Manu
(answering the OP)
Stupid me, I just adressed similar issues in
http://forum.dlang.org/thread/jlssvljpdgxlubdaj...@forum.dlang.org
("End user experience with D") because first I thought this here
was about game development (which I don't care about).
Short version: I agree with Manu. Completely. Strongly.
One warning tough: We should differentiate between "essential"
and "nice but not urgent". For a simple reason: Feasibility.
Sure, .chm help would be nice for Windows users but frankly, good
doc at all are more important; if it's, say, HTML that's good
enough for a start.
As for an IDE we should definitely favour something cross
platform (and non-java!!!) rather than caring about peculiarities
of Windows, OSX, lx/kde, etc. Those can be done later. Really
important is to have a base asap.
The killer: Debugging.
GDB is bare minimum and seeing adresses of structures rather than
the members and their values is basically worthless.
When trying a new language I usually do a small project:
"osnsort", i.e. a utility that takes (usually via a pipe from 'du
-h') lines with OS style numbers (like "4,3M" for 4,4MB filesize)
and then sorts those lie properly (unlike sort which lousily
fails).
While D nicely showed some of its strengths and the coding was
really enjoyable and the docs good enough for someone with a C
background) I soon found myself inserting lots of "debug"
statements as a "more elegant" version of the old "#ifdef DEBUG
printf(...) #endif". *** YUCK!!! ***
Putting it bluntly: D is so great, so powerful, so nice language
that it's worth a lot of efforts. With any other language I would
have turned away and taken it to be a hobby thingy.
So, we must not please everyone and take care of their OS issues
right now, but we definitely need some kind of, if somewhat
crude, working base in terms of debugger support, IDE and docs.
A+ -R
