Steve Teale wrote:
Can this group come up with a proper, sober (OK, I'm not), case for D.
This would clearly have to steer clear of the standard libraries, I can't see
how any outside observer is going to be impressed by the fact that we have two.
And D is a computer programming language. So we should deal with it as that
first.
Andrei's article had a lot of good points primarily revolving around the need
to concentrate on concurrency, but I suspect that we should probably stress the
basics.
When Bjarne Stroustrup was originally promoting C++, he made a strong point
that you could at least consider it to be a 'better C'. This point, it seemed
to me, was lost on many. Now we are looking for radical arguments as to why D
is a cool language. Maybe we should remember the basics, and concentrate less
on the vapor.
Bearophile made a counter-argument. But this also did not stress our basic
weaknesses. Most of us are using DMD, which on Windows uses a 20 year old
linker, and utilizes an antique object file format. Under Linux, it can't
produce the position-independent code that's required to create reliable shared
libraries.
Unless you use alpha-level code, you can't load arbitrary D modules at run-time.
There isn't a decent debugger for either Windows or Linux. There may never be
one if the potential authors see the constant focus on meta-programming - that
must make life hell for them.
I'm not advocating a return to D1, but I do want to see closure on D2, and an
ascent from the constant alpha state. Then after that, I'd like to see a more
formal system of RFCs for library proposals, and a recognized pattern for
voting on them so that anyone who kept up-to-date with the process would not be
surprised by what suddenly appeared in Phobos, or perhaps it should be the D
Standard Library (DSL).
When all that had happened I could forget computer programming and get on with
my woodwork relatively secure in the knowledge that I had chosen to support a
winner, and the Walter's efforts were not in vain.
I think the proper case for D has to be in the form of (several)
real-world applications/system-utilities that people outside this
newsgroup will *want* to use, not because they're written in D, but
because these programs actually do something useful that a lot of people
would really want.
Games are certainly a good use case for D, but developing a good game
takes a long time and in the end it won't compete with any of the
available commercial games.
There are (surprisingly) many small utilities that do something very
simple, yet they're very useful and are used by a lots of people,
Just to name a few ...
- Launchy
- safarp (simple and fast add/remove programs)
- Sumatra PDF viewer
- WinDjView
- Fotografix (extremely tiny image editor,
http://lmadhavan.com/software/fotografix/ )
- Rapid Environment Editor
- WinDirStat
- Unlocker
These are very useful "system" (kinda?) utilities ,IMO there's lots of
room for programs in this category to be written in D, either programs
that are not written yet, or written but have sub-optimal performance.
(for instance, launchy can be slow sometimes).
Of course, the end users of such utilities might not care much about
what language they're written in; but that's exactly the point.
