Re: Another patch (this one is important) to abc2ly

[Juergen Reuter]

On Mon, 2 Oct 2006, Laura Conrad wrote:

> I just tested it and in actual code, yours seems to do the same thing
> mine does,

N.B.: There should be a minor difference in the handling of white space 
before/after the hyphen, which however is not essential, I guess.

> except I find it harder to read.  What do you think is the
> advantage of yours over mine?

Your approach

>      a = re.sub ( '-', '- ', a)        # split words with -

applies a rule which holds for most cases, but not all cases. 
Therefore, you introduce another rule

> +    a = re.sub ( ' - - ', ' -- ', a)  # unless was originally " -- "

to compensate for the error in the first rule.  Why not fixing the 
original rule rather than adding another rule for compensation? 
Compensation usually increases the oevrall complexity and thereby makes 
the code difficult to understand.

As far as readability is concerned, well, this probably depends on the 
point of view.  Looking at your compensation rule, I find it difficult to 
understand, as it is not immediately clear to me, what happens if e.g. you 
have "---" or similar.  My suggested rule simply says: replace "-" by "- " 
if and only if there is not another "-" immediately before or after the 
"-", and no further compensation is needed (hopefully).  This is quite 
clear to understand, isn't it? ;-)

> The code I tested on was:
>
>   The normal way to type multiple syllables in ABC:
>
>   a = "mul- ti- ple syl- la- bles"
>
> Where your version produces:
> > > > a = re.sub ( '([^-])-([^-])', '\\1- \\2', a)
> > > > print a
> mul-  ti-  ple-  syl-  la-  bles
>
> ...

There shouldn't be a hyphen after "ple".  I hope, that's just a typo?

Greetings,
Juergen