On Fri, Oct 19, 2012 at 09:30:52PM +0700,
 Ivan Shmakov <[email protected]> wrote 
 a message of 60 lines which said:

>       It simply doesn't seem quite right to me to use an "authority"
>       for a content-derived identifier.  

It exists (but is optional) in the future ni: scheme. I say future
because the standard is approved but not yet published as a RFC
because it depends on other documents of the HTTPbis crowd, which are
not ready.

ni (Named Information) is very close to what many people on this list
are doing since the value in the URI is a hash of the content. You can
_optionnally_ add an authority and, in this case, it is the name of a
server you can use to retrieve the content (but the important part is
the hash).

So, <ni:///sha256;ZAywWlRKRClpbwpcHeNDo0PMFzhqPhbmDGvlwy9BiIg> is the
URI, the variant with an explicit authority is
<ni://www.bortzmeyer.org/sha256;ZAywWlRKRClpbwpcHeNDo0PMFzhqPhbmDGvlwy9BiIg>
and this variant can be automatically transformed in the standard
http: URL
<http://www.bortzmeyer.org/.well-known/ni/sha256/ZAywWlRKRClpbwpcHeNDo0PMFzhqPhbmDGvlwy9BiIg>
which is retrievable (and you can check the hash if you want, but do
so on the XML source, not on the HTML). So,
with ni: URI, you can have your cake (an address based on the content)
and eat it (retrieve the content from an actual server).

You can read the future RFC in
<http://www.rfc-editor.org/internet-drafts/draft-farrell-decade-ni-10.txt>,
I think it is very interesting for people here.

_______________________________________________
p2p-hackers mailing list
[email protected]
http://lists.zooko.com/mailman/listinfo/p2p-hackers

Reply via email to