perfect example of a 
> pattern of cognitive dissonance that I've witnessed before in this 
> forum. The irony is so astonishing, you were entirely right to frame it 
> humorously - although it's not funny because it's entirely true: there's 
> no time to get stuff done - it's all sucked by hand-wringing over it.
it's not about time, it's about motivation. the real thing is that i
have almost no motivation to make D better for others, and D is
already good enough for me. i'm writing all my pet projects in D
nowdays, and i'm ok both with the compiler (with some patches) and
with the documentation.

yet i have some thoughts that i mind to share, maybe in questionable
form. i wish D to became world-wide success, and although i not
motivated enough to write a code/documentation, i believe that
expressing my opinions is good for those who may think the same way but
didn't bother to write that. so they will see that they aren't alone.

i'm trying to help people in D.learn too, as this is the thing i can do
besides contributing the code/dox, so i'm not completely useless. ;-)

ah, and it's fun to answer the questions, 'cause it helps me to find my
weak points (and sometimes the questions themselves makes a nice
riddles to solve).

> You must be spending literally hours a week reading and composing 
> messages on this forum
this is the way i'm working, actually. when i designing the code in my
head, i can do various other tasks. singing, writing to forum, playing
Wasteland 2 and so on. i can even sing and play Wasteland
simultaneously!

> and my perception is that you are truly willing 
> and able to help move D forward.
yes, i love D. it has some things that i will never accept, and some
things that annoys me, but it's the language that hooked me. i really
with D to become widely-used mainstream language. but... see about
motivation above.

> Your posts are eloquent and well written. Yet when it comes to actually
> doing it, well, that's what doesn't quite happen.
that's 'cause i often want questionable things. like allowing `auto` in
`foreach`, or fixing bug #13670 (which is not-a-bug for everyone else),
or forbidding `new char[128]` syntax (or allowing `new char[128][32]`
for that matter). small improvements here and there, which will hardly
find their way to mainline.

ah, and i HATE github. ;-)

btw, you will hardly find me in DIP discussions too. i tried to design
some languages in my life and found that i'm very bad in it. so if i
don't see any obvious design flaws, i'm just lurking there.

> It seems to me that for some of us, contributing to D is like teenage 
> sex: everybody talks about it, everybody knows how it should be done 
> better than all others, but few are actually doing it.
this requires me being on github, and i don't want to be there. i
understand that my hatred towards github is silly, and github is a
great collaboration platform, but i just can't take it. and there is
virtually no sense in posting patches to bugzilla.

> My question is how we can break this pattern? What steps can the D 
> leadership take to convince folks like you to actually add, even with 
> the simplest things?
i honestly don't know. this is complex issue that involves, besides
other things, my lazyness and lack of motivation, which aren't directly
connected to the way D is developing/managed.

sure, i'm not expecting that i can direct D developement in the way i
want it to go. yet seeing the great language which is contaminated with
questionable legacy hurts my enthusiasm. i was there with C and C++:
"we have alot of legacy code, so we can't change it to be better or do
it in completely another way." i understand why not breaking people's
code is so important, but... there are alot of things that dfix can do
for us.

and sometimes even not so intrusive changes, like deprecating prefix
function attributes like `const`, aren't going to mainline. or
enforcing property syntax, which will fix the long-standing bug with
delegate properties (yes, i'm using that patches and i'm very happy
with them). or introducing `virtual` keyword, which was banned, but
desperately required to cancel `final` effect on some methods. i can go
with that "cosmetic issues" for a long time. they seems to be small,
but they are the source of everyday frustration. and knowning that even
unintrusive fixes or fixes for which dfix will be able to upgrade
source code automatically will likely not be accepted is decreasing
motivation too.

or take "magicport", for which i proposed to make a simplistic codegen
that will produce C or C++ code (not a source translation, like
magicport, but using C/C++ as "assembler" in codegen), so there will be
no problems with bootstraping D compiler on the box where only C++
compiler is available. i was told that this is unlikely to be included
in DMD, and immediately lost any interest in it -- as i don't need such
codegen, and working on it without any hope that it will be used is
just a wasted time.

i don't know what you or any other person can do with this without
turning D to "Ketmar's pet project". and i surely don't want to fork D,
as it will not do any good for both projects. i don't know.

i can stop ranting here in NGs, but this will force my overall interest
in D to fade away. i'm writing here 'cause i really care. not enough to
go to github and start contributing, but enough to highlight the things
i consider not very good. i may be annoying, but then it's easy to just
blacklist my posts in mail reader, or write simple userjs to hide my
posts on the web. ;-)


sorry for the long post which includes alot of homegrown phylosophy and
only adds new questions instead of answering to yours.
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 181 bytes
Desc: not available
URL: <http://lists.puremagic.com/pipermail/digitalmars-d/attachments/20141231/bfe09cb7/attachment-0001.sig>