Naked Science Forum
Non Life Sciences => Geek Speak => Topic started by: Nizzle on 14/08/2009 12:29:56
-
As far as my computer hardware knowledge goes (let's say from the tip of my nose to my upper lip) i know that the true basics of a computer is all about 1 or 0, power on or power off.
But I can't imagine this is really efficient. At some point, miniaturization will reach a limit (atom size), so my question is:
Are we looking for other options to replace the binary system? For example something like power off, power 25%, power 50%, power 75%, power 100%?
-
http://www.newscientist.com/article/dn17575-ditching-binary-will-make-quantum-computers-more-powerful.html
the site is down as of the post, but it is fairly reliable so i imagine it will be accessible soon :)
-
There are pros and cons to using higher base systems in computing. With binary we can get a result of either 0 or 1, whereas with trinary, for example, we could get 0, 1 & 2, or perhaps more usefully, -1,0 & 1. However, this would greatly complicate the underlying electronics as they'd either need to distinguish between different levels of state, or between opposite state polarities.
If you want to see just how efficient binary is, in comparison to other bases, see how high you can count on the fingers and thumbs of your hands in different bases. In base 10 you can only count up to ten, but in binary you can count up to 2047.
-
DNA sequence is a binary code, it has proved rather successful.
-
a bit off topic, but something i've thought about a bit. i've thought before that quantum computing was a bit hmm.. i don't know.. if it was a more efficient way of storing data, surely evolution would have stepped in and used a mechanism to harness it? quantum computing is extremely powerful, of course, so i'd have thought that evolution would have found some way of using it as a means of increasing processing power. then again, perhaps it is that life and using quantum states cannot work together for some reason.. maybe that is to do with incompatibility with organic substances, or constant observation of the system (thus forcing 1 state to appear)..
-
There are pros and cons to using higher base systems in computing. With binary we can get a result of either 0 or 1, whereas with trinary, for example, we could get 0, 1 & 2, or perhaps more usefully, -1,0 & 1. However, this would greatly complicate the underlying electronics as they'd either need to distinguish between different levels of state, or between opposite state polarities.
If you want to see just how efficient binary is, in comparison to other bases, see how high you can count on the fingers and thumbs of your hands in different bases. In base 10 you can only count up to ten, but in binary you can count up to 2047.
Do you have six digits on one hand?
Anyway,
If memory serves me some of the earlier digital electronic systems used other bases- notably base 10. They gave up because binary was more efficient.
Quantumn computing may alter that but, until then, my guess is that we wll stay with base 2.
-
There are pros and cons to using higher base systems in computing. With binary we can get a result of either 0 or 1, whereas with trinary, for example, we could get 0, 1 & 2, or perhaps more usefully, -1,0 & 1. However, this would greatly complicate the underlying electronics as they'd either need to distinguish between different levels of state, or between opposite state polarities.
If you want to see just how efficient binary is, in comparison to other bases, see how high you can count on the fingers and thumbs of your hands in different bases. In base 10 you can only count up to ten, but in binary you can count up to 2047.
Do you have six digits on one hand?
You'd need 6 digits on both hands. 2047 needs 12 bits. 10 digits will allow up to 512 & 11 digits, 1024.
-
Doesn't everyone have eleven fingers then? (Oops - 1023 - I've no explanation for getting 2047!! except double-vision and not being able to count properly - hopefully I'll get the hang of it one day)
With your palms facing you...
R Thumb= 1s
R F1= 2s
R F2= 4s
R F3= 8s
R littleF= 16s
L littleF= 32s
L F3= 64s
L F2= 128s
L F1= 256s
L Thumb= 512s
...so the total you can represent is the sum of those values = 1023
-
DNA sequence is a binary code, it has proved rather successful.
Surely DNA uses base 4? A, T, C, & G?
-
DNA sequence is a binary code, it has proved rather successful.
Surely DNA uses base 4? A, T, C, & G?
Ture, but biology also uses a binary modifier; they either are; or are not; methylated.
As usual biology is very clever but damned complicated.
-
Surely DNA uses base 4? A, T, C, & G?
C always connects with G, T always with A,
So the base pairs (http://en.wikipedia.org/wiki/Base_pair) are either, C-G or A-T, (two options, binary).
-
Surely DNA uses base 4? A, T, C, & G?
C always connects with G, T always with A,
So the base pairs (http://en.wikipedia.org/wiki/Base_pair) are either, C-G or A-T, (two options, binary).
Ah yes, but the orientation can be either C-G (A-T) or G-C (T-A), with any given strand of the double helix containing either C, G, A, or T (with the other strand being it's mirror). This gives us four options. And so DNA is a Quaternary system.
-
Hi RD
Yes, I accept that. But does it matter which way round the pairs are? Obviously that affects the order of nucleotides on either one of the single strands of DNA. At ‘A’ level the nucleotide sequence was always described as ‘a language of four letters’. Isn’t it the order of the one strand that expresses genetic information? I thought the only purpose of pairing is to allow the helix to unzip and replicate? This could well be one of those occasions where I find out that a little bit of information has lead to me making untrue assumptions. [;D]
Hi Bored Chemist
I’ve heard of DNA methylation but I don’t really know what it is. Doesn’t it affect some but not all of the nucleotides and stops them being expressed, thus forming an integral part of the epigenetic code? If the nucleotide sequence is a language of 4 letters then methylation could be the use of caps instead of lower case to double up the number of characters?
-
Ahh. Pmb00cs posted his reply while I was drafting up mine.
-
[:o] DNA is in no way binary!
It's a quadrinary system and A-T is most definitely NOT the same as T-A.
DNA methylation is still being researched, but one of the functions already discovered is the activating or inhibiting properties.
In order for DNA to be transcribed to RNA, all kinds of activators and promotors need to bind to the DNA first.
Some of these activators and promotors sterically require a methylated DNA strand in order to bind in a stable manner.
Another class of activators/promotors cannot bind if a -CH3 group is attached to the DNA due to sterical hindrance.
So the methylation/unmethylation process is a way for the cell to indicate which DNA should be transcribed and which should remain inactive.
-
Binary, quaternary, octal, hexadecimal, and more. Programmed in them all, converted them all. not difficult.
( if you do a search on octal machine code you will find the statement"I have very gray hair and not much left" is more believable, I encountered a 1960`s vintage custom processor PLC implemented in a huge frame of blue red and yellow and black lego DTL modules, ( ferranti? marconi?) and the code was listed in octal,and hard wired on a pcb as a sequence or links that was addressed by hard logic counters. Part was a state machine. I had a bit of a culture shock, even having built CPU`s in discrete logic as a university project to get you familiar with digital design.)
50% between 1 and zero volts or current or light levels or flow rates or......is uncertain you can decide the thresholds when you sit and programme the IC you want before you send it out to a chip fab, or do it in discrete stuff, analogue or digital NOT HARD.
The circuitry required is just a comparator with hysteresis to provide a dead zone around 50% to reduce the effects of noise.
For a multi valued input, ternary, quaternery etc a comparator is requred for each significant level value. If the input available is limited then binary will give the greatest noise immunity from interfering signals, going to a 4 level physical symbol set up will halve the noise immunity.
-
There are 4 base pair sequences, but the active sites are modified by adjacents so i can`t realistically say how many real combinations by property there are and not many others have even been speculated about properly by the people still researching this.
I will go and get my coat for saying what others won`t, as usual.
-
There are pros and cons to using higher base systems in computing.
If anything, the hardware acoutics of harmonics filters apply to just that,THE HARDWARE (period), that was done to accomplish speed and data reliability against machines corruption(never make a mistake of what was through-put).
The real point in base systems i believe is expressed in it either being stack data(instruction) or used Data itself(simply the user information that is operated upon that is fed to the CPU).
A base systems' relevence is only to the symbols sets quantification(number of as a tatal of group) "without making the total integer sum quantity of representation of symbol individuals units a fraction or place(alike decimal place / a floating point) or finite and singular complete units p/representation(integer pure math definition).
If a base number were relevent, it would be pre defined Object oriented to each instruction operation circuit filter and to assign an alike of higher language operation onto a chip, meaning its stack fed data may be required to be a pre defined set of bytes as a chunk (a set of fed assigned parameters), but its operation insignate call denoted/assigned relevent to the number of objects and specifically for the operation circuit carried in the pre made CPU system of circuits.
That would be one hell of an ordered abstraction of the higher language commands built into a CPU instruction set(architecture).
However pure base maths potentially could benefit accuracy against base conversion distortions at the math level.
-
Number sytems and abstractions from the physical are all very well as my tonque in cheek first comment was intended to say, ( and the offtopic aside was bracketed) but the question was about physical interface parameters and the hardware of logic gates.
Can you tell me where most of above has anything other on topic or informed off topic?
object oriented what? WHAT THE HELL CHILD!
THE FOLLOWING IS OFF TOPIC I AM OLD ENOUGH TO BE AWARE AND APOLOGISE FOR THE SHOUTING sorry
The structures and the compiler methods of the OO languages are structurally bloated and that renders them unsuitable for safety critical applications, assembler, Forth toolkit derived code , some verified C`s and ADA are all that are qualified for Nasa and military aerospace usage.
Assembler and any of the forth variants are extensible, compact and testable, but indisciplined programmers cannot work with these as they rely on their judgement and expertise rather than a restrictive compiler writer to contain their mistakes.
If I have 2 tonnes of gear doing 400rpm what do you think I will programme the motor hardware in? not an untestable, uninsureable object oriented thing.
I`m sorry but loading a piece of information with loads of tags to define its properties so it can be found and identified by routines is not the same as dataflow control at point of compilation, so that datatyping is implicit by one of a few methods of dictionary encoding and there is a minimal payload associated with the data item. A linked list dictionary is one used even by the non hardware aware.
It all starts if you think strong typing is important, it isn`t, self discipline is.
I have no Idea who you are or who you think you are, but if I expressed what I now think, I would be barred from this forum unfortunately moderators seem to be less attentive to your postings.
-
MODERATORS COMMENT
It is not the moderator's job to evaluate the quality of a post or pass judgment on ANY person or their opinion. We do not have enough experts in each and ever subject posted to the site, or subjects that crop up as asides from the given subject, to adjudicate the voracity of any given post made by anyone. This is not a forum of rigorous scholastic academics, but a forum that attempts to use the combined knowledge of its member to provide general answers on a multitude of subjects outside the purview of the moderators.
Disagreements are inevitable.
-
Nizzle: i know that the true basics of a computer is all about 1 or 0, power on or power off.
There are other versions of decisions not eveluating as 1 or 0 true or false in computing.
PERL(Practical Extraction and Report Language) is a great example of what can be returned from an "if" statement,
quite unusually, up to around 7 but in special situations only, for most 0 and 1.
Tringstar: was about physical interface parameters and the hardware of logic gates
So there should be 5 buttons(1 special , four others), on/of 25 50 75 100.I don't see how that a leap out of the modern age to the future.
-
I thought that the acronym 'PERL' stood for 'Pathetically Eclectic Rubbish Lister'
Sort of ironically appropriate to refer to it really [;)]
-
Back in the fifties when the storage of one bit requred two thermonic vacuum tubes three state logic was tried for a while but as transistors came in it was soon dropped.
-
It is an interesting bit of post hoc rationalisation that causes people to assume that there's something basic about the choice of binary logic for computers. Binary circuits happen to be very convenient to build. That's why computers mostly use binary logic. However, binary arithmetic is very lumpy and nearly always needs translating for us to understand the answers. Decisions (which are so frequently talked about in yes/no terms) in life and when presented to a computer are very seldom binary. They often need to be reduced to a set of binary decisions in order to be presented to a 'simple' binary computer but usually start off as either a multi-variable and multi valued input and involve a number of levels / values of output from a real life decision; a load of 'what if's and 'except when's. It's very often not a yes/no but a maybe and maybe not.
I wonder what sort of reasons would be given, post hoc, if a really small, cheap and fast tri-state circuit suddenly emerged.
-
I think Edster already made this point, but I think it's worth repeating.
At the hardware level, particularly in a big chip, it's bl**dy difficult to ensure that noise does not cause errors. To keep a lid on the heat produced as the speed goes up, we have to use cores with lower supply rails, so the noise immunity is further reduced, which makes everything even harder.
That's not to say that hardware systems other than binary are impossible, or worth considering, but I suspect that companies like Intel are unlikely to deviate too far from their current path unless they see some significant advantage, in terms of, for example, performance, power reduction, or cost.
Personally, I always thought analog computers were really cool, perhaps because they have an infinite number of levels.
BTW, I'll be happy to share my memories of Solidac (a germanium transistor 20 bit serial machine with a propensity for playing bagpipe music) with anyone with a strong constitution and too much caffeine in their system.
-
Thirty years ago I used to work with analog computers on HELL DC300 scanner/recorders they used 30 circuit boards weighed about 50 Kg and consumed several hundred Watts.
They needed frequent adjustment and the 50 or potentiometers became jittery and had to be lubricated with WD40 (some were sealed and you had to carefully drill holes in them).
Present day machines use a few chips all set up via a terminal, a vast improvement.
-
Well, sure, didn't you feel as if you were driving something more like a steam locomotive than a mini? Bet it had a great aroma too. All that ozone and hot amplifiers.
The thing I miss most about tubes/valves is the hot aroma they gave off. Maybe I could make a product that makes PC's smell as itf they are running on tubes? Probably get shut down for health reasons in a couple of days!
-
Decisions (which are so frequently talked about in yes/no terms) in life and when presented to a computer are very seldom binary
All very well to simply see boolean and 1 or 0(a bit), but the true underlying rationale of making computers is data and data handling through instruction circuits.
Truth tables have individual circuit instruction paths and patterns of data shunting sequence for each version of its operators themselves.
Its not actually booleans, thats only one small piece.
The true governor is the byte , or how to encode both data and instructions inside anything from 0-255(8 bits) or more until reaching 64 bit Alpha CPU counters.
1 and 0 in the binary system are really only convenient encoding.
Boolean is extremely important only because it helps make a clear choice in channeling to the next set of instructions to start at on the stack.
-
I'm afraid that made as much sense to me as a random sentence generator I once wrote, which used a large database of word-triplets and their corresponding usage statistics to string sentences together.
I was actually quite pleased with the results.
Oh what the heck - here are a few examples...
Subjects of photographs are often considered to be about three miles distant, from whence several of the most famous of these were built to promote development, even if not apparently viable in strictly financial terms.
According to the first governor of the 20th century, the lines between major cities were converted to standard gauge line was built in the same way.
After a tearless and comical funeral service, where he is always a bit of American and Canadian slang refers to statistical or econometric analysis of numeric data.
-
Alright then, for a computer to be affected by the state of one or zero the filter circuit needs to detect which it is(1 or 0) before it gets re-channelled for whatever its using that bit for what it does with it.
Becuase the number of instruction circuits(or re-channelling operation sequences) are finite, and because more accurate magnetic disc signal storage is possible, discs and CPU's could read signals and stored data(instructional or pure data) using some other base(e.g. see Chinese language symbols for numerics, has an individual character up to around 600 or more before it has to do representation rules for bigger numbers) well above base 2 logic, inclusively making some pure data have symbolisis representation inside that base not simply control the CPU sequences of use of the instruction circuits.
e.g. 1 oscillation beat to send a signal for a 0 and requires 8 beats in binary to read that stack number for its corresponding circuit.
Now 1 oscillation beat to represent a specific signal frequency for the CPU recognition filter and and only one beat used.
Best i get it down to physically in my mind is the disk being able measure the lay-down length of the written magnetic section or be extremely sensitive to the level of the stored magnetic charge that was layed down on the disc so it can recognise the level of the charge.
RAM much the same sort of thing.
unoff-sig: http://www.wild-canidae-conservation-forum.netne.net/index.php
-
If you had ever maintained machines with analog computers you would have been real happy when digital came in.
-
If you had ever maintained machines with analog computers you would have been real happy when digital came in.
LOL! I know what you mean. The first digital machine I worked on was bad enough. It was originally designed to be built with valves (er tubes?), but germanium transistors suddenly became affordable, so they made it "solid state" instead. It was a pain. It would blow transistors on power up at an alarming rate. My job was to track them down and replace them. Mind you, it taught me a lot about computers - I had no choice!