How much memory may a HPPL program use on a G2?

[komame]

(09-11-2023 05:39 AM)Tyann Wrote:  For crosswords, the grids are formed with UPPERCASE letters without accents, so if in the grid I have an F 2 empty squares then an E and an empty square, I enter f??e? the programme returns 'forest'(forêt) and 'ferret'(furet) as solutions.

I don't know your language, but in my native language, there are 9 diacritical marks (ą, ć, ę, ł, ń, ó, ś, ź, ż). They appear normally in crosswords, and you have to find an exact match for them. If in your language, crosswords always use letters without diacritics, then it's clear why you're doing this conversion. The question is, in such a situation, wouldn't it be better to have words in the dictionary with diacritics converted from the start? Searching in a dictionary without diacritics would be much faster. What about when a diacritical mark is the first letter of a word (if it can happen in your language)? In that case, it becomes even more complicated because you can't compare whether the last word in the list is smaller than what you're looking for to skip to the next list.
Anyway, I understand that during word comparison, you additionally iterate through the characters and sequentially replace diacritical marks with their Latin alphabet counterparts.

Best regards,