tag:blogger.com,1999:blog-12753102.post7936303141356967927..comments2024-03-28T15:41:37.170-04:00Comments on Ben's Journal: Cicada: Erlang Style Message Passing for SISCBen Simonhttp://www.blogger.com/profile/09833753747177544979noreply@blogger.comBlogger6125tag:blogger.com,1999:blog-12753102.post-58547229275667764102007-11-06T21:29:00.000-05:002007-11-06T21:29:00.000-05:00Mike -You're so right - the ability for SISC to li...Mike -<BR/><BR/>You're so right - the ability for SISC to live in both the Scheme world and the Java world is an incredible asset.<BR/><BR/>I doubt I could get away with implementing as many real work projects in Scheme as I do, if I couldn't run it in a standardized container.Ben Simonhttps://www.blogger.com/profile/09833753747177544979noreply@blogger.comtag:blogger.com,1999:blog-12753102.post-58041819480488241582007-10-23T11:47:00.000-04:002007-10-23T11:47:00.000-04:00I'd have to say SISC Scheme's main advantage over ...I'd have to say SISC Scheme's main advantage over Erlang is that it runs in a JVM.<BR/><BR/>That's the situation I'm in right now: I want to use a better language than Java, but I have to play the "deployable" game. SISC's conformance to R5RS and native Java interaction is really killer, such that I can make an app or library callable from Java, and no one is the wiser that I'm using Scheme.<BR/><BR/>I like Erlang's concurrency model, and the particular problem I'm trying to solve fits very nicely into the domain. I'd like to explore it to see how closely it maps, but I can't deploy Erlang on Java.<BR/><BR/>Thanks to Termite and Ben with Cicada...this is getting a lot closer. If I have free time (yeah right!) I may look into using UBF for serialization between Termite or Cicada nodes to see if that will reduce some of the marshalling overhead.<BR/><BR/>Thanks again Ben! Bravo!Mike J. Bellhttps://www.blogger.com/profile/15836560206877247391noreply@blogger.comtag:blogger.com,1999:blog-12753102.post-5393173303057724002007-05-28T16:11:00.000-04:002007-05-28T16:11:00.000-04:00Good question Pierre.I'm going under the assumptio...Good question Pierre.<BR/><BR/>I'm going under the assumption that most APIs tend to be available in main stream languages, such as PHP, Perl, Java and C.<BR/><BR/>If I pick and arbitrary topic, say integrating with <A HREF="https://www.paypal.com/cgi-bin/webscr?cmd=xpt/cps/general/SoftwareDevKit-outside" REL="nofollow">PayPal</A> or calculating <A HREF="http://www.merriampark.com/ldjava.htm" REL="nofollow">Levenshtein Distance</A>, there's bound to be an API available in Java. Would there necessarily be one in Erlang?<BR/><BR/>True, I don't think Java has quite the library base of say Perl (via CPAN), but it still seems pretty significant. Especially given projects like <A HREF="http://jakarta.apache.org/commons/" REL="nofollow">Jakarta Commons</A>.<BR/><BR/>Can you give me some examples where Erlang had a library and Java didn't?Ben Simonhttps://www.blogger.com/profile/09833753747177544979noreply@blogger.comtag:blogger.com,1999:blog-12753102.post-37971935694535902302007-05-28T16:00:00.000-04:002007-05-28T16:00:00.000-04:00I am a bit puzzled by the main advantage of SISC o...I am a bit puzzled by the main advantage of SISC over Erlang ... I have been working with java for years and my humble impression is that java fails miserably on the libraries side (compared to other languages).PierreRhttps://www.blogger.com/profile/07709580305967202558noreply@blogger.comtag:blogger.com,1999:blog-12753102.post-53334691474273390572007-05-28T10:22:00.000-04:002007-05-28T10:22:00.000-04:00Dominique -Wow, that's all good news.The SISC cont...Dominique -<BR/><BR/>Wow, that's all good news.<BR/><BR/>The SISC continuation idea is a good one. We'll see which happens first: I want to use this in a larger context that pushes SISC's capabilities, or a portable version of Termite appears.<BR/><BR/>I'm hoping for the latter :-).<BR/><BR/>Using this lightweight message passing as a way to interconnect schemes would be awesome.Ben Simonhttps://www.blogger.com/profile/09833753747177544979noreply@blogger.comtag:blogger.com,1999:blog-12753102.post-27532771294704087082007-05-28T07:01:00.000-04:002007-05-28T07:01:00.000-04:00Nice!Regarding the threading issue: if SISC contin...Nice!<BR/><BR/>Regarding the threading issue: if SISC continuations cheap enough, maybe they can be used to implement a threading system on top of them. That's how Gambit-C threads work.<BR/><BR/>Also, I heard that Marc Feeley (the author of Gambit-C) has a student who works on a portable implementation of Termite. It should be available later this summer. I hope we'll be able to code distributed applications using heterogenous Scheme systems this way.Dominique Boucherhttps://www.blogger.com/profile/18059572110310581530noreply@blogger.com