Since we seem to be hurtling toward oblivion at an accelerating pace,an obvious question is: how much time do we have left on our subjectivescale, i.e., how much longer will life seem to last?
There are better codes than the Huffman code given above. For example, we couldassign Huffman codes to pairs of digits. There are 100 pairs each with probability0.01. We could assign 6 bit codes (000000 through 011011) to 00 through 27, and7 bits (0111000 through 1111111) to 28 through 99. The average code length is 6.72bits per pair of digits, or 3.36 bpc. Similarly, coding groups of 3 digits using9 or 10 bits would yield 3.3253 bpc.
PAQ7 (Dec. 2005) was a complete rewrite. It uses logistic mixing rather than linearmixing, as described in section 4.3.2. It has models for color BMP, TIFF, and JPEGimages. The BMP and TIFF models use adjacent pixels as context. JPEG is alreadycompressed. The model partially undoes the compression back to the DCT (discretecosine transform) coefficients and uses these as context to predict the Huffmancodes.
PAQ8F (Feb. 2006) adds a more memory efficient context model and a new indirectmodel: The byte history within a low order context is modeled by another low ordercontext.
PAQAR (v1.0 to 4.0, May-July 2004) by Alexander Rhatushnyak is a PAQ6 fork whichis the basis of several winning submissions to the Calgary Challenge. The primarydifference is a greatly increased number of mixers and SSE chains.
While you have it out of the pull handle hole in the lower front cowling, take a look for any wear. If so it will usually be on the top of the hole and will be pear shaped with the narrower part where the rope has worn. This motor does not have a brass sleeve like the later models do, so it may be best to ream that out a bit to keep the sides from wearing (creating more drag and chafing of the new rope). You can ream it by drilling, or just using a round powersaw file to clean it up. If you do, be sure you do not leave any sharp or even rough burrs which will damage the new rope. I see no real need to now try and install a brass rope sleeve through this cowling as what wear you see only took 40 years to accomplish.
Instead, there is an interactive program (not requiring script!) whichproduces tables of relative ages for a visitor-supplied pair of ages tobe compared.
"(1) Pull the rope out as far as you can, then allow it to retract less than one revolution until the rope end of the spool faces the port side of the motor. Lock the starter in this position by lifting the starter pinion gear upward to engage the flywheel ring gear, and slide the slightly spread handles of a pair of pliers under the gear, holding the gear upwards and into the flywheel gears. (2) remove the starter rope handle. (3) remove the rope from the spool. (4) Replace with a new rope, which needs to be 56 1/4" long, then burn the ends with match for about 1/2" to keep them from unraveling plus to be stiff so they will hold in the spool or anchor. Thread the rope thru the slot in the spool AS IT CAME OUT OF. (5) Thread the rope thru the motor cover and install the handle. (6) Hold the starter rope handle to keep it from rewinding too rapidly, remove the pliers allowing the rope to rewind normally.
This site, dated 1997 (earlier than the Logtimesite), starts with the same premise as Logtime, and even usesstartlingly similar phrases (see below), but does not use logarithms.
For those readers who have had some calculus, or at leastare familiar with the elementary properties of logarithms, a derivationof the mathematical relationships of Logtime is presented below.
Whatwill be shown is that a simple, plausible assumption about theestimation of time intervals by the human mind ("psychochronometry")implies a mathematical model that can explain common observations, atleast to the extent allowed by the unmeasurable subjective nature ofthose observations.
This is not a "proof" of Logtime: the basic premise is notonly unproven but probably unprovable because of its subjective natureand the difficulty of experimentation.
To reassemble the cleaned starter unit, (1) place the upper bushing and starter head with the pinion gear spring onto the spool. (2) Insert the spring retainer in the spool. Align the slot in the gear with holes in the spool and spring retainer. Insert the roll pin with its split seam in a horizontal position to avoid dragging against the slot in the pinion gear. (3) DO NOT oil the gear or spool, this will attract dirt, causing pinion to bind on spool. (4) Insert the spring into the bottom of the spool, turn to engage with the slot of the spring retainer.