Friday, November 28, 2008

startup midwifery and RAI*

i seem to find my life interleaved with helping new spinouts get going without being actually a founder or shareholder (number 12 yesterday - more news later) - its fun, but not-for-profit:-)

meanwhile thinking about RAI* (Redundant Arrays of Independent Stuff):-
An alphabet soup of these leads to some possibly neat ideas such as

Redundant Array of Independent Architectures = Virtualisation+Emulation
RAI Bytes - well this is just FEC
RAI Computers = Derek Murray/Steve Hand's Spread Spectrum Computing
RAI Disks (the original one:)
RAI Electricity (power supply backup/battery/diesel/solar etc)
RAI Frames = layered coded video
RAI Gods = polytheism (even more playing safe than Agnosticism:)
RAI Help = Google
RAI Internets = Virtualise the Internet
RAI Jobs = academics:)
RAI Kludges = s/w reliability technique used in aerospace quite a lot
RAI Links = Multipath
RAI Memory - obviously
RAI NOtifications - what you get from signing up to too many social net sites
RAI Objects = Eternity File System
RAI Projectors - coping with speakers who have Macs or Windows or Linux laptops
RAI Queues - being english
RAI Redundancy = Recursion - or maybe redundant array of indepenent recursions
RAI Security - what we really have - insecurity
RAI Testtubes = life sciences:)
RAI Users - Sys Admin view of the world
RAI Virtualisation = see recursion
RAI WOrk - see Academics
RAI Xen - 4.0, 5.0 etc
RAI Years - life, pain in diodes, left hand side, etc
RAI Zeros = fault tolerant implementation of /dev/null

we can replace Independent with Guardian, Times, Express

Redundant Array of Guardian Actions
for example...even makes sense and is a groovy Indian musical

Wednesday, November 26, 2008

Great talk by Andrew Odlyzko here today

A talk by Andrew Oslyzko is always worth seeing/listening to! Indeed he posed this very question to the highly engaged audience - would we have attended if there were no slides/pictures? Answer, 100% yes. And would we attend if noone was allowed to speak (including him)? answer almost 100% no.

Several things I thought worth noting...

internet growth has slowed (from 100% to 50% or so - although some agencies still have it a bit higher) _ my take on this is that the internet growth is a series of step functions, and we've now got to the point where the only way to go up is a Very Big Step (i.e. fiber to the home) - my view (I put this to the speaker) is that a really worthy New deal project for the EU and US to do to invest our way out of recession would be to incentivise a set of parties to fix all the national utilities (e.g. 30% of UK water is lost in leaks) by employing a bunch of building industry people who are rapidly getting unemployed otherwise, to go dig up roads and put in new pipes including new fiber

the speaker claimed that there's little growth potential in the fixed internet - this may be true (modulo FTTH) but there's serious growth potential in 3G (High SPeed Packet)...

he claimed cloud computing doesn't really add up - his figures depend on someone uploading their entire disk - i think this is a bit misleading (his figures - a 300G drive, being uploaded over a 300kbps uplink takes 3 months) - in reality, IP backup is incremental - i don't back up the drive - jsut the deltas - over the lifetime of the machine (in my house, about 2-3 years) that might be a bit over 100% of the whole drive - no problem there then.

his coolest point was video download - everyone talks about streaming - andrew asked
"how many people want faster than real time video" - around 30% of the Cambridge audience said yes! this is unusual - most people don't get it - but the reality (e.g. a video flash player from youtube shows this in the grayed out bar ahead of the current play point) that a lot of the time you do get faster than streaming - at least 2 major reasons this is good
i) its essentially increasing the number of overall costumers (conservation law and work) and so you get a bunch more videos on average -
ii) an extreme case i that i am about to go on a trip and want to download several movies to my player - clearly i dont want to stream them - i want them ASAP!

I pointed out this is just the inverse of the TiVo time shifting case (or as Sky put it pause (and rewind) "live" TV - this is basically
Fast Forward (name of a startup from Berkeley a while back, who were too far ahead of their time:)

Last but not least - it isn't the sustained rate you want on your access link - it is the peak you get for bursts and the lower latency for those bursts - for human impatience (read, Game Players) this is key....

I couldn't agree more!!!

Tuesday, November 25, 2008

today's puzzles

1. what is the maximum _width_ of a mobius strip?

2. why is my life pants? - women keep saying what sort of underware I should port. How is that fair?

3, if I was to say to you "eldritch", and "nameless horror" you would probably cry
"H.P.Lovecraft". But what about a band name generator based on the name?

We have "sauce type" "emotion/profession" and "vehicle"

so you could have
ketchup witchcraft
mustard loveboat
mint loveseat
brown likecar .... <- fail :)

4. Finally, (for today), is
transcendental physics a category error
like the God Delusion (or the Dog Allusion)?

Monday, November 24, 2008

accelerationistas! be careful what you wish for

So its my belief that the current economic woes of the west can be blamed on the accelerationistas! of the Internet era - basically, the old fogeys who "run" things in the financial service world, around the time of Clinton and Blair, actually believed what they were being told by the mad people who said - economics is not a zero sum game (c.f. Clifford Stross's novel, Accelerando, and many writings of people who read (but don't understand) Cory Doctorow, Lawrence Lessig, and the madder moments of the net e.g. Agassi's free cars)

So the spin gained traction but had no torque.

Especially with the .com fiasco at the end of the last boom....although there's evidence of cell phones and computers and even some drugs being close to free

But then
the governments of the world thought they'd try it in the realm of finance - now as far as I am concerned, the governments of the world look after my desiderata - as an English person, this is my home (castle), health, and shopping (food/consumables) and (I suppose) transport and heating/lighting - i want them defended - this means both against invaders real, and invaders economic -

Homes, food and transport/power/fuel are basically very a close to a zero sum game, and the price system should really be a convex optimisation problem - this is almost certainly precisely soluble through central planning on a Very Big Computer, or through a decentralised system such as a free market -
this all depends on people running free markets in these things unconstrained, but they don't.

Unfortunately, oil isn't a free market (the price is set in dollars which favours US businesses energy costs and set by a cabal called OPEC). food isnt a free market (the EU and US both conspire to run protectionist worlds).

On the other hand, where the whole non-zero sum game does operate (hi-tech industries such as computers, networks and bio-tech) is where there shouldn't be a normal market - indeed, there should be NO IPR either. basically,
entertainment (fulfillment and well-being) and health (longevity and well-being)
are aribtrarily consumable and so innovation can spiral.

But food, energy, housing - these are not amenable to the accelerationist dream,
and that's I guess what the world just found out.

Economics 2.1 alpha should include this concept

(yes, yes I know when we have nano-tech, then these other bastions of traditional scarce resource allocation may too fall before the burgeoning "technosaur").

So what I am saying is that the spiral (on house loans exceeding 2nd order derivative of the house price increase) is a simple lesson in controls on resources whose dynamics are limited by traditional physics. But that shouldn't mean we abandon the cool madness of hyperinflation on valuation of new tech work - just that we separate these two regimes until later....much later...

Friday, November 21, 2008

damn - got me in one, again

how not to keep your job

too much too young - flaws in early internet research

so looking back at some early (not as early say as the arpanet, but early compared to web 2.0) internet research, it looks to me like it is really worth re-vidsting some of the "well known" results from between (say) 1989 and 1999 (i.e. the decade before the last 10 years or so).

I can see several "well known" results (a.k.a. folk knowledge/science) that are probably wrong - often, these are architectural in scope, and are because someone (often from a well known institution) wrote a preliminary paper on a prototype (lets not mention names, but early RSVP code or IPv6 spring to mind) and drew conclusions which were taken, for want of a better word, as seminal - so some comments on scalability (e.g. of intserv and rsvp) and security (e.g. of 8+8), put people off working further in promising directions, and were not at all correct - we don't know what the right answer was because no-one has re-done the work in the context of later knowledge e.g of switch router design, or of crypto-assigned addresses, or of hardware support for fancy queueing and scheduling - there are lots of examples - the problem historically (and for the community and services and products) is that with an exponentially growing network, the tipping point is past, at least as far as today's IPv4 internet is concerned. But I worry that newcomers building new systems (e.g. IPv6 internet in china with Huawei) may take the old research as correct when it isnt.

Saturday, November 15, 2008

gender bender
reports that I (who write that blog) and probably a woman...
on the other hand
says this is by a man,
as is
my other blog - so right 2/3 of the time...

Friday, November 14, 2008

self plagiarism in code...and ip spoofing

has anyone ever run a plagiarism detecter (there are lots) on (say) the linux source tree? it'd be interesting to see how much redundant code there is:-)

oh so with ip spoofing - can we ask someone near by a source to attest that they have seen the host we are speaking to (i.e. they attest to seeing a packet with a low rtt)?
as a cheap and cheerful prevention of wide area ip soofing bots?

living in a material world

to quote someone who isn't a physicist,
we are all living in a material world,
and in that world we are limited to light cones
within which there are _causality_ principles
(as well as entropic) - hence relative locations have
a bearing on infrmation flow between objects

the ordering of messages on this mail list
(and the rate of increase of entropy in the universe)
Is bounded at least in part by latency
latency shows up quite a lot in distributed algorithms (routing,
location/mapping, memory/cpu speed/power, BFT schemes, etc etc:)

the internet to date lives in a virtual world and data oriented
mnetowkring likes to fool itself it is in some steady state in an
"infosphere" where we can plonk down copies of data willynilly, and
therefore see no latency.

reality bites - a lot more data goes in from, and even more usefully,
out to the real world. the points it goes in and out (ingress/egresss_
better be near people/devices/sensors/actuators that want the data

since some data ingresses from the real world (my typing) and out (you
reading on the screen >-here<-)
we better route messages on paths that don't spend too long going
arond the 13 dimensional infosphere too many times before they get

oh, and the folks "out there" (i.e. the 4billion cellphones) have a
better handle on how to do this than the folks "in here (the mere
1billion internet backwoodspeople)....

time to get more real...which is why i started
this thread on e2e which got led astray by some philosopher/wizard/physics/guru peeps:).

indeed if we think of this as an Onion
with RL as an outer layer, and SL as an inner layer
(and as many people have been doing, mining
interactions between RL and SL via footprints in the sand
in the SL) then one consideration about how one
creates mappings from id (in RL) to location (in RL/SL interface)
might be privacy, so the onion immediately brings to mind

of course, if all you are interested is SL, then let me onion
route you this:

Wednesday, November 12, 2008

brainy bumbershutes or the sentient umbrella

in this age where we have free bikes available in all intelligent (civilized) cities (e.g. Paris, Copenhagen, Amsterdam), and we are busy outfitting them with wireless sensors to report where they are, when, and what the weather is like (skidding, lights) and pollution (air/noise), we should note that these said same cities also haveunreliable humidity - so we need a citywide free umbrella system - the umbrellas could be kitted out with torches in the handles - these torches could be powered by batteries which contain sensors (so that we can find an umbrella)

a sentient umbrella city app on an [ig]-phone would pop up and buzz when the weather in the owners locale was about to turn rainy, and tell them "100 meters down the street on the right is a city-free umbrella - you can see it as I have just start ed flashing its light...and warned other people that this one is now yours for the next 1km/hour... ... ...

[if anyone can find the origin of the word "bumbershute" as an alternative for umbrella, please let me or tim griffin know asap:-)

automating creativity by application of ancient greek prefixes

It's not all Greek to me
automating creativity by application of ancient greek prefixes

Whenever you have an interesting problem, you hold it in your hands and look
at it, turn it around, look underneat (see if it has a makers marque),
peer inside, knock on it with your knucles and so on.

One thing you can do is to try and make a variation of the problem, and then
see if the variant is easier to solve....

The Greeks have a "system" of prefixes for words, which modify the meaning, in
a systematic way, to explore all the alternative "views" of that meaning.
Other languages are less systematic (obviously, English, being effectively
"panglossian", has lots of prefixes for all these including the Greek ones,
but then the most common one might be somethign else (even, ugh, Latin:)

So here's a (not comprehensive) list - think of them as functors, or
even illocutionary acts....

eu-meta-lateral thinking, so to speak

an - not

amphi- both

ana- back

anti- against

apo- away from

dia-, di- across

dys- ill, (dystopia v. utopia)

ex- out

ecto- on the outside (ectoplasm:-)

en-, em- in

endo- within

eso- inward

exo- outward

epi- upon

eu- well

kata- down

meta-, among/between

palin-, back again

para-, beside

peri- around

pro- before,

pros- to

syn-, together

hyper- above

hypo-, under

Monday, November 10, 2008

© God

so lets say creationists are right and intelligent design is the order of the day.
then, quite simply, patenting genes is obviously wrong, since they are derived products of the intelligent designers original. Hence all well and good.

but lets say darwin is right. then they are evolved and you can't patent them then either

nor can google

© 0000 God

Friday, November 07, 2008

poisoning p2p nets - does the RIAA pay for the copies?

so if some agency poisons a p2p file sharing system, e.g. to try to limit copyright theft,, presumably the use some sort of degraded version of the actual file (else it would be trivial to detect that they just have a completely incorrect check on the file, surely)

but if they use something that has even some of the original copyrighted material, it is a derived work

so do they pay the artists?

if not, aren't they just as bad (worse:)???

Sunday, November 02, 2008

hard sf fashion victims prisoners of history

i've noted before that GUI toolkit designers betray their age since the look and feel of given guis (think windows 98,nt,xp,vista, MacOS in various guises, or tcl/tk or X10 with athena widgets, xerox smalltalk browsers, etc etc,)
all look like hifi equipment from different eras - my hypothesis is that the GUI has the fashion of the designer's first hifi (60s clunk and wood, seventies brushed aluminium, eighties matt black, etc etc - same for knobs/buttons etc)

so in SF we see the same thing - when authors describe future scenes (in life, bars, clubs, etc) clothes, music, and so on all reflect the authors' teenage years - whether its 60s fun, or 70s glam (star wars), 80s punk/newave (stepheson/thrash skateboarder music), 90s goth (accelerando) etc etc

it would be nice if just for once, someone in the tech/geek/futurist world had the imagination and wit to do something less retro (as well as maybe, just for once pretty please, having characters and plot that don't reek of derivista).

Blog Archive

About Me

My photo
misery me, there is a floccipaucinihilipilification (*) of chronsynclastic infundibuli in these parts and I must therefore refer you to frank zappa instead, and go home