[cap-talk] Can we stop discussing Butler Lampson's opinions now?

[Jed Donnelley]

On 4/10/2008 9:13 AM, David-Sarah Hopwood wrote:
...
> ...Frankly, I don't see what this cult of personality is around Lampson
> and his opinions. "He's just zis guy, you know?"

I brought his views up to begin with because, while he is only
"zis guy" as you say, he is "zis guy" who:

1.  Is regularly invited to give Keynote and invited talks at
well attended conferences and at computer science departments
around the country (world?) - e.g. the repeat of his "Gold
and Fool's Gold" talk as noted being given again:

Joint meeting of Boston/Central New England Chapter of IEEE
Computer  Society and GBC/ACM
Thursday, 4/17/2008, 7-9 pm
MIT Room 54-100

and he:

2.  Takes every opportunity that is presented to him to
absolutely tear into POLP and in particular capabilities.

He argues that POLP (POLA) never has worked and never
can work.  He furthermore argues that capabilities
have never "pulled their weight" (e.g. his Cambridge
Cap example) and that they never can.  And finally
he says (at 8:55 for Sam Mason in:
http://www.usenix.org/events/usenix06/tech/mp3/lampson.mp3 )

"Things that looked very good and into which a lot
of effort was put that have been failures.  I hope
there are people in the room that are working on
some of these things and that I will be successful in
dissuading them from continuing to work on these
things.  Some of these things might possibly work
at some time in the future, but I think the weight
of the evidence is against it.

Capabilities seemed like a wonderful notion ..."
(continued in:
http://www.eros-os.org/pipermail/cap-talk/2008-April/010843.html
)
where the top of his hit list is capabilities.

and

3.  While Butler is just "zis guy", he is widely
respected and not isolated.  This is part of the
reason he's invited to give so many talks.  He has
a large number of colleagues who respect his opinions
and in particular consider his opinions on POLP and
capabilities as fairly well established "mainstream."


I consider the above a problem for anybody working
on POLA with capabilities, e.g. the Fred Spiessens
of the world.  Why should people have to start working
in this area from a position of bucking the status
quo?

I agree there's little point discussing it further
on cap-talk.  I think the valid reason for discussing
it to begin with is to have our arguments tuned
the next time we run into somebody who was influenced
by Butler or by the Lampson "school".  For example, I
think the discussion of the results of the Cambridge
Cap work is important.  I also think it interesting
and important that Butler doesn't attack capabilities
because of their lack of "control" over delegation
(what we've recently called the "capability propagation
control myth").  I wonder if that means he understands
the issues with communicating conspirators and the
necessary role of confinement in blocking delegation?

He's zis bright guy, zis learned guy, zis visible and
respected guy, and because he comes with a whole
collection of colleagues and like minded people
he isn't quite 'just' "zis guy."  Unfortunately he
strongly believes that work on POLA/capabilities is
counter productive and worse than a waste of time.
And he won't answer my email ;-)  Maybe I'll get
to attend one of his presentation some day.

The first time I hear a keynote or even an invited
talk at a computer security conference where somebody
says something to the effect that, "There are those
who've thought that POLA/capabilities are counter
productive, but recent work in this area has shown
that they can have significant value in a number of
areas such as providing more flexibly controlled network
sharing, improving system integrity and security, and
generally providing discipline for object oriented
programming."

(pick your own words, any positive mention will
do) then I will feel that capabilities have "arrived"
(or perhaps re arrived)

<might there have been such a talk already?  Perhaps
David Wagner?  We seem to be getting closer.>

and that we can stop worrying about how to combat
the Butler Lampson's of the world blocking work in
the area with their "haven't worked and can never
work" school of thought and dissuading people from
working in the area, and we can just get on with
the work of applying capabilities where they can
provide the positive value.  There may be something of
a generational issue to this, but I believe people
in this area, even Butler Lampson and his school, are
subject to demonstration and engagement.  Even though
he considers capabilities as an area where he has done
"serious" work, his thoughts on the topic seem to me
dated, inaccurate, and misguided.  At some point I
think they need to be faced head on.

In the mean time, I'm done with Lampson for this round,
certainly on cap-talk.

--Jed  http://www.webstart.com/jed/