[AFS3-std] rxk5 - summary of hackathon issues

[Marcus Watts]

I'm glad Simon entered some items about rxgk.  He reminds me
that I owe some similar things for rxk5, starting with a list of them.
Here is that list: of the items that came out of the hackathon for rxk5;
some of these are in part common with rxgk.  I will post separate
more complete messages on each of these following, so please don't
use this message to start discussing these!  I will stagger them
so as not to defocus discussion from simon's messages too much.

In the following, I'm going to add 3 things:

risk
	0 = much safer than falling off a log
	4 = should be safe, but needs care to do right.
	9 = hopelessly insecure/liable to breakage
difficulty
	0 = no work at all.
	4 = "one week" project
	9 = moon mission
inv-util
	0 = it slices, dices, feeds the cat, and takes the paper in.
	4 = meh.
	9 = wallpaper is more useful.
	should be utility, but I'm inverting this to better match
	the other scales.

all 3 scales are purely my subjective flash judgement based on my
preceived impression of the current state of afs.  You might well
assign different values.  I might too, tomorrow.

1. authenticator max_calls 4
	risk: 1.  difficulty: 1.  inv-util: 7	[ cf rxgk ]
2. prf rfc 4402
	risk: 3.  difficulty: 4.  inv-util: 3
3. initial packet encryption
	risk: 7.  difficulty: 7.  inv-util: 5	[ cf rxgk ]
4. k5ssl
	[ this is purely an openafs software issue: will post in afs-devel ]
	risk: 3.  difficulty: 1.  uselesness: 4		[ sxw>rxgk ]
5. ubik_SRXServerProcV2 and afsconf_ServerAuthV2
	[ this is purely an openafs software issue: will post in afs-devel ]
	risk: 1.  difficulty: 1.  inv-util: 2	[ also rxgk? ]
6. cell_max 256
	risk: 1.  difficulty: 1.  inv-util: 2	[ also rxgk? ]
7. cm properties
	risk: 2.  difficulty: 3.  inv-util: 6	[ ?? rxgk ]
8. way for app to add data to authenticator
	risk: 2.  difficulty: 3.  inv-util: 2
[ This was mentioned at the hackathon, but wasn't on the original list of 8.
[ However, at the european afs workshop, it became clear this was going
[ to be a roadblock:
9. one file server at a time conversion process
	risk: 5.  difficulty: 5.  inv-util: 4

( about 6 of the above items are on the "we'll probably do them" list,
so that means at least some of the above should be pretty perfunctory
in terms of discussion. )

Again, I'll post more detailed notes on each: so please
wait for that before telling me why you would pick
different numbers, or anything else about these.
Useful followup messages to this: additions or changes
to this list of 9 items.  Or, if you think my numbers are
way off: your own set of 27 ranking numbers.

					-Marcus Watts