Weekly Planning Meeting, 9/25

[Marc Stiegler <marcs>]

Highlights 9/25:

The big congratulations this week go to michael, hill, dean, 
ravi, hugh, and rob. Michael has the snarf receiver working, 
a key milestone in getting to the point where we can write to 
disk. Dean is now reclaiming stamps off the disk correctly. Ravi 
has IPos's working, which is another critical part of the process 
of doing bert & stamp reclamation. And rob has engines working 
properly with elk.

Mr. Hill achieved an amazing milestone in the performance engineering 
of the backend. This week's 7 percent per day got him to the 
point where the C++ version of the server now runs faster than 
the smalltalk version. For some of us, this is quite a relief, 
having proof at last that C++ really CAN be more efficient than 
smalltalk :-) 

Hugh achieved a similar breakthrough: with the newest, fastest 
version of incremental text redisplay, you can't type faster 
than the screen displays--at least not if you're using a souped-up, 
top-of-the line Mac :-)

I guess the best news is that we still have so many performance 
improvements yet untapped; it seems increasingly likely that 
we will really be able to make xanadu fast enough to be useful 
(though we still have some opportunities for performance LOSS 
coming up, in the permissions system and the disk writing).

Congratulations also go to chris and andy. Andy has varargs fixed 
for elk, while Chris has insertions and deletions, extracted 
from the ent, represented in smalltalk. Now Chris has only the 
little matter of displaying them for the user :-)

Meetings: Hugh, andy, macElk. Roger, markm, backup vlad. Bill, 
hill, bug on comm. Ravi, roger, hugh, teach hill fm tests. Roland, 
ravi, michael, hill, binding com.ents in formic. Ravi, marcs, 
merge spire. Roland, michael, the return of a const parameter 
in C. Michael, dean, discuss resthome. Ravi, dean, garbage coliect 
canopies. Marcs, dean, time for bertstamp reclamation. Hugh, 
chris, comparison tools in fm, Thursday at 2. Michael, dean, 
second snarf packing meeting. Roger, ravi, where font info lies. 



Anyway, this week's task list is as follows:

A.       dean

     10/02/90     remember what doesn't work in signals
x    9/27/90     fluidly bound grandmap
x    9/27/90     stamps reclaim correctly
x    9/27/90     implement the destroyer upon reboot.
x    9/27/90     test stamp deletion tree walk
    10/02/90     berts gc done
    10/09/90     canopies gc correctly

B.       hill

x    9/19/90     help roland with stubble
    10/02/90     look at changeclass in zortech
    10/02/90     turn delete to destruct in smalltalk
     9/27/90     think about stepper for grandhash

C.       hugh

x    9/27/90     new selection object
x    9/27/90     test tools for permissions
    10/02/90     permission test differences display
    10/02/90     multiple runs per line

D.       markm

    10/02/90     absence of metaclubs means self meta
    10/02/90     testers on existing permissions
    10/02/90     message on overloading
    10/02/90     create new readclub with members null
    10/09/90     constructor bombs in smalltalk
    10/16/90     club desc add shared edit club loc, use everywhere, 
                       create shared edit club as needed
    10/22/90     create read club's shared edit club with gatekeeper's 
                         club as member
    10/30/90     fix nonclub berts to have the 2new berts the 
                       read club of the stamp share
    11/27/90     tests for clubs

E.       marcs

     10/09/90     design for minimal spire complete
     10/11/90     message on videotaping
     10/28/90     collect sqa notes
     10/28/90     start second round of patenting work
     10/30/89     figure out whether links are standalone objects

F.       michael

x    9/27/90     snarf receiver working
x    9/27/90     snarf transceiver working
    10/02/90     stubble methods for shepherd stubs
    10/02/90     interface to urdi
    10/09/90     viewhandler: connect to dean's stuff

G.       ravi

     9/27/90     enhance smalltalk to ignore groups of conflicts
x    9/27/90     ID, Ipos
x    9/27/90     generate an IObjectAsPosition coordinate space (called IPos's)
x    9/28/90     split uses of IDSpace to IDSpace vs. IPos spaces.
x    9/28/90     implement interning table for IPos's 
     10/02/90     spire design
     10/02/90     add wait for open to gatekeeper since sensors work (from markm list)

I.       roger

     10/02/90     merge hugh's stuff, ravi's whole application

J.       roland

     10/02/90     start writing translator for C++ to documentation 


K.       Bill

     10/02/90     entry receiver recipe

L.       Joel

     10/21/90     find replacement for bobp

M.       Chris

     9/27/90     insertions, deletions identified and ready
     10/02/90     figure out text comparison
     10/02/90     mail message, compile error with new proxies
     10/02/90     get move stored, ready for presentation

N.       Rick

     10/02/90     paper for carl

O.       Paul

x    9/18/90     fix garbage collectin with engines
     9/18/90     ld-r working
     9/18/90     trace command

P.       Andy


x    9/28/90     varargs fixed     
     10/06/90     Investigate SIOW for main tty window
     10/06/90     Port Rick's sunview interface

Q.       Rob

x    9/18/90     find and destroy bug in continuations

L.       Small Tasks List

    add newlines to the smalltalk transcripts
    Cleanup region code
    Sequence # in smalltalk, dean & michael versions
    Test files
    fix stubble for 1-way connections for dean


L.       Deferred Meetings:

     After First Dev Release, markm, chris, deletion of actual 
                object, propagate comm properly
     Markm, paul, rick, zstep debugger
     Roger, hugh, reproducible Tapestry tests.
     discuss stubble for ADI with autodesk
     Dean, michael, structured debug on the ent.
     Michael, dean, markm, ravi, efficiency of internsets.
     Roger,, hugh, dean, ravi discuss system independent attributes
               for the frontends  after fm does paragraphs.
     Roland, dean, discuss xlint commands in smalltalk.
     Markm, hill, discuss explicit deletion of orgls & refcount deletion.
     Markm, hill, discuss exception handling on sensors between comm handlers.
     Dean, roger, michael, markm, discuss sequence numbers of size 4 billion.
     Dean, markm, discuss intersecting backfollow ideas.
     Ravi, roger, discuss common view.
     Hill, markm, matching on underscore.
     Markm, dean, does not propagating unreadable clubs optimization
              overlap delayed installation in ent?
     Michael, dean, markm, LRU on RAM objects (snarfs, flocks, etc).

M.       Unassigned

     merge contact databases
     scenario documents (markm, dean
     tutorial for users on the frontend
     reference manual for users on the frontend
     test scaffolding, code exercisers

--marcs