> Becuase our needs will change. I really doubt that we won't come up with
> some new cool thing we can do to SSKs. Overloading the SSK syntax is quite
> frankly a nieve thing to do in the long run.

Your system should be as complicated as it needs to be to meet your known
needs. You can't realistically design for things you haven't thought of
yet. Doing so leads to systems which are so overgeneralized as to be
difficult and unpleasant to actually use. DBRs are a fundamental thing for
Freenet publishing, not a nifty but unnecessary feature like
splitfiles. So it's okay to merge them into the key. Everything else can
still go in metadata like before.

The difference between fundamental and nifty features is that fundamental
features are added slowly and have to be supported by everyone. So it's
okay to do things like change the URI syntax to include them. How often do
they come up? In 3 years we've come up with 4 fundamental things: names,
signatures, content hashes, and date-based indexing. New fundamental
things are going to be added slowly enough and will be important enough
that we should add new key types for them. We do, after all, have an
expandable key type system for this reason. In the particular case of
DBRs, they are only reasonably used with SSKs, so it makes sense to merge
them into the SSK URI format. If this were not the case then the
reasonable thing would be to either expand the overall URI format (if it
was a really pervasive and fundamental thing, such as with docnames) or to
add a new key type (such as with SSKs).

For not so fundamental things, such as splitfiles, the appropriate thing
is to define a new metadata message and put such information in a control
document.

URIs need to be kept simple.



_______________________________________________
Devl mailing list
Devl at freenetproject.org
http://lists.freenetproject.org/mailman/listinfo/devl

Reply via email to