On Monday, 20 August 2018 at 12:26:25 UTC, Laeeth Isharc wrote:
On Monday, 20 August 2018 at 11:55:33 UTC, Joakim wrote:
"So how do you change paradigms? Thomas Kuhn, who wrote the
seminal book about the great paradigm shifts of science, has
a lot to say about that. In a nutshell, you keep pointing at
the anomalies and failures in the old paradigm, you keep
coming yourself, and loudly and with assurance from the new
one, you insert people with the new paradigm in places of
public visibility and power. You don’t waste time with
reactionaries; rather you work with active change agents and
with the vast middle ground of people who are open-minded."
(Quoting from the article I think).
Kuhn and Lakatos. Paradigm shifts don't take place when the
dominant paradigm is defeated by logical or empirical means.
Paradigm shifts take place when for some reason people say "how
about we stop talking about that, and start talking about this
instead".
Not sure why you'd call that anything other than defeat. :)
I think he described certain political changes in the Western
World beginning in the mid to late 60s rather well. I don't
think it describes how changes in the sphere of voluntary
(non-political ie market and genuine civil society) activity
unfold. Supposing it were a good idea (which it isn't), how
would one be able to to insert people in places of public
visibility and power who put forward a point of view that is
very different from the prevailing one? Only via a program of
entryism, and I don't think that in the end much good will come
of that.
By convincing those with power/visibility that the contrary view
is worth integrating? Look at Microsoft's about-face on open
source over a couple decades, going from denigrating it to buying
open-source producing or supporting companies like Xamarin and
Github and open-sourcing several of their own projects, as an
example.
So I think the original author has cause and effect the wrong
way around (not too surprisingly because he is talking about
things that relate to politics and activism). [NB one
shouldn't mention the Club of Rome without mentioning what a
failure their work was, and it was predictably and indeed
predicted to be a failure for the exact same reasons it failed].
It isn't that you insert people representing the new paradigm
in positions of influence and power.
It is that people from the emerging new paradigm - which is
nothing, a bunch of no-hopers, misfits and losers viewed from a
conventional perspective - by virtue of the fact that it has
something useful to say and has drawn highly talented people
who recognise that start ever so slowly to begin things and
eventually to accomplish things - still on the fringes - and
over time this snowballs. After a while turns out that they
are no longer on the fringes but right at the centre of things,
in part because the centre has moved.
The best illustration of this phenomenon was I think in a work
of fiction - Neal Stephenson's Cryptonomicon. I never expected
someone to write a novel based on a mailing list - the
cypherpunks. It was about as surprising to me then as it would
be to see Dlang - the movie - today. And of course that itself
was an early indication that the ideas and ways of thinking
represented by what was originally quite a small community were
on the ascent.
I agree that she's looking at it from the point of view of
governmental change for her environmental agenda, whereas the
market is more likely to have entirely new institutions- it used
be new _companies_, but with the internet it's now increasily
decentralized operations like the community behind bitcoin or bit
torrent... or D- form that become much more important than the
old ones: creative destruction. So, significantly open-source
Android replaces mostly closed Windows as the dominant OS used by
most consumers for personal computing, rather than Microsoft
really getting the new religion much.
This pretty much reflects what Laeeth always says about
finding principals who can make their own decisions about
using D. "Places of public visibility and power" for D are
commercial or open-source projects that attract attention
for being well done or at least popular.
Well - I understand what you mean, but I don't recognise this
as being my point. Principals who can make their own decisions
probably aren't today highly visible and visibly powerful. The
latter comes much later on in the development of a project,
movement or scene and if you're visible it's a tax that takes
time away from doing real work. By the time you're on the
front cover of Time or The Economist, it's as often as not the
beginning of the end - at least for anything vital.
You're misreading what she wrote: she only said that you place
new people in positions where they have some visibility or power,
again because of her emphasis on government change, not that you
convince the already _highly_ visible/powerful to go your way.
Since the former can be almost anything depending on the context,
I gave examples of who that might be for D.
We're doing both: most of the material on the D blog and my
own D interviews are not with corporate representatives. We
could stand for more of the latter though, especially the big
successes, because people are more influenced by them.
I'm not saying it's a bad thing to go for big stories. But
it's a mistake to place the attention people today naturally
tend to. It doesn't matter what influences most people - it
matters what influences the person who is poised on the edge of
adopting D more widely, adopting D as a beginning, or would be
if they knew of the language. The latter is quite a different
sort, I think.
To know about D in the first place, you might have the sort or
programmer who's not going to actually use it become aware of it
and mention it to the guy who would use it but never heard of it.
In other words, there are different ways of getting to the people
who would use D: I think this is one of the better ones, but we
need to use several ways.
One of the best is a track record of successful projects using
the language.
Liran at Weka picked up D because he saw Kent Beck post on
Twitter about Facebook's Warp written in D (or maybe it was a
linter) and it seemed like an answer to a particular problem he
had (if I am remembering correctly). It wasn't because of a
grand thing - it was because of a little thing that seemed like
it might be a creative solution to a real problem.
Signal:noise is much higher away from the limelight too. By
far better to have a high share of attention in some specific
domains or interest groups than to have a low share of
attention of some enormous market.
Sure, we need to try several ways to attract attention.
Many devs use large corporate deployments as a litmus test of
whether they should explore a new tech. To the extent that
we've never published a blog post about Weka, only offhand
mentions like when Andrei visited Israel, that is a big
marketing failure for D.
I know the Weka guys are very busy, but the further success of
D will only help them too, so they're undercutting themselves
by not making sure that blog post gets done.
Well, someone could just take the key insights and experiences
from their talks, write them up, check with them and post. The
latter are a considerable commitment already for a startup
that's hitting a revenue growth phase. There are lots of
things for them to be busy with beyond just the technology.
Yeah, that could work.
Finally, regarding leverage, I keep pointing out that mobile
has seen a resurgence of AoT-compiled native languages, but
nobody seems to be trying D out in that fertile terrain,
other than me.
I did try, but it's not exactly easy to make a complete app
in D, even on Android. It would be great if there were some
way to automatically wrap the APIs.
Right now, the Android port is more suited for writing some
performant libraries that run as part of an existing Android
app. The kind of polish you're looking for will only come with
early adopters pitching in to smooth out those rough edges.
If we had autowrap for JNI and could dump the types and method
prototypes as part of the pre-build process, what would the
next stage be to be able to just call Android APIs from D and
have them work? JNI isn't that bad (I know it's deprecated)
and I used it already from D in a semi-wrapped way. So I
wonder how much more work it would be to have autowrap for JNI.
I didn't use reflection on the Java side because I wasn't
wrapping that much code. Are there XML descriptions of Android
APIs you could use to generate wrappers?
No idea, not something I've looked at nor plan to, as I mentioned
to you in over email recently. I'd like to use mostly D with not
much Java on Android, so I'm not too worried about the interface
right now.