On Sun, Aug 20, 2000 at 04:17:37AM -0500, Brandon wrote: > > > > Brandon rejects this elegant solution because of semantic objection (as > > I reject this because using an anomalous condition (zero length data) as a > flag for special data (freenet control entry) is bad protocol design. It > is not immediately obvious when looking at a message that a message with > no data is a freenet control entry. > > I prefer to explicitly state such a condition with a field in the message > saying hey, this is a freenet control entry rather than having it be > implied.
Actually orthogonality is the key to good design, so if we already have a combination of fields (DataLength and meta-data length) that imply that something is a control message or not, putting in another field to say so explicitely is bad design. What is "immediately obvious" to you is not that important, humans are not going to be looking at these messages. > > > > _______________________________________________ > Freenet-dev mailing list > Freenet-dev at lists.sourceforge.net > http://lists.sourceforge.net/mailman/listinfo/freenet-dev > -- \oskar _______________________________________________ Freenet-dev mailing list Freenet-dev at lists.sourceforge.net http://lists.sourceforge.net/mailman/listinfo/freenet-dev
