Re: intent of package seti@home
Kevin Dalley writes:
> There are subtle differences between most free software and a distributed
> program. There are many buggy versions of the Linux kernel around.
> (2.2.1, for example).
Which was released by Linus: the equivalent would be SETI releasing a new,
improved client and having the hackers quickly discover some serious bugs
and report them with patches. How many kernels that were not approved by
Linus are in widespread use?
> A bug in one copy of a distributed system could affect the final results.
Then the system is hopelessly fragile. What if the box some client is
running on has a hardware problem? Are the clients statically linked? I
assume they use floating point. What if the box has broken floating point
emulation?
> Not at all. But there may be advantages of knowing which data was
> analyzed with which bugs.
Which is one reason why each result report should include the version of
the client that generated it.
--
John Hasler
john@dhh.gt.org (John Hasler)
Dancing Horse Hill
Elmwood, WI
Reply to: