fish wrote: >>Also, how does the fact that this "defeats the purpose of implenting >>updatable keys" make it "no better than most of the ideas going around for >>editions right now"? > > I am forced to assume that you incorrectly phrased the question. either > that, or you havn't tried to run a freesite ;). But I will answer it with > aquestion: > > what's different about TRK's to progressivly checking each previous time > periods DBR url if today's isn't found?
All the unsuccessful requests for the old DBRs would go the full HTL. A TRK just takes a single request that goes the full HTL, even if the site was not updated in months. This is not really expensive: On the latest version of an edition based site the image link to the next edition causes an unsuccessful request. One could get rid of that with TRKs. >>2. Minimal request length (requests that must go the full HTL are >>undeseriable) > > agreed, at least for browsing, and provisionally agreed for FNP traffic > internally. You can let each reader decide, if he accepts a possibly old version in exchange for lower latency. - Thomas Leske _______________________________________________ devl mailing list devl at freenetproject.org http://hawk.freenetproject.org/cgi-bin/mailman/listinfo/devl
