My book waits until the ninth chapter before introducing these cruel words for which gender must be explicitly remembered, rather than derived from form.
It's funny everyone dislikes this sort of thing, and yet everyone agrees to them.
I was trying to be positive, but the truth is I don't give a damn about the spirit that lives in the computer. The machine is a tool which has no sentimental value to me whatsoever. If my style successfully kills a spirit, somehow, then it means nothing to me. Let it be replaced by its better.
I'll briefly try to rephrase one last time. In principle I'd like the machine to save me time allowing me to do things I wouldn't otherwise. Most of what I want to do otherwise is understand. Programming can help me do this by encouraging me to refine and manipulate my structural and algorithmic models (it makes these more clear like writing to distill thoughts), moving towards some hypothetical correct understanding. This would be my justification for looking for alternative solutions after a problem has been solved. Perhaps this is the crux of the cultural difference, but I will avoid conjecturing your motivations.
That's my point. The machines shouldn't compute as we do. It's foolish to have the machine compute as a human would, when this be wholly unnecessary.
Being essential is different from being how we compute. Computing definite integrals is an illustrative example. Humans compute definite integrals by finding an anti-derivative. You can hardly say however that such a form is more essential than Riemann integration. In fact the the proof of the fundamental theorem of calculus is based on the definition in terms of Riemann integration. Most methods of numerical integration are not so conceptually different from Riemann integration, and if circumstances demand would not be unwelcome just as symbolic integration is not unwelcome for humans. Graphing would be another but I won't elaborate.
Only I've an accurate mental-model of what Elision will be and may be.
I don't disagree of course, but guess what I meant by "can't be fully structured meaningfully" is that ultimately the allocation of phonemes and their representation is arbitrary. This is randomness you have to squeeze into a model.
I was recently reviewing some mathematics and realized the part of Elision that compresses the character tables may be partially related to what is known as super permutation. It's obvious how Latin declension and conjugation tables can be represented as tables, but the compositions of those tables can be represented as tables and so on and so forth.
I hadn't heard of super permutations before. I do see some similarity between your ideas and that. I'm not sure I understand what you mean by composition of conjugation tables though.
My solution is as follows: Get the index of the first possible such word, and of the last possible such word; now the searching is a bounds check on each word, needing no more dictionary access. Searching the auxiliary dictionary would mean just an additional bounds check. Isn't this neat?
Ah, I fell into the incremental search trap. Your solution is dramatically more elegant than the alternative (I wrote some pseudo-Scheme for heathen text using some fancy tricks and it still wasn't pretty at all).