Talk:Distributed computing

From Citizendium
Revision as of 18:03, 26 February 2007 by imported>Markus Baumeister
Jump to navigation Jump to search

BOINC

I don't really have any expertise in this area, so I can't help much with the article, but you'll probably want to include projects other than seti@home that run on the BOINC platform - maybe in a section specifically about BOINC? - see here: http://boinc.berkeley.edu/projects.php --Joe Quick | Talk 11:20, 23 February 2007 (CST)


I think projects like SETI, FOLDING@HOME and others are perfect examples for its deployment. Robert Tito | Talk 11:21, 23 February 2007 (CST)

recursive calculations

these are the most used calculations used, based upon a small dataset and a simple formula trying to calculate a "best Value average" for a property at a certain point of time in the flow of the total process Robert Tito | Talk 17:18, 23 February 2007 (CST)

  • If you said that recursive functions are one of the typical aims of distributed computing, I would agree. But the calculation is often done iteratively (as the function is end-recursive). --Markus Baumeister 17:35, 26 February 2007 (CST)
  • True but the iterative partial results are part of a recursive total. Robert Tito | Talk 17:43, 26 February 2007 (CST)
  • Well, if you say so, I'm not a mathematician. I just wanted to point out the misunderstandable formulation (i.e. the calculations done on the nodes will often not be recursive). --Markus Baumeister 18:03, 26 February 2007 (CST)

Can we link this page to Programming language? preferrably GNU gcc, java are used in these programs. Robert Tito | Talk 14:33, 26 February 2007 (CST)

  • (this got lost by overlapping edits) I would suggest not to force linking. In the end every program needs a a programming language but we don't want to refer from nearly overall to that article. BTW, distributed computing and especially parallel computing profits much from not using imperative languages like C(++) or java but automatically parallelizeable ones (like I don't know). --Markus Baumeister 18:03, 26 February 2007 (CST)

Distributed computation?

Sorry to be the one constantly criticizing titles. But I know this topic as Distributed Computing. And several web pages do so too, like: http://www.distributed.net/ (the original cow cracking DES (I participated)), http://www.distributedcomputing.info/ and of course everyone's new darling http://boinc.berkeley.edu/intro.php. --Markus Baumeister 17:35, 26 February 2007 (CST)

I can agree to that, I will move the page to distributed computing - if none will object. Robert Tito | Talk 17:43, 26 February 2007 (CST)


Drawbacks and disadvantages

The first two paragraphs of this very clearly refer to general problems of distributed systems (it even uses that term) not specifically computing. The quote (I also don't know if it is really from Lamport) is typically related to systems like NFS which, at least in former times, managed to bring down whole institutes despite slave servers, etc. .

The third and forth paragraph then refer to general problems of parallel computing (although worsened here by the size and lack of speed of the connecting network).

I'm not sure what to do about that. Maybe mentioning that we have this join of problems may be enough, maybe more changes are necessary. --Markus Baumeister 18:00, 26 February 2007 (CST)