PDA

View Full Version : And to think I suggested something better.


Loose rivets
5th Dec 2013, 04:09
Incredible. Despite being the author of a book which suggests an even more faultless code, I sometimes wonder if I don't do the blueprint to our beleaguered form an injustice when I suggest we are just a sub-set of perfection. My novel is just a yarn but the code to life on this planet often leaves me just plain bewildered.

While you're delving into the BBC site - you will, won't you - just look at the 400,000 year old DNA which shows more than a melting pot between species supposedly separated by time but also by continents.*

I hope one day I shall be allowed to see the truth.


BBC Science - DNA: the 'smartest' molecule in existence? (http://www.bbc.co.uk/science/0/22199991)

Just one tidbit of this fascinating article.




Different sequences of the four types of DNA bases make 'codes' which can be translated into the components of proteins, called amino acids. These amino acids, in different combinations can produce at least 20,000 different proteins in the human body.

Think of it like Morse Code. It too uses only four symbols (dot, dash, short spaces and long spaces), but it's possible to spell out entire encyclopaedias with that simple code.

Just one gram of DNA can hold about two petabytes of data - the equivalent of about three million CDs.

That's pretty smart, especially when you compare it to other information-storing molecules. Using the same amount of space, DNA can store 140,000 times more data than iron (III) oxide molecules, which stores information on computer hard drives.


*
BBC News - Oldest human DNA may hold answers to man's evolution (http://www.bbc.co.uk/news/science-environment-25224095)

arcniz
10th Dec 2013, 12:57
And to think I suggested something better.

Well, Rivets, back in the day when the Ethernet standard was still an unlaid egg, about the era when Disco appeared on scene, from inside the selection process one proposed 4-state bits as the base topology for networking pc's. The idea awakened peripheral interest, but quickly fell to the force of binary all or nothing axers. Were it other than that, maybe, maybe our machines could-would be communicating in DNA-speak now... for what that's worth. A few bazillion extra ways to say "maybe" woulda been a plausible consequence from that. Opportunity missed for...lack of interest.

balsa model
10th Dec 2013, 17:44
4-state bit
That would be an oxymoron.
"bit" stands for "BInary digiT".
(At least that's what they say, now, in schools.)
Incidentally, I presume that you are not conversant with modern signals theory (and practice) but multilevel "symbols" are used quite frequently.
Some common examples:
American HDTV is broadcast with 8 levels / symbol (which is then modulated on RF).
Your Giga Ethernet uses them; that's why you can use it with Cat 5e cable, same as for 100 Mbit/s Ethernet. You don't need a cable rated to Gigahertz.

Lonewolf_50
10th Dec 2013, 17:51
arcniz: a few years back, I read about the next "revolution" in computing power, which was based in the use of color/light rather than on/off (binary) combinations for computer instructions. This would lead, IIRC, to a basic 8 bit baseline, by using the seven rainbow colors of light and "off" as the base 8 derived standard.

It's been a few years, and I am not sure how far along that ever got, or if there were density issues that consigned this idea to the graveyard. One of the things that made this an appealing approach was the lesser amount of energy needed to drive the circuits/transistors/logic gates using light versus electron flow/current.

Hydromet
10th Dec 2013, 20:59
one proposed 4-state bits as the base topology for networking pc's.
That would be an oxymoron.
"bit" stands for "BInary digiT".
So a four state system would have Quaternary digITS or quits?

I'll quit now, while I'm ahead.

Loose rivets
10th Dec 2013, 21:43
Well, not an entirely new concept.:rolleyes:


Arfer C Clark. Octospiders

Google Image Result for http://zhippo.com/TrentEdwardsCatTattooHOSTED/images/gallery/octaspider.jpg (http://www.google.com/imgres?imgurl=http://zhippo.com/TrentEdwardsCatTattooHOSTED/images/gallery/octaspider.jpg&imgrefurl=http://www.trentstattoos.com/tattoos/tattoos_37496.html&h=535&w=600&sz=350&tbnid=aTwpN8_GdbX7eM:&tbnh=91&tbnw=102&zoom=1&usg=__a0PARIHgR1jUnR9HxE4mjlh2BRQ=&docid=bw_G03GTb-K2rM&sa=X&ei=NIenUojkNMH6oASakYDICw&sqi=2&ved=0CC0Q9QEwAQ)

arcniz
11th Dec 2013, 02:32
That would be an oxymoron

Thank you for the corrective suggestions. In the metaliteral dreamworld of Pprune, precision and acuracy are sometimes scored at transgalactic scale, and sometimes looked at down to a muon on a gnat's ass. Oftentimes, the net difference between these extremes is negligible in value and usually also is moot in effect.

Rather than oximoron, I am more of an omnimoron, mself -- working on the theory there's nothing so simple or so perfect that a persistent fool cannot screw it up a bit further than the collective history of humanity has done with manifold facts and fictions.

If someone could point one to an accessible authoritative fulltext copy of the relevant source specs for HDTV, that would be greatly appreciated. Am familiar in part with fractal image construction-deconstruction, but not specifically with subset details of how it is applied in HDTV.

Watching HDTV signals off air from sources more than a hundred miles distant, one sees many variant modulation vagaries from weather inbetween self and source, and from particular history streamming in from the sun, so on stormy nights one has ample chance to ponder the craft of those silly little squares that make no sense at all when broken in sequence, but kiss up to imagic perfection when conterranian signals are strong and nearby stars recently are mostly sleeping.

arcniz
11th Dec 2013, 05:51
Lonewolf_50 says:

I read about the next "revolution" in computing power, which was based in the use of color/light rather than on/off (binary) combinations for computer instruction

One of the less-publicized limitations of semiconductor technology is dimensional stability. Correct widths and layers are absolutely essential when laying down semiconductor materials on chip wafers, and all the art of manufacturing focuses on dimensional uniformity in the planes and places that enable circuits as we presently know. I suspect process limits there are a substantial deterrent to n-state digital circuits as production products internally co-operating in parallel at various light frequencies. Sometime, maybe, this may change as process controls improve. Testing would be an issue also... my info is that testing comprises more than HALF the cost of many higher-complexity binary chips. Hard to figure how multi-wavelength logic would impact that, but not a plus, surely.

arcniz
11th Dec 2013, 05:54
Hydromet says:
So a four state system would have Quaternary digITS or quits?

I called 'em "quats".


A real limitation in the practical number of concurrent states for operators and variables in deterministic computers is the number of concurrent things the developers can get their heads "around".

When one started designing serious mainframe computers, there was zero design automation available to support development, greatly limiting the tolerable level of complexity. Acres of fellows in white-shirts and ties at drafting tables were the norm. After a decade, most of them had gone, but design-automation was still very primitive -- no simulation or autoplacement or path-timing analysis, for example.

Now, a generation later, one notices tMEshat there's still not a lot of tools for n-ary alternative analysis of timing and path behavior. Message seems to be that the people who drive design have limits -- along with the equipment.

One can readily foresee an oncoming age of non-deterministic automata that are mostly analog in style, though mostly digital in design. Perhaps that will foster n-ary logic --- or perhaps not.