On Thursday, 15 January 2015 at 03:19:10 UTC, brian wrote:
On Wednesday, 14 January 2015 at 05:10:05 UTC, Joakim wrote:
I don't think it's that unusual for a native compiled
language: can you find C++ snippets to write tweets? I doubt
it. A compiled languages is just not the tool people usually
grab to write such things.
Pardon my naivity, but I don't know why I *wouldn't* want to do
this in D? I want something to pull tweets, analyse them
quickly and produce results. And then I graph it in something
else. Isn't this what this language would be *really* good at,
as it provides efficiency with minimal coding?
You originally talked about _writing_ tweets by using a D tool,
not pulling and analyzing them as you do now, which is usually
what people use a native language for. For writing tweets,
tossing off something in a scripting language is usually much
quicker and easier, as you see with all the perl/python/php/ruby
libraries listed in the official Twitter libraries link Ola
provided above.
However, D could be used for both writing and analyzing, and if
you look on dub, you'll see two libraries that advertise twitter
support:
http://code.dlang.org/packages/twitter4d
http://code.dlang.org/packages/graphite
D would like to be good at all these things, and some of them
are enabled on dub. But as a newer language with a smaller
community than Java, obviously the amount of APIs covered is
going to be less right now.
I'd love to help it on this journey, with my limited knowledge.
Where can I start?
Simple, write the kinds of libraries or API wrappers you'd like
to use.
This is the normal chicken-or-egg problem with a new language
like D, ie people want libraries or API wrappers to get their
job done easier but if they don't spend time writing them,
they never get done.
Agreed. But if I'm looking to use an API library and their
isn't one there, writing a new library entirely isn't exactly a
trivial task.
It doesn't have to be a new library. It could just be a D
wrapper for an existing C library, because D provides for
wrapping C easily. I ran the automated C->D header translation
tool dstep and modified the output a little to provide a D
wrapper for the native Android C APIs:
https://github.com/jacob-carlborg/dstep
https://github.com/joakim-noah/android
Also, maybe the D community doesn't put their code up for
public consumption as much as some other language communities.
This is my belief and main point.
Tough to say.
I don't doubt that these issues exist, I just consider them
normal for a newer language.
As for recommending D, I think it depends on the type of user.
If they care at all about efficiency, I don't know that I'd
recommend a language other than D. If they don't really care
how fast their code runs and prefer to stay at a higher
scripting level, then the advantages of D fade compared to
other languages.
Yes the language is efficient but what's the point in having
efficient code if I don't know how to get it to do what I want?
Generally, the kind of people who write efficient code don't mind
writing and figuring out such stuff themselves, without relying
on existing code examples.
While I'm not sure if you agree or disagree with me, I think
you also clarify what I was trying to say.
My point wasn't that there aren't ways to do things in D.
My point was that there are fewer examples of *how* to do
things in D.
This will discourage the new user, which will prevent it
becoming a more popular language.
So if I'm looking for a new language to use, I'd probably stick
to the ones that have more examples.
I agree that most users learn from examples, so to the extent D
is missing those, it will not get as many users. But any new
language without corporate support has this problem, as language
users generally don't bother taking the time to slap all their
support code and examples online.
dub has been an effort to remedy that, and taking your twitter
example, appears to have worked, based on the two links I found.
Obviously, D still has a ways to go to reach ruby or Java levels
of example/library availability.
