Dynastic character numbering - r1433

[Magnate]

Evening all - with the help of a lurker I've implemented roman numeral name suffices, so those of us who die a lot can have automatic numbering for new characters if we want.

So, if you give your character a name ending in " I" (that's a space and then a capital i), it will notice that, and put a counter in the savefile. All subsequent characters will default to the same name (even if you change the race/class) with the suffix properly incremented (up to 255). If you change the name to one that doesn't end in " I", it will turn off the numbering and all subsequent characters will get the same name.

It does not work with any suffix other than the first - that would need an entire "de-romanify" function which we didn't write. This means that you cannot change the suffix of a dynasty once you've started (e.g. if one dies an embarrassing death) - you either have to carry on or restart at I.

Please test and report any bugs in this thread. We used a spare byte in the savefile, so there should be no savefile compatibility issues.

(Edit: do not forget the space. It won't work without a space as the penultimate character, and a capital i as the last.)

[Atarlost]

A number arabicizer wouldn't be too hard to write and if you cutoff at 255 you could just read look for a last word with nothing but CLXVI in it. No counter needed.

[Magnate]

A number arabicizer wouldn't be too hard to write and if you cutoff at 255 you could just read look for a last word with nothing but CLXVI in it. No counter needed.

If you had an arabiciser/deromanify, you wouldn't need to stop at 255 (that's a limitation of the 1-byte counter). If anyone wants to write one and send it to me, I'll happily expand this functionality to accommodate changes to the suffices, starting at numbers other than I etc.

[Cauldron]

A number arabicizer wouldn't be too hard to write and if you cutoff at 255 you could just read look for a last word with nothing but CLXVI in it. No counter needed.

I think you mean CCLV.

Roman Numerals!? They never even *tried* to teach us that in school!

[Marble Dice]

CCLV is indeed 255, but when he said CLXVI he was referring to the set of characters necessary to recognize roman numerals no higher than 255 i.e. you wouldn't even need to look for the characters D or M. If fact, the set [IVXLC]+ would get you all the way up to 399 or CCCXCIX.

[Cauldron]

CCLV is indeed 255, but when he said CLXVI he was referring to the set of characters necessary to recognize roman numerals no higher than 255 i.e. you wouldn't even need to look for the characters D or M. If fact, the set [IVXLC]+ would get you all the way up to 399 or CCCXCIX.

OK that makes sense. I think 399 is properly CCCIC, though. Which is the problem with roman numerals. The problem is incrementing the roman numeral. Rather, converting from arabic to roman. Though someone has probably written a conversion function.

[Marble Dice]

Some sources consider it improper to prefix a larger roman numeral with a smaller one to indicate subtraction when the larger roman numeral is more than one order of magnitude larger than the smaller one. If you go by that rule, then IX (9) is the last time you can use I as a prefix, and the only prefix you would use on C would be XC (90).

But there's tons of inconsistencies like that. Sometimes you see IIII instead of IV, sometimes you see IIX instead of VIII, and then sometimes people make an exception for numbers over 1000, since the characters after M (1000) have bars over them, which is difficult to render. So instead of M,V-bar for 4000, sometimes you'll see MMMM.

Arabic to roman is significantly easier than roman to arabic, in my opinion, although you can find algorithms for either. If you're doing arabic-to-roman you can just pick a consistent set of rules and run with them, but when you're doing roman-to-arabic, you may want to accept multiple methods of writing the numbers, or you might want to exclude poorly-formed numerals.

I'm glad I never had to do math with roman numerals.

[d_m]

I'm glad I never had to do math with roman numerals.

I remember hearing some opinion from a historian (or maybe some random wacko on the internet) that ultimately Roman engineering was limited by how hard doing arithmetic in Roman numerals was.

[Magnate]

I remember hearing some opinion from a historian (or maybe some random wacko on the internet) that ultimately Roman engineering was limited by how hard doing arithmetic in Roman numerals was.

I think the fact that they didn't have the number zero also made engineering pretty hard ...

... and here's me thinking people would just say "Gee, I've wanted this functionality for ages - thanks"

[Colbey]

I've hoped for this feature for a long time. I express my gratitude!

[Magnate]

I've hoped for this feature for a long time. I express my gratitude!

Heh, thanks. eatomyx sent me a de-romanifier, so the next update will enable people to use suffices other than I, should they so wish. It could also dispense with the savefile byte, by parsing the suffix directly.

[Nick]

...to use suffices other than I

Suffixes suffices.

[Magnate]

Suffixes suffices.

Yeah I wondered about that. I was assuming suffix is like index, where indeces suffices. But maybe it's more like prefix.

[Magnate]

This functionality has been overhauled in r1512 (thanks to Marble Dice and Peter Denison). It should now increment pretty much any roman suffix, including insane ones like MMMDCCCLXXXVIII (which I've tested). Anyone who reaches 4000 consecutive characters should take a break from playing Angband and go and find a character set with the roman numeral for 5000 in it (M with a bar over the top).

If for any reason it can't deal with your suffix, it should print an error message. If you get this error, and think your roman numeral suffix is valid, let me know.

You still need the space before the suffix, in all cases.

[myshkin]

Anyone who reaches 4000 consecutive characters should take a break from playing Angband and go and find a character set with the roman numeral for 5000 in it (M with a bar over the top).

Actually, that should be a V with a bar. Wikipedia also makes reference to a notation with parentheses, but I haven't seen that before.