Devon wrote: > The order of the dictionary makes a lot of intuitive sense the way it is: > the pure symbols are at the top, the lettered symbols are at the bottom. As > you go down the first column, the pure symbols start with boolean functions, > proceed to simple math, then more complex math and non-math functions > affecting shapes, sizes and orderings of arrays; finally, the pure symbol > section ends with the more complex function composition operators.
> Putting it in a more lexicographic order according to "a." would make sense > if we thought people would be familiar with the order of punctuation > symbols, something that seems very unlikely: off the top of your head, do > you know if comma is before or after semi-colon? I agree. Also, since there are only about ~100 entries, and they all fit on one page, ordering of the DoJ isn't as serious a problem as a real dictionary (because even a linear scan will complete in reasonable time). > The useful redundancy might be to offer alternate groupings on other pages, > say by categories like "set membership", "complex numbers", etc. Yep, we could have one "tab" per collation or ordering; the standard one is the default tab, parts-of-speech ordering could be a tab, group-by-use (sets, arithmetic, etc) could be a third, and we can add more as we come up with them. -Dan ---------------------------------------------------------------------- For information about J forums see http://www.jsoftware.com/forums.htm
