Hi Ian, Thursday, March 07, 2002, 5:44:47 PM, you wrote:
IC> On Tue, Mar 05, 2002 at 02:34:40AM +1300, David McNab wrote: >> Re-code the whole fucking thing in C or C++ !!! IC> You seriously believe that writing software in C++ is faster than IC> writing software in Java? I have never met anyone equally familiar with IC> both C++ and Java who would agree with you. The simple fact that you IC> don't need to worry about bounds-checking in Java alone gets rid of a IC> significant percentage of bugs which plague C++ software. Freenet needs a large installed base to really come into its prime. The complication imposed by the need to separately download and install a JVM will deter a large percentage of windows users (who outnumber linux users at least 12:1), thus robbing Freenet of 90+% of its potential installed base. Also, refer earlier comments by others on this mailing list about the problems inherent in all the available free JVMs. Also, there are low-level bit-bashing buffer-munging areas where Java is just plain inefficient. One specific nagging problem with Fred is the failure of the bandwidth throttling. This is echoed by a few Slashdot comments I've seen, where people object to allocating the lion's share of their bandwidth for a lot of stuff they're unlikely to use personally. Because there's no effective bandwidth throttle, I don't run a freenet node myself, and haven't done so for several months. In the times when I do fire up my transient node, 8 out of 10 timss I get repeated 404's on CofE - a site which by virtue of its popularity, should be easy to find. Sometimes when I re-gen my datastore, I can't even get [EMAIL PROTECTED] >> And clean up the design in the process. IC> What is wrong with the design? Sure, it isn't perfect, but I am curious IC> as to what specific suggestions you have (you obviously have some if you IC> are implying that there are serious flaws in the current design). See above. Also, in my experience, reliability of key retrieval in 0.3 was better. David _______________________________________________ freenet-tech mailing list [EMAIL PROTECTED] http://lists.freenetproject.org/mailman/listinfo/tech
