C++ to catch up?
Laeeth Isharc via Digitalmars-d
digitalmars-d at puremagic.com
Sun Feb 1 15:20:14 PST 2015
On Monday, 5 November 2012 at 18:20:23 UTC, Jonathan M Davis
wrote:
> The closer that C++ gets to D, the less interested that many
> people will be in adopting it, particularly because of the
> large user base and the large amount of code out there that
> already uses C++. Programmers have to be convinced to move to
> D, and for many C++ programmers, the improvements to C++11 are
> enough to make a move to D not worth it, even if D is a better
> language.
(He goes on to point out that nonetheless D will always have the
edge because legacy and installed base).
One should be careful about superficial translation of instances
from the purely commercial world to the world of languages, but
it strikes me that Clayton Christensen's Innovator's Dilemma does
apply somewhat to the case of D vs its peer languages. His
central point is that in the beginning disruptive innovation very
often tends to commence as a niche thing that may well be
globally inferior - he uses the example of Honda motorbikes that
allowed them to gain a foothold, and that once they dominated
this niche and gained succour from it were able to use to expand
their footprint to the extent that they posed a serious threat to
the established dominant players. But for many years, these (and
later the cars) were seen as products of clearly inferior quality
that had the advantage of being cheap.
The interesting thing is the emotional aspect of perception -
nobody would have taken you seriously had you predicted in the
early stages that Japanese auto makers would become what they
subsequently became. And one could have pointed out some decades
after the war ended that they had been in the business for years,
and why should anything change. This is exactly what people say
about D - it's been around forever and hasn't taken off, so why
bother. (see recent Slashdot thread for an example of this).
It is a basic insight of gestalt psychology that perception is
shaped by emotion (really it's affect, which goes much deeper -
emotion is the tip of the affect iceberg), and one way to know
when this is occurring (my background is as an investor and
speculator, so I have devoted a couple of decades to applying
this in a practical way) is that on the one hand you have an
emotional intensity out of proportion to the importance of the
topic, and on the other the reasons people put forward to justify
how they feel are observably not in accordance with the facts.
See the Slashdot thread...
So in any case, D is not competing on price, but has other
strengths that are of very high appeal to a certain group (if you
want to write native code in a productive way) even though one
must honestly acknowledge its imperfections in a global sense -
reading back through the forums a dozen years, this seems to
occur quite regularly in waves. "When is D going to be
finished?" even a decade back. To be upset by the imperfections
is missing the point, because languages - even programming
languages - have a certain innate pattern of development (that
resembles Goethe's observations about the metamorphosis of
plants) that can't be forced, no matter how much one grumbles or
stamps one's feet.
Furthermore, people tend to extrapolate superficial trends even
though history tells us this is a poor guide to the future.
Japanese cars really took off once crude exploded in the early
70s (and again towards the end), and auto-makers were slow to
respond. Perhaps they did not organize their business on the
basis of a prediction abuot energy prices, but the point is they
were ready to take advantage of this shift when it occurred.
I do not want to attempt to be a pundit, but it is interesting
that the notable use cases of D - at Sociomantic, Adroll, and
Facebook are all aligned with certain salient and very powerful
underlying technological drivers and trends. It's no longer true
in many applications that programmer time is expensive compared
to machine time, and large data sets encountering the challenges
of memory vs CPU trajectories create new challenges and require
new approaches. And it is a positive for D that some of its
competition does not take D seriously at this stage - one thinks
for example of Guido and his insistence that execution speed
ought not to be a factor given work is I/O + network bound, even
though this is less true for numerical computing and some kinds
of data crunching. (Not that D is mature here, but there is much
that can be done within the existing framework).
In any case, dissatisfaction channeled in a constructive
direction is a positive thing, because it is the opposite of
complacency and is the edge of the challenger. The point isn't
how people feel, but how they respond to the challenges in front
of them.
As a newcomer, it is very satisfying to see the progress made on
documentation, ecosystem, and C++ integration and I have quite
some respect for the difficulty of the roles of Walter and
Andrei. One is so short of time and attention, and no matter how
hard one works and, whatever decisions one makes, it is
impossible to keep everyone happy. If one isn't being
criticized, one isn't doing it right. (Which is not to say that
some of the criticisms will not have merit).
Here is a table from an article exploring Christensen's ideas.
There are some resonances with past and current questions in the
development of D, although as I said one can't map things
perfectly because it's a different situation. And the original
book is better than what has been written based upon it.
[Table doesn't format well, but you can see it here. The
surrounding text is less relevant].
http://recode.net/2014/01/06/the-four-stages-of-disruption-2/
Laeeth.
More information about the Digitalmars-d
mailing list