RE: Optimism [Was: flame wars]

From: Emlyn O'regan (oregan.emlyn@healthsolve.com.au)
Date: Thu Jul 17 2003 - 21:43:39 MDT

  • Next message: Spike: "RE: flame wars"

    Harvey wrote:
    > I wish more people understood this. Moore's law IS a marketing ploy.
    > Are there any real engineers that quote or use this "law", or even
    > believe it qualifies as a "law"?

    It's more of a social observation than a law. However, isn't it still
    holding true, in some rough sense? Don't hardware designers take it into
    account when balancing increase in capabilities with time to market? I've
    always though that it was a very interesting self-supporting meme; the more
    people believe it, the more real it is, and so the more people are forced to
    believe (and thus continue to support it). It reminds me of the
    self-reinforcing effects of confidence and of pessimism in a market, but
    seemingly without the dualism of those two (and so without the bust to
    offset the boom).

    > Another interesting parallel to technology here. Abstraction
    > is a good
    > tool, but in practice it has lead too many people to design bloated,
    > inefficient, unreliable software.

    Off topic and aside, I'll take issue with this point as too one-sided. The
    other side of the coin is that without abstraction, we'd have so little
    software, compared to what we do have today, that we'd be very much poorer.
    Also, the abilities of that software would be miniscule; no distributed
    enterprise apps talking to multiple datasources over many layers of
    protocol, heterogenous networks, on many and varied operating systems on
    varying hardware with wildly different capabilities. Gone are the days of
    understanding the flow of every last byte of information. And that's not so
    bad... for a start, you get a job for life, Harvey :-)

    > Similarly, I think our ideas about
    > technology have been abstracted to the point that we produce bloated,
    > inefficient and unreliable expectations.
    >
    > > This, at least, is what
    > > I can salvage from my side of those copy arguments we used to
    > > have. (I concede.) And finally, we shovel all the real
    > > problems under that carpet we call "software." This is part
    > > of that larger Myth of the Computer Age: universality. We
    > > can do anything in software, given enough speed. This is not
    > > true in any practical or useful sense, however.
    >
    > Agreed! The myth I keep running into is the seven-layer model of
    > networking. It says that layers are abstracted independently of each
    > other so that the physical, data-link, network, transport, session,
    > presentation, application layers can all
    > mix-and-match/plug-and-play/and
    > be vendor neutral. It sounds good in theory, but it isn't
    > real. If it
    > were real, we should be able to run a Linux program displaying IBM
    > EBCDIC characters over an http web session via a non-TCP
    > protocol on an
    > appletalk network running Microsoft netbios on an ATM link!
    >

    Maybe. Does the layer model say that each model is entirely independant of
    each other? Or the truer view, that each layer is based entirely on the
    layer below? And of course there is the realistic addition to this, that the
    idiosyncracies of lower layers bleed through to higher ones when we would
    rather they didn't. Still, I seem to be able to send email from anywhere to
    anywhere, oblivious to the software+hardware used by others (and even mostly
    oblivious to what I am personally using!)

    > > I say the following with complete confidence: there will be
    > > no Yudkowskian Singularity, the copy is not the original, the
    > > creation of the first assembler will not cause an immediate
    > > revolution in manufacturing. These are science fiction
    > > pipedreams.
    >
    > I believe that these things will happen eventually. But they will be
    > carved out with great difficulty at every step of the process.
    > Technology gets more complicated as it grows exponentially.
    > Layers upon
    > layers of abstraction make system unpredictable. Complexities make
    > seeing the complete picture more and more impossible. As technology
    > solves our problems, we are presented with more numerous and more
    > difficult problems. Like a video game where you never win,
    > we just keep
    > advancing to more and more difficult levels. The work gets
    > harder, the
    > rewards get bigger, the disasters get worse, and the process repeats
    > itself over and over. There is no end to this process, and
    > I'm not sure
    > why so many people on the Luddite side and on the technology
    > side think
    > there is.

    You are dead on about this progression; it's why I've said before that I
    think software development is far more difficult now than it has ever been,
    although our tools and our abstractions race along keeping us barely above
    water (maybe). We are getting to the point, I think, where the increasing
    number of abstraction layers (and interactiosn between orthogonal systems of
    abstraction) is starting to require meta-abstractions to group and hide many
    layers at once, in new ways. Otherwise, the software crunch that we've been
    staving off (quite well!) for so long is going to bite us eventually.

    Also, and interestingly, this is one of the major arguments about
    "progress"... that it is really just very expensive wheel spinning, and that
    socially, we impose ever increasing burdens on ourselves for no useful
    offsetting result. That's one of the reasons people look to technological
    singularity in my opinion; an answer to this criticism of us becoming
    increasingly busy to no effect is the promise of a discontinuity, which will
    actually make all the work leading up to it worthwhile, after all.

    Personally, I feel that the improvements we make, the increases in
    complexity of software, the crazy new challenges we face, are definitely
    worth it. We get little discontinuities all the time; I'm sending this email
    over one of them. The world really has been getting measurably better. The
    next one I'm hanging out for is ubiquitous wireless internet access, with
    large market penetration on GPS (or similar) enabled devices. That is,
    people themselves (rather than desktop PCs) begin to go online in a big way.

    >
    > > Yes, and as well as realising that this stuff requires hard
    > > work, we need a *critical* approach to technology. We need
    > > to take our heads out of the sand, lose the ridiculous
    > > "luddite" talk, and realise that, yes, technology does effect
    > > peoples lives, and that, no, not all technology effects all
    > > lives in a positive way. Technophilia doesn't cut it.
    >
    > Agreed! I have come to believe that the Luddites are right in their
    > fears. This stuff is dangerous and will cause problems. However, I
    > don't believe they are right in their solutions. We can't just ignore
    > the stuff or get everybody to agree to relinquish any use of
    > technology.
    > We must strive to make sure that safety and security are
    > included in all
    > technology. We need to make sure that dangerous technology doesn't
    > accidentally destroy the world just as we have to make sure that
    > dangerous terrorists don't deliberately destroy the world. This takes
    > auditing, accountability, transparency, negotiation and a whole lot of
    > attitude adjustment. People who risk massive disasters without taking
    > proper precautions are like terrorists. They risk other
    > people's lives
    > to achieve their own goals. They don't care if other people die if it
    > will help them reach their own goals. There must be a way to
    > work with
    > Luddites, identify real problems, and then develop solutions
    > to address
    > these concerns.

    Ooh, you are advocating the complex, intricate, middle ground solution,
    rather than "Everything's going to be great" vs "Everything's going to fall
    into crap". There's a hard ask. It's a good path for transhumanism to walk,
    though.

    >
    > > Technology is ideological in the strongest sense.
    >
    > You make a lot of excellent points that I agree with. Mike Lorrey is
    > also making other points that I agree with. Anders said it
    > best when he
    > said we are in a phase of "growing up". Our college-days of idealism
    > are gone. We are in the real world now. Things aren't as rosy as we
    > thought. We live in a Dilbertesque world, and there are too many
    > pointy-haired bosses calling the shots who don't understand what they
    > are talking about. We want to make a difference, and still want to
    > pursue our dreams. But we have to adapt to the real world
    > and the real
    > future. The future isn't turning out to be what we expected it to be.
    >
    > --
    > Harvey Newstrom, CISM, CISSP, IAM, IBMCP, GSEC
    > Certified InfoSec Manager, Certified IS Security Pro, NSA-certified
    > InfoSec Assessor, IBM-certified Security Consultant, SANS-cert GSEC
    > <HarveyNewstrom.com> <Newstaff.com>

    Geez, everyone is so down about this stuff. Admittedly, things aren't up the
    way they were. However, given that the world is at a real low point, the
    fact that it is still difficult to be anything other than at least mildly
    optimistic (in my opinion) about the future of the world (particularly in
    relation to technology over the coming decades, and the transhuman vision),
    is something to cheer about.

    Excepting, of course, if this isn't as bad as it gets...

    Emlyn



    This archive was generated by hypermail 2.1.5 : Thu Jul 17 2003 - 21:53:22 MDT