Khayyam's verse, as rendered into English by FitzGerald, presents a very nice opportunity to demonstrate what can be done. As one appreciative reader commented "I wouldn't have thought what you are doing would have been possible." Or as an old friend of mine wrote after receiving a copy of The Anagrammed Bible: "Thanks for sending me your fascinating book. It gets marveled at by everyone who sees it. We're all amazed to see that our intuitive notions about the space of all strings seem to be very wrong. But I don't know whether to think that the set of meaningful strings of characters that can be made from a given string of characters is (in English at least) much larger than I would have thought, or whether you just have a very efficient method of finding the few that exist." Actually, it's a little of both. But more important are the literary implications, which are two-fold. 

First, a whole new territory of creativity is now opened up. As Mike Keith stated in the Introduction to our Anagrammed Bible: "By imposing arbitrary textual or semantic constraints, writers create significant literary works that would probably never be imagined were it not for the rigorous discipline of the constraint" and: "this often results in stating things in ways that we wouldn't otherwise think of in a million years." This of course runs counter to the whole free-form thrust of the 20th Century, with its wholesale abandonment of all constraints, where anything is regarded as poetry as long as it is different from ordinary speaking - and sometimes even that, as long as it is broken into a poetry-looking format rather than written out in paragraph style. This democratization has lead to the ultimate absurdity of a web-site that claims to be publishing 3.1 million "poets"!

Second, the peculiar fascination of an anagram, which is to see the same or nearly the same thought re-expressed using exactly the same set of letters but none of the same key words, is an occasion for calling renewed attention and bringing new life to old masterpieces that may have fallen into neglect.

I have chosen to do the Rubaiyat on a quatrain-by-quatrain basis, rather than as a whole. My personal preference is for the smaller unit, because I like the way a tighter constraint results in a more identifiable "anagrammatic flavor". If taken on as a whole it might very well have been possible to have maintained the aaba rhyme, but it was my feeling that 3 rhymes in just 130 letters would leave too little maneuvering room to accomplish the other goals I had for meter, alliteration, grammar, aptness, and style. Even as it is, I will sometimes slip in an extra syllable, like FitzGerald also did (or omit one - but only at the beginning of a line). The equivalent of these departures from a perfect iambic is the occasional repetition of a key-word. Sometimes this is done intentionally, to preserve a parallelism, but more usually it is for the same reason that has always motivated authors to allow minor constraint deviations, namely that a slight relaxation is required to make possible a particularly fortunate expression that is in the vicinity.

In anagramming, it is of the utmost importance to make sure one is starting with the correct text. With the Rubaiyat this is complicated by the fact that FitzGerald did so much revising. There are three possible approaches: 1) select a single edition; 2) do all variations; 3) try to come up with a complete set by drawing from all the versions. In spite of the fact that it leaves me open to criticism from those who might have preferred different choices, I have decided on #3, because of its appealing compromise between the simplicity of #1 and the comprehensiveness of #2.

The anagramming of poetry involves more than just the ordinary constraints of: aptness to the original; using the same letter set; and avoiding re-use of any of the original's key-words (unless for special effect). To these it adds the following constraints:

1) Ending up with the same number of syllables. Actually there is some slight wiggle room here. In the case of a rubai, containing four lines of iambic pentameter, there are normally 40 syllables. However two of FitzGerald's have some lines with an additional unaccented ultimate syllable - #33 containing 43, and #99 containing 44. I have not taken it as a requirement to necessarily maintain precisely these counts. For example my anagram of  33 has 40 syllables, while my anagram of 100 has 44 syllables. These are infrequent exceptions, and 95 per-cent of the time it is 40/40. The thing you have to be careful of while in the process of composition, is not to start deviating too far from the vowels-per-syllable figure that characterizes the original. This can vary from 1.07 (almost every syllable containing only one vowel) to as much as 1.46 (nearly half of the syllables containing more than one vowel). 

2) Maintaining a certain rhyming pattern - in this case abcb. Wherever I may have ended up with aaba (for example 93), it has been purely coincidental, just as, I suppose, were FitzGerald's aaaa's (29, 35, 69, 76)

3) Getting the accent's all in the proper place. Again, I have taken liberty, in roughly the same degree of frequency as did FitzGerald, to replace an iambus with a trochee or an anapest, occasionally.

As to my method of proceeding, sometimes I will start by doing a rewrite without very much thought at all for the available letters, just to get some ideas going. Then I'll determine how far off it is from being an anagram, and begin deleting words that cannot possibly be used and replacing them with other, less obvious ones. Here are three examples, showing first the original, second the non-anagam paraphrase, and third the final anagrammed result:
 

I think it is fairly obvious how relatively vapid the non-anagrammed result is, compared to the anagrammed one - pretty much on a par with some of Talbot's and Whinfield's less inspired efforts. This is because of how the anagram constraint makes you really have to stretch to come up with ways of saying things that are vastly different from what would naturally first pop into your mind.

Actually, I don't normally do it this way. It takes longer to completely rework something that winds up most of the time having to be nearly totally discarded. Instead I'll start by locating a pair of rhyming words which have some relevance to the topic, and that use up some of the excessivley over-abundant letters, if there are any. Then I will look for available key-word synonyms, appropriate modifiers, etc., letting the constraint control my choices right from the start. If I begin with a non-anagrammed paraphrase, I run the risk that if the letter-set isn't sufficiently challenging, some of the relatively uninteresting starting material might wind up getting preserved.

Other works you may also find interesting can be located by visiting RichardBrodie.com