On 06/10/10 13:08, Todd Walton wrote: > On Sat, Oct 2, 2010 at 11:55 PM, Volodya <[email protected]> > wrote: >> Freenet is in active development. > > The good old standby argument. > >> Stability is good, but would you rather have a buggy insecure node? > > So, Freenet is buggy and insecure? I understood that it was caveat > emptor, but being so buggy and insecure that a fix can't wait a mere > two weeks or so is something else entirely. >
(I meant to reply to this much earlier, forgot, sorry.) I agree, the "in-development" argument shouldn't be used as an excuse, and it's wrong to use it as an excuse. Frequently *forced* updates are a *bad* thing and we should aim to reduce them. The reason for the "forced" updates is *protocol changes*, due to freenet being a distributed network. This means that newer nodes won't talk to, or will have trouble talking to, older nodes. This is what's unavoidable and "forces" an update - we don't directly force people to update; you can turn auto-update off. I don't know exactly how Matthew chooses when to make a particular build mandatory, but I'd guess it would be based on a subjective judgement of when to force all nodes to switch to the new protocol, rather than have parts of the network talking different languages. I'm pretty sure that fixes to buggy/insecure code don't *force* updates (at least, we shouldn't do this - no other project forces updates, even for security reasons). The changelogs don't actually explain this. I guess it would be best to explain exactly why a build is being made mandatory, such as highlighting the particular protocol change that's triggering it. X _______________________________________________ Devl mailing list [email protected] http://freenetproject.org/cgi-bin/mailman/listinfo/devl
