Is there a map reduce implementation in common lisp? like hadoop?
gavino
D Herring
> http://hadoop.apache.org/core/
Why do you ask?
Probably not. Common lisp is a very hard language to use. It is only
good for 1970's AI technology. As a result, today's user must rely on
old, musty libraries which lack modern features.
You would probably have better luck at comp.lang.c++ or even better at
comp.lang.python. Their techniques are much more advanced. Think
about it: since C++ and Python are modern languages, their designers
were able to learn from and avoid Lisp's mistakes. Therefore, they
are vastly superior for all modern tasks such as map-reduce.
:)
Vetle Roeim
> http://hadoop.apache.org/core/
I don't know, but Hadoop can be used more or less with any language,
via Hadoop Streaming[0] and Hadoop Pipes[1]. Personally I've used it
both with both Java and Perl (don't ask), and it worked fine with both.
Actually, I found it to be a little more work to use it with Java,
even though that is supported "natively", so to speak. It might be
more flexible with Java, as you have easier access to internals.
--
It's not a bug, it's the future.
[0]: <URL: http://wiki.apache.org/hadoop/HadoopStreaming >
[1]: <URL: http://hadoop.apache.org/core/docs/current/api/org/apache/hadoop/mapred/pipes/package-summary.html >
George Neuner
wrote:
>I can't figure out if you realise how lame you sound....are you
>european? I mean do you think trying to talk down to someone does
>anything but make you look like a complete tool? Show an app that you
>have created that does something......then talk.
If you had actually learned any Lisp in the last 2 years, you'd know
that Lisp has a whole family of higher order map functions (map,
maplist, maphash, ...) and also a reduce function.
Had you known this, you would have realized that Herring misunderstood
your intent and would have refined your question appropriately.
As for a parallel implementation of map/reduce, I don't know of an
existing one for Lisp - perhaps someone else does. However, it would
not be terribly difficult to implement. There are 2 possible models -
why don't you implement them both and tell us which works better.
George
--
for email reply remove "/" from address
gavino
> On Wed, 27 Feb 2008 09:35:16 -0800 (PST), gavino <gavcom...@gmail.com>
woa do you think I could if I learned lisp?
awesome!
gavino
don't ask re perl lol, comedy, perl seems to be like porn, everyone
uses it to get job done but no one wants to admit it!
Vetle Roeim
> On Feb 27, 11:43 am, Vetle Roeim <vetler+use...@gmail.com> wrote:
>> gavino <gavcom...@gmail.com> writes:
>> >http://hadoop.apache.org/core/
>>
>> I don't know, but Hadoop can be used more or less with any language,
>> via Hadoop Streaming[0] and Hadoop Pipes[1]. Personally I've used it
>> both with both Java and Perl (don't ask), and it worked fine with both.
>>
>> Actually, I found it to be a little more work to use it with Java,
>> even though that is supported "natively", so to speak. It might be
>> more flexible with Java, as you have easier access to internals.
>
> don't ask re perl lol, comedy, perl seems to be like porn, everyone
> uses it to get job done but no one wants to admit it!
Hey, I admitted it! ;)
William James
rober...@gmail.com
> http://hadoop.apache.org/core/
I remember 30-40 years ago a LISP machine was created with hundreds of normal processors so it can manage map-reduce functions in a transparent way for the programmer.
For those that think LISP is old, bad and complex, they must know that the memory management model in Java is a poor copy of the one in LISP. So to me, it seems that other language designers have not learnt too much from LISP.
I've been working with LISP during years and I haven't found any other computer language and development environment so powerful, even today.