On Thursday, 24 September 2015 at 09:48:24 UTC, Chris wrote:
On Wednesday, 23 September 2015 at 23:00:29 UTC, Laeeth Isharc
wrote:
On Wednesday, 23 September 2015 at 16:22:35 UTC, Joakim wrote:
To break out to an early majority, D will have to prove
itself, ie the innovators and early adopters have to show
empirically that it is working better for them and allowing
them to do more.
I think you are spot on.
I agree. Conventional marketing won't get us far at this stage.
We agree (although the theory of marketing I think points in the
same direction I described, whether or not that's what people
understand by marketing).
Thank you. Nicely put. Mind you, a lot of complaints are not
related to the language itself (as others have said), but are
secondary issues like IDEs
(one-click-debug-compile-run-deploy-go-for-coffee-magic) and
libraries, which are logically a step you take _after_ a
language has been created.
It's an ecosystem and things co-evolve.
Why do the documentation and IDE options fall short of what many
newcomers might expect? Because the community seems to me to be
comprised of serious programming types, and for such people the
importance of documentation and IDEs is less than it might be for
others. That's a strength of the community, not a weakness, but
it just means at this time those aren't especially selling points
of D (although it's improving every year). At some point the
dynamics will change and either existing D IDEs will become good
enough, or someone will be motivated to write one.
These things do make a difference. At least for the Python
crowd. But be prepared that people might attack you saying that
with C++ it would be 10-20% faster than D, because D has GC
blah blah blah.
Yes, but the reason it takes him an hour today whilst he is
putting money to work behind this strategy is that the
alternative to the internal scripting language is C++, and that
will cost time and money. Having to make a business case for
something often means that projects with a high return on
investment don't get done, or take a long time to be done,
because of the human factors.
And if you have to wait 3 minutes (remember, this is on my home
machine with dmd debug mode) or 2.7 minutes, it's not an
important difference. Because I still remember what I was
thinking when I ran the study. But after an hour I have
completely forgotten and will be doing something else.
Of course you can scale up to more machines, but the cost of
adding complexity for a small tool isn't zero, even if the cost
of raw horsepower is close to zero.
The amount of random criticism that is thrown at D, confirms,
imo, that it is really good, else people wouldn't bother to
attack it so passionately. Only really good creations are
attacked with a passion - be it in art or technology.
Yes - that's very insightful. I wonder why that is.
Laeeth.
