Texas comes into the 21 century

TI has been giving me good service since before s.e.d. existed.

Problem? All the stuff I've designed lately works.

John

ou

=A0So

The Fall is one of those after-the-fact explanations that raises more questions than it explains.

-- Bill Sloman, Nijmegen

But they've got about the same level of scientific credibility.

John Larkin hasn't yet got it into his head that any single instance of the genome doesn't know anything about any other genome, which makes it rather difficult for it do anything planned or systematic. The genome doesn't include any documentation of its modification history - the system was set up before ISO9000 came into force.

-- Bill Sloman, Nijmegen

Maybe we are the evidence.

John

You make this stuff up. Genes switch one another on and off, transpose, repair themselves, modify one another, cooperate to synthesize stuff. The complexity is barely understood, and every revelation is yet more complex.

The simple Darwinian, and neo-Darwinian models, are hopelessly naiive.

I pity people who refuse to be astonished.

John

But the basics are now pretty clear. Some patterns of RNA are capable of catalysing their own reproduction when in the right brew of raw materials. We now know that most of the required raw materials are more common on the early Earth than had previously been thought. Liquid water in a reducing atmosphere may well be the critical requirement.

And the first self replicating autocatalyst gets to win the race since it rapidly mops up the pool of available resources. Then you are down to copying errors, photo degradation, bits breaking off and mutations to generate diversity and the game begins again with a new pool of more complex components some of which are already part way there.

We accidentally created a new form of self replicating pseudo-life with the BSE protein folding problem in cattle turned into cannibals. Comparatively simple molecules can have surprisingly complex behaviour - particularly if they are catalysts.

It is probably no coincidence that DNA/RNA using four bases is the quantum equivalent of a binary code. That is you get a match in a four way quantum comparison in exactly one attempt. The same is also true of the mapping of 3 base comparisons to ~20 amino acids which is again on the physical limitation of quantum mechanical comparisons.

They seem to work pretty well though and on about the right timescales in terms of number of generations required for a major improvement.

Amongst other things that have evolved are retroviruses that snip bits of other creatures DNA and put them elsewhere. It is these molecular scissors that have been sequestered to execute controlled GM research.

I don't doubt that there are still parts of the genome presently labelled "Junk DNA" that have yet to be understood by science, but I do not see any reason why further progress cannot be made.

Remember that genetic programming and closely related simulated annealling algorithms are now used in computing to solve approximately otherwise intractable problems of global optimisation. The "solution" found this way may not be the perfect global optimum but it is good enough for all practical purposes more often than not.

You could cite as evidence for this the fact that the optic nerve of the human eye creates a blind spot in the main field of view. What "intelligent designer" would ever do that? Dumb as a rock.

There is a big difference being astonished and throwing hands up in the air and explaining everything by saying "Goddidit". The latter approach is neither rigorous or intellectually satisfying.

Regards, Martin Brown

's

The - rather remote - possibility can't be excluded, but life on earth all seems to use much the same molecular machinery, so if anyone did seed the planet, it would seem to have been with single-celled organisms around a billion years ago, so your hypothetical intervention would be remarkably hands-off, more along the lines of environmental pollution than an ecological upgrade.

-- Bill Sloman, Nijmegen

Sure. But any system engineer should be able to recognise that if a system is going to do intelligent design, it has to have some idea of where its been, and that information simply isn't available in the genome.

Hand-waving baffle-gab about how complicated the system is doesn't get around the basic point that evolution works by changing things at random and throwing away the changes that don't work, without making any kind of record of what the unsuccessful changes were.

Jerry Fodor's book "What Darwin got wrong"

makes much the same - valid - point, then fall flat by claiming that there hasn't been enough time for evolution by selection of random variation to create the variety of life that we see, which is essentially the creationist "we don't understand how it happened, so it can't possibly have happened" argument which the creationists famously - and foolishly - applied to the evolution of the eye.

Jerry Foder and Massimo Piattelli Palmarini have enough sense not revive the argument about structures and functions which fossilise as well as eyes and proto-eyes, but they don't have enough sense to realise that an argument based on ignorance isn't actually decisive or even persuasive.

One can be astonished about what evolution has created without feeling any necessity to make implausible claims about how it might work.

-- Bill Sloman, Nijmegen

And so - after about an hours or so's fiddling around - did the circuit where the TLC2201's inadequate data sheet shafted me. Most circuits take some fiddling to get them to do exactly what you want, and if you weren't in the habit of reading data sheets carefully you might never notice when one was less forth-coming than it should have been.

-- Bill Sloman, Nijmegen

Those of your ilk, astonish me constantly !!

you

at

s,

=A0So

hat

That's why we have a blood-brain barrier - to keep the magisteria non- overlapping.

-- Joe

I'd expect a low-noise, slow CMOS opamp to have a fair amount of input capacitance. If it mattered, and it wasn't specified, you might have paid more attention. Sounds like you assumed a few pF.

Fiddling doesn't scale.

John

I didn't throw up hands or invoke a diety to explain the complexity of life. Cellular machinery is an incredibly complex state/computing machine that can manage pretty well on its own. But it's not scheduled by a passive string of base pairs (four? seven?) that waits around for spot mutations to change the design.

John

You can't possibly know what's available in the genome. Nobody does. It's a state machine of unimaginable cleverness and complexity. You are looking at flipflops and gates and saying that there's nothing there to make Windows from.

That's probably all wrong. "Junk" DNA turns out not to be junk at all. Sequences are switched on and off. THINK about it.

Random variation is too slow, and too damaging. That's why DNA has repair systems.

John

Those of your ilk don't astonish me at all.

John

heir

nd

h.

nd

So would I, now. When it happened, around 1995, I'd not used one, and while it was obvious after the event that a low-noise MOSFET was going to be big, with relatively high stray capacitances, I had - at the time - the slightly optimistic idea that an op amp data sheet would tell you if any of the op amp parameters were unusual.

What was irritating was that input capacitance was on my mental check- list, and I'd tried to find it on the data sheet. Since I couldn't I'd used the typical op-amp value of about 3pF and - if that had been right - I wouldn't have needed to put a few pF in parallel with my feedback resistor, because the extra phase lag wouldn't have been be significant within the - low - bandwidth of the op amp.

Quite explicitly.

The sort of fiddling that I had in mind takes place in getting the design nailed down for production, and in that sense it does scale - the 3p3F capacitor I had to add to my circuit should have stabilised any subsequent copy of the circuit.

-- Bill Sloman, Nijmegen

Well they often do, but the unusual+bad things are typically not in the big type. Some of those TI CMOS op-amps actually have an array of something like 100 transistors for each of inverting and non-inverting inputs. By interspersing the transistors in an array they get lower Vos and TCVos, at the expense of input capacitance. I would expect a fairly high differential capacitance as well- I don't think that's normally specified on the data sheet.

It's usually not clear what the input capacitance specs really mean. Is it each pin to the power rails? Pin to pin? Under what conditions and common-mode voltage?

Having done some self-power-bootstrapped opamp circuits lately, we've become wary of this stuff. If it matters, it's best to test for yourself. The PSRR curves give some hints.

John

In fact that is exactly what you are doing. Your implicit claim that single nuclear polymorphisms aren't the engine of our recent evolution is correct, but your claim that the system doesn't wait for random mutatios to change the design is nonsense.

In primates (which includes humans) Alu elements

are 300 base-pair long segments of genetic information that get copied out of the genonome then pasted back in somewhere else, entirely at random. We've got about a million of them, and they account for some

By picking up and pasting in these relatively long chunks of data, the genetic system seems to have an appreciable chance of grabbing a base- pair sequence that codes for something useful - presumably the cost benefit ratio is better than you get from single nuclear polymorphisms, when only one pair of bases in a base-pair sequence gets changed, though this is also going on all the time.

This kind of cut and paste mechanism can explain why many of the crucial development genes show up in multiple copies in multicellular organisms

Arthropods have two cluster sof Hox genes, chordates (including us) and echinoderms have four.

It is still a totally random system for making changes, and most of the changes made are neutral, and don't affect anything. Of the insertions that do have an effect, the vast majority are deleterious, and probably - amongst other things - account for the fact that in humans some 30% of conceptions are spontaneously aborted in the first three months after conception. Every now and then something useful happens, that makes a real difference. Whatever the mitochondrial Eve had, some 200,000 years ago

and Y-chromosome Adam

some 142,000 years ago, seems to have been useful enough to have lead to tolerably rapid population growth amongst their off-spring (which is why we have a relatively low genetic diversity, when compared with longer-established species).

Culturally modern behaviour is usually considered to have got under way about 50,000 years ago, but there now seem to be enough isolated examples of some modern behaviours from earlier times to suggest it took a while for our ancestors to build up a sufficiently comprehensive cultural tool kit to make it useful enough to be likely to spread.

-- Bill Sloman, Nijmegen