Path: csiph.com!v102.xanadu-bbs.net!xanadu-bbs.net!feeder.erje.net!eu.feeder.erje.net!news.albasani.net!.POSTED!not-for-mail
From: Sebastian <news@seyweiler.dyndns.org>
Newsgroups: comp.lang.java.programmer
Subject: Lazy primes with Java 8 (was Re: Java 8 Streams and Eratosthenes)
Date: Mon, 10 Jun 2013 01:27:54 +0200
Organization: albasani.net
Lines: 58
Message-ID: <kp331c$n68$1@news.albasani.net>
References: <kolnme$g3e$1@news.albasani.net> <ZYtrt.80089$cD1.53480@newsfe29.iad> <komnrp$3sh$1@news.albasani.net> <komq8g$8hc$1@news.albasani.net> <IpNrt.2$He1.1@newsfe05.iad>
Mime-Version: 1.0
Content-Type: text/plain; charset=ISO-8859-15; format=flowed
Content-Transfer-Encoding: 7bit
X-Trace: news.albasani.net f4QEoV3LMyBwVkBL08EevuRacC0V21XukJ0Y1/oqO2xtA66bb6Z4aPKVZg51BExTHrCE+HALsexJLYrCJQej8w==
NNTP-Posting-Date: Sun, 9 Jun 2013 23:27:40 +0000 (UTC)
Injection-Info: news.albasani.net; logging-data="ElYOJjKofkYViY13SAUzt/yA7I7m7WeUgQRuGswZe71UnNYtd3eWdfWG0+Beq2325D+EFAp5t5H676FlFD2s4J3hsi+pl58G8s9mGpk+rm03NIDn3dXG5ePAfidGqEDa"; mail-complaints-to="abuse@albasani.net"
User-Agent: Mozilla/5.0 (Windows; U; Windows NT 5.1; de; rv:1.9.2.15) Gecko/20110303 Thunderbird/3.1.9
In-Reply-To: <IpNrt.2$He1.1@newsfe05.iad>
Cancel-Lock: sha1:RQ5sGdhZ3y42BCTsI2/9uACit8A=
Xref: csiph.com comp.lang.java.programmer:24347

Am 05.06.2013 22:43, schrieb Arved Sandstrom:
> On 06/05/2013 04:44 AM, Sebastian wrote:
[snip]
>>
> Me I wouldn't re-invent the wheel in this case. Implementing this type
> of laziness is not that straightforward.
>
> And why not use interop between JVM languages? This way you leverage the
> strengths. I usually have Java as the main language, but for some
> functionality in some apps I use Scala. Maintainability goes up, for
> starters, as does reliability and speed of development.
>
> It's all ultimately JVM bytecode.
>
> AHS
>

It's not everyone has a management who sees it this way. In our large
project (we're talking several hundred developers and a really
complicated build process), I don't see anyone advocating more
complexity by adding another compiler and language.

Anyway, it's turned out not to be too complicated in straight Java.
One can simply extend a straightforward implementation of the usual
cons-list with the following features, giving one a
FilterableConsList<T>:
a) accept a Function<T> to generate the elements in the tail
    in sequence
b) accept a Supplier<FilterableConsList<T>> to supply a tail
c) accept a Predicate<T> to filter out non-matching elements
    from the tail
and make the implementation class LazyConsList<T> lazy for the tail
(but eager for the head).

It's also easy to implement AbstractCollection by providing an iterator
and make the thing streamable by providing a spliterator of unknown
size.

With that, the driver for generating the first 100 primes becomes:

LazyConsList<Integer> s = sieve(LazyConsList.newList(2, x -> x + 1));
s.stream().limit(100).forEach(x->System.out.print(x + " "));

where the interesting sieve() method is defined like this:

private LazyConsList<Integer> sieve(FilterableConsList<Integer> s) {
   Integer p = s.head();
   return cons(p, () -> sieve(s.tail().filter(x -> x % p != 0)));
}

Now I think that looks pretty neat (and almost exactly like the
corresponding Scala code.)

-- Sebastian