[Gzz-commits] journals benja

[Benja Fallenstein]

CVSROOT:        /cvsroot/fenfire
Module name:    journals
Changes by:     Benja Fallenstein <address@hidden>      03/05/16 23:15:51

Modified files:
        .              : benja 

Log message:
        stop working *yawn*

CVSWeb URLs:
http://savannah.gnu.org/cgi-bin/viewcvs/fenfire/journals/benja.diff?tr1=1.35&tr2=1.36&r1=text&r2=text

Patches:
Index: journals/benja
diff -u journals/benja:1.35 journals/benja:1.36
--- journals/benja:1.35 Fri May 16 12:24:46 2003
+++ journals/benja      Fri May 16 23:15:51 2003
@@ -45,10 +45,45 @@
 Past:
 -----
 
-2003-05-16 [5.5h, 13.00-18.30]:
+2003-05-16 [14.5h, 13.00-20.30; 22.00-05.00]:
 
 - work towards making libvob and fenfire work on Kaffe
 - submit patches to Kaffe
+- more thinking about vobs, trying to come up with
+  way to get started on structured graphics vobs
+- learn about XML signature: can we come up with a way
+  to use one-time signatures with XML sig? (register
+  our own algorithm) -- actually doesn't require 
+  registering: XML sig uses URIs! There are free
+  implementations of XML sig; however, we may consider
+  XML sig as too heavyweight (need to think).
+  We could also say: only RDF blocks are signed
+  (for RDF, meaning is more or less well-defined);
+  so the sig block simply gives the hash of the RDF block.
+  -- It's interesting that there is no standard
+  one-time signature algorithm-- I mean there are algorithms,
+  but abstract descriptions, not on the level that specifies 
+  which exact bytes the output is for a given input.
+- Think about timestamping; to provide non-repudability
+  (one can't say one signed a given block) one would need
+  to be able to prove that a block was signed before
+  the key was revoked? Otherwise to say you didn't sign
+  a given block, you'd just need to revoke the key that
+  signed it... and claim that the signature was created
+  after you've revoked the key. However, non-repudability
+  is not a central goal of Storm; it may be ok if it is
+  only archievable if a trusted third party timestamping
+  service is available.
+- Figure out how at the expense of some more space & time,
+  one-time signatures can be made any-time; need to think
+  about whether that is useful... :-) [may not be new,
+  but if it is, may be publishable]
+- Note that there are *many* introductions explaining that
+  one-time signatures can only be used one or a limited
+  number of times. Therefore, it seems this is new.
+- Start writing this up. Notice that ACM CCS deadline
+  is tonight and make brave-hearted but folly-ish attempt
+  to write an article. ;-) Give up.
 
 2003-05-15 [8h, 14.00-24.00 with breaks]: