Thanks, Steve, for the comment!
I agree with most of what you said, although I wish to add that even though phonemic decoding is slower than logographic decoding, as you pointed out, the latter takes longer to attain (3 years versus 3 months to 1 year), it is worth noting that once a person has practiced reading words using the first method, s/he does not use the latter method. In other words, I don't think that these 2 methods are mutually exclusive. Do you agree?
I would favor a mild regularization. Something less radical than house stile. Just dropping the surplus characters would have a major beneficial effect. Some silent letters are markers so these are not surplus. The e in rope is not surplus since it marks a long vowel [róp]. No real need to do much more to remove the roadblocks to reading. Spelling is another matter.
/'siz*rz/ has over 1000 plausible spellings. The first phoneme /s/ in the initial position can be spelled s, c, and sc. 85% of the initial /s/'s are spelled one of these ways. The short i has a number of spelling patterns in English. i is clearly the best choice but not necessarily right.
/z/ might be ss, zz, s, z, ... The syllabic R could follow any vowel letter but er and or are the most likely candidates. The plural /z/ is usually s in English but there are other possibilities.
To get the number of probable spellings, just multiply 4 x 4 x 4 x 1 = 64. That provides lots of ways to misspell but not as many as when one uses all of the possible spelling patterns per sound. Alexander Ellis found over 10,000 different ways to spell 6 phonemes.
If children learn the high frequency spelling patterns they would be able to read 85 % of the words in the dictionary. There is no need to learn all of the possible ways to spell a sound. 4 ways to spell a vowel and 2 ways to spell a consonant is usually sufficient. I have a list of ways to spell or represent each spoken phoneme (patterns of spelling) I need to separate the high frequency spelling patterns from the others. Most phonics programs have done this but usually don't indicate the phoneme being spelled in English.
English is a flawed phonemic writing system. It is not a flawed logography. As a system of meaning signs, there are few flaws. The spelling of a meaning sign (or word sign) does not have to be phonemic or consistent.
There are very few writing systems that are as good as a dictionary key. Spanish and Italian are 85% phonemic or 85% as good as a perfect one and only one symbol per sound dictionary pronunciation guide. There is more than one way to measure phonemicity but they all give us a way of comparing writing systems.
Those who speak the language can learn how to write it in a near dictionary key in about 3 months. To learn 4000 words in a logography often takes more than 3 years. Those who have learned the system can read aloud any written word. They will not necessarily understand the meaning of the words but they can understand it as well as when someone read the text to them.
Logographies work quite well once learned. In fact, logographies can be read quicker than an alphabet where each word is read as a string of sound signs and converted to a spoken word or its sub-vocal equivalent. Reading by "sounding out" is so slow that this strategy is quickly dropped for high frequency words.
Logographies have a down-side, they take forever to learn and require good memory skills. Almost anyone can learn 40 paired associates (or grapheme -phoneme relationships). Some can master this learning task in a day others may take a couple of months.
Learning English is not a simple paired-associate task. Each phoneme has to be associated with multiple spelling patterns or graphemes.
If you feel that English (it spelling system) is perfect, please provide your rationale for asserting this. Thank you!