[Tutor] Roman to digital (pseudocode)

[Clay Wiedemann]

I am new to python and programming, so not sure if I should weigh in
here. Nonetheless . . .

Creating a dictionary seems fair, but I wonder about using pairs
rather than single letters. Roman numerals are constructed from strict
rules other than the allowable letter set -- here is the relevant
wikipedia bit:

Rules regarding Roman numerals often state that a symbol representing
10x may not precede any symbol larger than 10x+1. For example, C
cannot be preceded by I or V, only by X (or, of course, by a symbol
representing a value equal to or larger than C). Thus, one should
represent the number "ninety-nine" as XCIX, not as the "shortcut" IC.
However, these rules are not universally followed.

Although the shortcut may be easier for people, the universal notation
may be easier to define programmatically. I am wondering what would be
the best way to create the rules

 - A dictionary will help you look up values, but not rules. It does
not retain its order and order is essential. Instead, create a tuple
of the roman numerals in ascending order (roman). Create a paired
tuple with the base 10 value (baseten).

Now get an element from the string you want to test -- use the index
value in roman to test the rules -- you will know what is higher,
lower, and the same

Then look for things like the following:

- does the following element have a lower index in roman? if yes, you
know the value (match indexes from the tuples) -- step forward
- what if the follower is the same? then check the one after, if it is
the same (and the following digit is lower) then you have a value. --
step over the last matching digit
- if the following value is higher, it must be by only one unit, if so
you have a value, but also a new rule: the following digit must be
lower than the last digit of the pair.

and so on.
not pseudocode I know, and I am certain there are better ways to do
this, but maybe something here helps.

-c














On 3/13/07, Bob Gailer <bgailer at alum.rpi.edu> wrote:
> Deep breath .... and big step back.
>
> The problem is not just how to code this in Python, but how to parse a
> language (in this case Roman Numbers).
>
> I have studied formal language theory but am not an expert. So here's my
> take on an algorithm that would save a lot of stress.
>
> Create a dictionary with
>     keys for each roman letter and each "pair" (iv, ix, xl, xc, etc)
>     values are the corresponding decimal values.
> Starting at the left of the roman_input and repeating until the end of
> the roman_input
>   Take the next pair of roman_input.
>   If it is in the dictionary
>     add the decimal value and step to the next letter.
>   Else
>     take the leftmost letter of that pair
>     If it is in the dictionary
>       add the decimal value and step to the next letter.
>    Else
>      complain about bad letter
>
> --
> Bob Gailer
> 510-978-4454
>
> _______________________________________________
> Tutor maillist  -  Tutor at python.org
> http://mail.python.org/mailman/listinfo/tutor
>


-- 

- - - - - - -

Clay S. Wiedemann

voice: 718.362.0375
aim: khlav
wii: 3905 4571 6175 2469
twitter: seastokes