Investigating the Origins of Dreamspell

Updated 20 Jan 2010


At Whats New item 166, Jaguar Prophet V Closer of the Cycle, we saw that although Carl Calleman's criticisms of Jose Arguelles' Dreamspell system apply largely to his own invented version of the Maya calendars, he did raise some interesting points that need to be addressed. These points are included in the following list:

1. Dreamspell is nor "the Mayan calendar", but an invented system that blends the 260-day Tzolk'in calendar with a non-Maya 28-day "lunar" count, and also changes the day names so that, for example, 13 Eb (translates as 13 tooth, 13 grass or 13 road), becomes Yellow Cosmic Human and 9 Ben (translates as 9 reed or 9 cornstalk ) becomes Red Solar Skywalker.

2. The energetic qualities of the 260 days are completely different from those of the Q'iche Maya, who have kept the count unbroken since the time of the Classic Maya, and that this discrepancy is due to the Dreamspell practice of freezing the Tzolk'in count for a day every four years, on the Gregorian leap year-day, February 29th. So in Dreamspell, Yellow Cosmic Human (13 Eb) will occur on June 4th 2005, but according to the unbroken count of the Q'iche Maya, 13 Eb will occur on 49 days later on 23 July 2005.

3. The Maya kept a 29.5-day lunar calendar (of alternating 29 and 30-day lunar months), not a 28-day one.

4. The Maya did not celebrate New Years day on the 26th of July every year, since this would necessitate adding a leap day into the Haab (365-day calendar) and there is no evidence of this.

5. The "Loom of Maya" or "portal days" that create a pattern of darker squares on the 13 x 20 Tzolk'in grid, as shown in many depictions of the Tzolk'in, was an invention of Tony Shearer. The Tzolk'in grid itself was not used by the Maya but is a convenient modern way of depicting the Tzolk'in for easier understanding. Thus, the convention of starting with 1 Imix is not a Maya convention.

Jose Replies

In the Rinri Project Newsletter Vol. 2 No.5, Jose Arguelles has recently answered some of these points

"There continues to be confusion and controversy over what is called the Dreamspell Count and what is called the Traditional Mayan count. Let it be made unequivocally clear: the Dreamspell count, based on the Gregorian calendar synchronization date of July 26, according to the sequence of the Gateway family year bearers, Seed (Kan), Moon (Muluc), Wizard (Ix), and Storm (Cauac), is not an invented fantasy. It is a precise expression of the only recorded prophetic tradition of the Maya, the Chilam Balam. In fact, the Dreamspell synchronization system in the tradition of the Chilam Balam is called "Cuceb," literally "Squirrel," or "that which turns." By this system of correlation, which dates back at least to AD 1513, the sequence of days of the Tzolkin marked Kan, Muluc, Ix and Cauac always synchronized the day 0 Pop, beginning of the Haab 18-Vinal + 5-day Uayeb Solar Calendar, on the Julian ( pre-1582) Calendar date, July 16, and the Gregorian (post-1582), July 26.

In the fulfillment of the prophecies of Pacal Votan, the mysterious occult source of the Chilam Balam tradition, the Gregorian Calendar is replaced by the Tun Uc, Thirteen Moon 28-Day civil calendar so that Magnetic Moon 1 (old Gregorian July 26) always correlates 0 Pop of the Haab with a Seed, Moon, Wizard or Storm day of the Galactic Cycle or Tzolkin. This is the meaning of Cuceb, that which turns.

The purpose of the Tun Uc (Moon Count or Seven Count) is to establish a pure equivalence of civil time governed by the 13:20 frequency of the Law of Time, where: 13 moons (4 x 7 days) = 7 days (4 x 13 weeks) and 28 Wavespells (13 days) = 13 Moons (28 days) The Tzolkin grid which is utilized in the Mayan Factor and all Dreamspell play is also derived from the Chilam Balam, where it is known as the Buk Xok permutation table. The gift of Tony Shearer was to show the Loom of Maya grid of 52 activation portals within this matrix. Arranged as 13 quartets, each quartet having a sum of 28 tones, this also demonstrates the coded existence of the Thirteen Moons in the Loom of Maya Tzolkin-Harmonic Module grid. Finally, the sequence of the Thirteen Moons or Tun Uc with their totem animals is also pure Mayan and corresponds to the sequence of thirteen Mayan constellations as re-discovered and presented by Hugh Harleston en El Zodiaco Maya (1991). Since Harleston does not account for the Day-Out-of-Time, there is but a one day discrepancy between the thirteen totemic constellations and the corresponding Thirteen Moons correlated according to the Cuceb-Dreamspell synchronization point of July 26."

In order to check these claims, I looked first at the  online copy of the Chilam Balam of Chumayel. There is not alot to go on there, but in Appendix D, the Chumayel notes that the Chilam Balams of Tizimin and Mani have more to say on the subject of the subject of the Cuceb or Squirrel. At considerable expense, I managed to locate a copy of Maud Makemson's translation of the Chilam Balam of Tizimin, The Book of the Jaguar Priest (1951), and a copy of  The Codex Perez and the Chilam Balam of Mani, translated by Eugen R. Craine and Reginald C. Reindorp (1979) I also got hold of a copy of Tony Shearer's Beneath the Moon and Under the Sun (1975) and the new edition of Shearer's 1971 book, Lord of the Dawn. I present my findings below.

The Chilam Balam of Tizimin

The earliest writings of the Chilam Balams, or Jaguar Priests date back to 1593, when they were written in the Maya language, but in European characters. The writings were compiled into a collection in 1752, and named after the town where each version was found. Although the Chumayel was written after a complete conversion to Christianity, the Tizimin was written before that had happened ((p.218). The Tizimin records that the Itza overlords, (before the Spanish invasion) who had taken on the roles of the priesthood, had not retained the knowledge of the original priesthood, knew nothing about the days and katuns of the Maya calendars (p.18), and did not even know the names of the gods, to invoke them correctly (p.40).

Although the Spaniards ridiculed the Maya calendar (p.50), these "wretched men" (p.60) continually tried to find out more about the it, since their Bishop, Diego de Landa had organised the destruction of all the Maya hieroglyphic records. The calendar had "fallen into confusion", (p.54), particularly in connection with the year-bearer.

There is an interpolated section (see p.107) that was inserted by persons unknown, to put the "Jesus Christ count" above all Maya calendars (p.67). This is followed by "a pitiful attempt to explain the details of the Maya calendar" (p.108), in which Katuns are listed at 24 years in length instead of 20. It seems that, from 1752 onwards, the katuns were thought to consist of 24 Haabs (365-day years) instead of 20 tuns (360-day years), and they were named after their beginning days, rather than their end days (p.132). Makemson has calculated exactly how "the 17th century Maya who wished to reconstruct the katun-count" probably came to this false conclusion (p.136). In the commentary, Makemson concludes that the Tizimin is not a factual record (p.113) and the year-count had been lost by the time the earliest versions were written down in the late 1500s (p.223). However, from the vague information correlating Maya to Gregorian dates in the Tizimin, Makemson calculated a correlation that implies that the Long Count date occurred on June 22nd 1752. She also reasons that the July 16th (Julian calendar) or 26th (Gregorian) date of New Year given supposedly by de Landa is based on a misunderstanding of what he actually wrote, and goes on to explain how it could have happened.

Eric Thompson has discussed Makemson's conclusions in his book, Maya Hieroglyphic Writing (1978 edn. p.306-307), giving a convincing counter-argument (the GMT correlation advocated by Thompson puts on 21st December 2012).

The Chilam Balam of Mani

The Chilam Balam of Mani was collected together and translated into Spanish by Don Juan Pio Perez, and published in1843. One of the first realisations upon reading the Chilam Balam of Mani is that it is riddled with errors in the calendrical information, from omissions of regular periods (e.g. the "burner" period on p.33), to serious breaks in calendar sequences (p.32, 34, 35). Perez confesses to leaving out the material that offended his religious beliefs, calling it "Aristotelian foolishness about the creation of the world, the formation of matter and the seven heavens" (p.58). On p.69, it is admitted that the Spanish priests and conquistadors had attempted to interpolate their own material into the Books of Chilam Balam, resulting in " intermittent sections of absolute nonsense and frequent blurrings of the finer meaning as written by the Mayas."

The Mani, like the Tizimin, gives a mixture of 20-year katun sequences and 24-year ones, (p.79), which the editors describe as "confusing and conflicting statements regarding the length of the katun"(p.157), and agrees on the Itzas' ignorance of Maya calendars (p.104-107). It admits that the count of years was lost (p.110) and that nobody could read the old hieroglyphs any more (p.107). The interpolated passage about the Jesus Count being above the Maya counts is also included (p.126). This implies that the Gregorian calendar takes precedence over that of the Maya, being based on Jesus' supposed birthday (p.154).

The Cuceb section starts on p.98 and gives a Maya - Gregorian correlation:

"13 Kan is the first day of Pop and the stone [idol] for Katun 5 Ahau was taken [engraved] in the year 1593, which passed on 15 Tzec"

The editors conclude that this is "a strangely confused date" and agree with Roys (who translated the Chumayel) that this is a" fictitious katun". The year-bearers given in the Cuceb section (p.99)  that agree with the calendar round part of this date (but not the katun) are correct for the GMT correlation (in this case, the "first day of Pop" is 2 Pop, due to the slipping of year-bearers already referred to - the best software to check these post-conquest dates is Peter Meyer's Maya Calendrics, since he gives the varying post-conquest haab correlations for different areas- see his article for an explanation), so these cannot be the dates to which Arguelles is referring as the origin of the Dreamspell correlation. In fact, there are several other date correlations given in the Mani, including another confusing one (p.115 & note 41 p.187-188): 15th February 1544 = 18 Zac 11 Chen. The first thing we notice is that 18 Zac and 11 Chen are both haab dates, so one must be wrong. However, when we cross-reference the Tizimin, we find this date is also mentioned, but as 18 Zac 11 Chuen (Makemson p.26). Using Peter Meyer's Maya Calendrics software, if we try various correlations that have been suggested by Mayanists over the years, then 18 Zac 11 Chuen would have been 15/2/1544 in the Weitzell correlation which has a correlation number of 774078 and puts the end of the 13-baktun cycle ( in the year 2532 AD. However, this does not agree with several other correlations given in the Mani (if the Weitzel end-date is jumped back by calendar-round multiples, thus keeping this haab-to-tzolkin correlation, but adjusting the Long Count, we arrive at 16th December 2012 as the end-date - a correlation of #584278).

On p.98, just before the Cuceb section, 21st July 1596 is equated to 4 Ix and "the first of Cumku", and the NewYear for 1596-1597 is given as 3 Cauac. Although the second date agrees with the list of year-bearers, (a set of 4 possible Tzolkin daysigns that occur on the first day of the Haab calendar) which support the GMT correlation, the first date does not. However, there is a correlation that supports both of these: using Peter Meyer's Maya Calendrics software, which allows any correlation to be used, we find that a correlation number of #584315 agrees with the first correlation above, and gives 15th August 1596 for the Maya NewYear of 1596. Amazingly, if we use this same correlation  number to check the 15th February 1544, we get 3 Kan 11 Chen, which implies that it was the 18 Zac part of the double-haab date that was in error. This correlation number is 32 days on from the GMT and gives an end-date of 22nd January 2013*, but nobody has ever suggested it before, and I am not putting it forward as a viable correlation number (I'll get to the point I'm making shortly). As Thompson points out, a valid correlation number should be based on "the positions of the 260-day almanac, the years, the katuns, the moon, Venus, pottery sequences and architecture" (Maya Hieroglyphic Writing p.307), not just one source, or even three, as Makemson did.

Another correlation given in the Mani is 14 February 1793 = 6 Chicchan (p.89). The closest this could be to the GMT correlation (#584283), (in which 14 February 1793 corresponds to 7 Ik), is 57 days away from it, at #584440. In the Makemson correlation, 14 February 1793 is after the end-point ( but if the count is continued, it does not correlate to 6 Chicchan, but to 5 Manik, 15 days away from the GMT correlation, considering only the Tzolkin and not the Long Count).  In the #584315 correlation mentioned above, 14 February 1793 equates to 1 Oc, 32 days away from the GMT correlation. On p.127 a correlation for 10th May 1756 is given: 7 Cauac 17 Muan. In terms of only the Tzolkin, this is 26 days off the GMT correlation.

The point I am making is that the various correlations given in the Mani can lead to at least 5 different correlations between the Tzolkin and the Gregorian calendar. Now it is time to check Jose Arguelles' claim that the Dreamspell count comes from the Cuceb section of the Chilam Balam (of Mani, since the others don't have a Cuceb section). Using Gary Cook's Dream Spell Calendar Viewer program, 14 February 1793 equates to 3 Wind (3 Ik), 42 days away from the Tzolkin day given in the Mani, so there is no agreement there, but this is about 9 pages before the Cuceb section starts. Trying the other correlations given, 21 July 1596 equates to 2 Wizard (2 Ix) in the Dreamspell count, which is 80 days away from the given Tzolkin equivalent in the Mani, but this date is given on p.98, just before the Cuceb section. If we go to the first date given in the Cuceb section, (13 Kan = first of Pop - or New Year in the Haab calendar in 1593) and look for 13 Kan in the year 1593 in the Dreamspell count, we find that it does not correlate to 26th July, when Dreamspell celebrates New Year, but the 3rd November.

However, the list of year-bearers given in the Cuceb section (p.99), does agree with the Dreamspell (Kan, Muluc, Ix, Cauac) so the mystery deepened. Then I spotted how the error could have been made. On p.98, just a few lines before the start of the Cuceb section, the scribe states that July 21st 1596 was 4 Ix and first of Cumku. Then it is mentioned that 2 Ix was the year-bearer in 1595, and "the year-bearer for 1596-1597 will be 3 Cauac followed by year-bearer 4 Kan..." This agrees with the list of year-bearers given in the Cuceb section. I entered 21st July 1596 into the Dreamspell calculator and got 2 Ix. This was the year-bearer given by Mani for the previous year. Further checks showed that if the 21st July was entered into the Dreamspell calculator for all the years given in the year-bearer list, all the Dreamspell dates corresponded to the year-bearer given by Mani for the previous year. It seems that, since the Maya year started in the middle of the Gregorian year at that time, and that the list of year-bearers was written as 

13 Kan  1593-94

1 Muluc 1594-95

2 Ix       1595-96

3 Cauac 1596-97


...that the Gregorian year corresponding to the second half of the Maya year had mistakenly been taken as corresponding to the said day-sign. In other words, in the extract given, Mani gives 13 Kan as year-bearer in 1593 and 1 Muluc as year-bearer in 1594. Dreamspell has 13 Kan as 21st July in 1594 and 1 Muluc as 21st July 1595. However, in the same paragraph, Mani says that 21st July1596 was 4 Ix and first of Cumhu. That means that New Year 1596 (0-2 Pop) was about 25 days later. See UPDATE at the bottom of the page for the latest attempt at an explanation from the Dreamspell followers (PAN or the Planet Art Network)..

Dreamspell Correlation, New Year and Leap days

So, it seems that the Dreamspell count was supposed to be based on the correlation given in the Cuceb section of the Chilam Balam of Mani, in which the New Year of 15th August 1596 (0-2 Pop) equates to 3 Cauac, (correlation #584315), but the first mistake was to equate the year-bearer to 21st July. The second mistake was to equate it to 1597 instead of 1596. The third mistake was to then presume a New Year of 26th July - regardless of whether or not New Year in 1596 or any other year was celebrated on 26 July, if the Cuceb is taken as the "only recorded prophetic tradition of the Maya", then the integral New Year correlation for 1596 should surely be part of that. However, it seems that the 26th July as New Year has been taken from the notes of the editors (note 178, p111-112). Makemson, as already mentioned, did not agree with the 26th July as New Year, since it contradicted the correlations given in the Tizimin. The same note records the concept of a "Box Katun", in which two methods of adding leap days are given. These are adding a day to each of the Cauac years and adding 13 days at the end of a 52-haab Calendar Round. Archaeologists generally do not accept that the Maya added leap days, even after the conquest, and it seems that this was probably added to the text along with the Jesus Christ count and the 24-year katuns, but at least we can now see that the components of the Dreamspell Count did originate in the Chilam Balam of Mani, even though the various components are not consistent with each other.

The whole of the Cuceb section is summed up by the editors (note112 p.100) as follows:

" The entire section entitled "the squirrel" is an infelicitous [unfortunate] example of the old Maya art of prophecy, if it is prophecy".

Regarding Arguelles' claim that the correlation dates back to at least 1513, there is only one piece of evidence I found to possibly support this: In the Tizimin (Makemson p.32), it says that "four katuns ago" there was an understanding of the Moon. If the statement was written down in 1593, along with many others, this means that 4 katuns, or 80 years before 1593 was 1513. However, this statement is not directly applied to any of the several possible correlations in the Tizimin and Mani, much less the Dreamspell one that was seemingly based on a misinterpretation of the given dates.

28-day Lunar months

It seems that on p.19 of Mani, we have an explanation as to the inclusion of 28-day "months" in the Dreamspell system. Note 1 p.19 reads as follows:

"There is little agreement on whether the Mayas had the zodiac. Spinden suggests that a Maya zodiac is shown in the Peresianus codex and tells us that it has thirteen divisions represented by "animal figures holding the sun un their mouths associated with 13 sidereal months of 28 days each." Hebert J. Spinden. The Reduction of Mayan Dates 55, Fig 25."

In this case, it seems that the observation of a possible sidereal calendar (star calendar or zodiac) - divided into 13 signs through which the Sun passes over a year, has, in Dreamspell, been turned into a lunar calendar, by misinterpreting stellar periods as "months", or "moons".

Tzolkin Grid/ Buk-Xoc

In regard to point 4 above, we find that, just as Jose Arguelles claimed, there is a "Buk-Xoc permutation table" consisting of a 20 x 13 grid, in the Chilam Balam of Mani (p.181). It was originally located on the inside cover of the Mani manuscript, and consists only of the Tzolkin number grid (using Arabic numerals, not Maya bar and dot numerals), and there is no inclusion of the 20 day-names or glyphs down the side, although the editors note that

 "This permutation table becomes particularly useful and effective if the names of the twenty days are listed down the left side of the table. When this is accomplished, the day-name and its numerical prefix, for any given 260-day year, is readily available."

There are very few representations of the entire Tzolkin in the surviving codices, but the Aztec Codex Fejérváry-Mayer has 260 dots located around its perimeter, and there is a similar figure in the the Maya Madrid Codex, but these do not show the sequence of signs. However, the Madrid Codex also has an almost complete Tzolkin sequence running across several pages, and this sequence begins with 1 Imix. There is a similar, though harder-to-identify complete sequence running across several pages of the Dresden Codex, and this also begins with1 Imix.

The "Portals"

Nowhere is there any sign of the "portal days" in any of the Maya texts, and, just as Arguelles said, this was contributed by Tony Shearer. However, in The Mayan Factor, and also in Shearer's book, Beneath the Moon and Under the Sun, it is implied that the pattern was discovered by observing that the squares add to 28, in groups of 4. This is true for the pattern, but is also true for all squares on the Tzolkin, with the exception of the central column; every square, when added to its 3 other symmetrical equivalents - one in each quadrant of the grid, adds to 28. This reduces the meaning of the pattern to a totally subjective level.


While Arguelles' claims that the grid arrangement of the Tzolkin as shown in the Mayan Factor, originates from the Chilam Balam has been verified, some of his other claims have been found to be in error. In particular, his claim that the Dreamspell correlation originates in the Chilam Balam, while technically true - we have seen how it was generated from the information in the Chilam Balam - does not follow any correlation in the Chilam Balam of the 5 or so to choose from, but combines several mismatching bits of information to produce a correlation that doesn't agree with any of the calendrical information in the Chilam Balams.

The so-called "Tun-Uc" or 28-day lunar count seems to have originated from a stellar or zodiac period, which could, at a push, be called a "month", but not a "Moon" as in Dreamspell, since it had nothing to do with the Moon.

The July 26th New Year date continues to be disagreed on by various authorities - Malmstrom, for example, thinks it was fixed at July 26th before the conquest, others presume it was fixed after the conquest, and some think it was just the date of New Year at the time of the conquest, for 4 years, and some think it was just the date of zenith passage of the Sun. Makemson thinks the whole thing was a misinterpretation of Landa's work.

The pattern of "portals" superimposed on the Tzolkin grid is one of many possible superimposed patterns with equal meaning or lack of it.

Jose Arguelles has  posted an article called "There is No Conflict Between Dreamspell and the Long Count". In this article, Arguelles claims that the purpose of Dreamspell is to "displace the artificial and irregular Gregorian calendar", and that it "does not recognize leap day and so maintains the perfection of the synchronic order", but if it does not recognize leap days, then how come every February 29th, it freezes the Tzolkin? If leap days were not recognized, then the Tzolkin and Haab would run continuously, as if no Gregorian leap day had ever been heard of. The article concludes by saying that "People who believe that there is a conflict between the two systems are still immersed in dualism and are fixated on being right", and goes on to infer that anyone who disagrees is acting counter to evolution. So black is white and if you disagree, there is a pigeon-hole specially prepared for you. I guess that means I'm a arrogant dualist chimp, then. Oh well.


For more, see Mike Finley's updated analysis of Dreamspell:

 Jose Arguelles' Calendrical Dreams

* Here's an interesting discovery: 22nd January 2013 is only one day before the end-point implied by the Great Seal of America see The Eye in the Pyramid. 4th July 1776 -date of the Declaration of Independence as inscribed on the bottom of the pyramid, plus exactly 13 Katuns at 7200 days each - (93,600 days, which is one cycle of the Short Count) - one katun for each level of the pyramid, gives a termination point of 23rd January 2013, since the Declaration was signed 33 days after the start of the katun - (December 21st 2012) plus 33 days = 23rd January 2013.

McKenna's Timewave Zero originally terminated  on 17th November 2012, 33 days before ( 21st December 2012!


UPDATE: (December 2005)

Dreamspell follower, Treefrog has made some startling claims regarding this 18 Zac 11 Chen date - he claims it is written in a code known as the "language of the Zuvuya". Sharer (in The Ancient Maya) describes this on p.406 as "the language of the Zuyua", and says it was used by the Putun Maya (a ruling political group in the Yucatan at the time of the Spanish conquest), "to ensure that no Yucatec Maya impostors had crept into their ranks". Treefrog points out that the two haab days, 18 Zac and 11 Chen are 47 days apart:  He then points out that in 2012, there will be a 47-day discrepancy between the Quiche count (=traditional count or "true count") and the Dreamspell count, and infers some sort of connection ( here):

 "i'm not saying that the dreamspell was actually planned by the chilam balam scribes and that Arguelles found the 'code'... i imagine that both the scribes and Arguelles went where study and intuition took them, and the result is something of great value."

Treefrog goes on to point out that the Tzolkin days 11 Chuen and 3 Kan are also 47 days apart:

 "It simply came to me: 11 Chen and 18 Zac occur 47 days apart. 3 Kan and 11 Chuen occur 47 days apart. Therefore in a Lunar Wizard (2 Ix) year 3 Kan will occur on the day 11 Chen and 47 days later 11-Chuen will fall on 18 Zac....In the year 2012 the traditional count and dreamspell with have a 47 day difference so that the day 3-Kan traditional will be 11-Chuen (Jose's galactic sig) dreamspell." From 13moon tribe

UPDATE: April 7th 2006:

My original interpretation of exactly what Treefrog meant by this, has been rejected by him, since he says he was not saying anything about July 21st, which was an integral part of what I thought he meant. Since he has rejected this, (and I have now removed the interpretation, as he requested), the statement seems to be even less meaningful than it originally seemed to be...

As further evidence that the Maya froze the year to July 16th (Julian), Treefrog says that

"the Chilam Balam Books of Chumayel takes the initial calculation date from July 16 (26) 1555, While the Tizim starts on July 16 1581."

However, the Tizimin actually starts in 1593, (but does not give the Julian day or month), and there is no mention of 1581, so this is not actually evidence after all.

Treefrog cites a Jose Arguelles article called Adjusting the Discrepancies within the Cuceb-Chilam Balam Correlation , which explains that there is a 5-day difference in "tone" - the number that accompanies the daysign, between the Chilam Balam and Dreamspell. Dreamspell has 2 Ix as 21st July 1596 so according to this article by Arguelles, the equivalent date generated by the Cuceb correlation should be 10 Ix, but it is actually 4 Ix. Therefore, the Arguelles article is irrelevant, since it does not explain the discrepancy after all.

UPDATE : April 7th 2006:

Treefrog has sent this: "...also, for accuracy's sake you may wish to change the part that says Jose's essay on 'adjusting the discrepancies in the cuceb/chilam correlation' is faulty because it explains a change of five not four... the change of five is to that of part three of the codex perez. I can see how you thought 'unadjusted' meant original, but what he means is unadjusted from it's latest recording."

Treefrog has now written a huge article that can be found here (UPDATE Jan 2010: the link to the 38-page article has now been updated) .

UPDATE: Jan 2010:

Treefrog 's link above has been updated, and he has now provided a shorter 6-page condensed version HERE