International quotes

[Michel Fortin]

Le 2005-10-21 à 20:11, A. Pagaltzis a écrit :


> Keyboards have no way to enter bottom doublequotes at all, which

> is no wonder considering the common non-Unicode charsets do not

> even contain such a character. So the end result is that everyone

> writes German with straight doublequotes for quotations.


This sadden me, I find them nice these german quotes.

(I've found out that on my keyboard I can type shift-option-^ to get  
bottom double quotes. On the downside it seems I can't type an  
opening single quote.)



>> The other unanswered problem is that simple angle quotes could

>> be mistaken and replace a *true* greater than or less than, so

>> it may be better to forget about them.

>

> Not only that, but at least the right-pointing single angle mark

> <em >clashes </em >with HTML tag parsing. Granted, so long as the

> angle marks point “inward,” German-style (opening angle mark

> points right, closing points left), you can probably disambiguate

> 95-99% of the cases correctly.


I specifically ignored the HTML tag problem because I don't think it  
is a problem. SmartyPants tokenize the tags prior translating  
anything. If it wouldn't, think about what would happen to tag  
attributes :-) Anyway, my point is SmartyPants doesn't search for  
quotes inside tags and wouldn't mistaken a tag for a quote.

Parsing of `<<` wouldn't be easy either. It would probably be  
necessary to convert `&lt;&lt;` and `&gt;&gy;` to quotes, as a tool  
like Markdown will automatically escape less-than characters with  
entities if they are not a tag. And what about `<<b>>`? Markdown will  
convert it to `&lt;<b>>` and you would have to figure out a way to  
make a quote out of that. Or is it a tag?

I haven't realized until now how much complexity angle quotes (simple  
or double) could be.



> But from where I sit it seems impossible to disambiguate single

> angle marks that point “outward” from HTML tags. (F.ex., what

> about the letter <p>?)


This wouldn't work with `<p>`... it could work if you leave spaces as  
common French typography suggest (`< p >`), but then how do you know  
it isn't part of some mathematical stuff?

     1 < p > 2 <=> p > 2



> So I’d say a transliteration for single angle marks right out of

> the window.


One side problem I think of, and it applies to double angle marks as  
much as single angle marks, is that `<<` and `>>` are somewhat common  
operators in many programming languages. Just like the  
emphasis_by_underline problem with paths and programming symbols in  
Markdown, converting `<<` and `>>` could lead to display problems in  
comment boxes where sample code is outside a `<code>` tag. This  
wouldn't be catastrophic (and certainly a smaller concern than the  
underline problem), but it may be a good idea to think of this in  
advance.



> Actually you’re mixing up the styles. English would be

> transliterated like so:

>

> [...]


Right, my error. And all is really missing in the current SmartyPants  
is a special rule for `,,` as a lower double quote.



>> A notable exception is ”...” where both quotes are curled the

>> same  way as the closing quote in English (used by Swedish,

>> Finnish, and  Dutch (alternative quoting style) [^1]).

>

> This could be written in transliteration as

>

>     ''...''

>

> which results in two right doublequotes.


... and happens to work perfectly already with current SmartyPants. I  
somewhat thought SmartyPants tried to be "smart" in this case, but it  
does not. Great!

  * * *

Later:


> I don’t see the harm in having transliterations available. They

> would be useful when you really want to do something that the

> language settings would not allow. SmartyPants would just be

> useless to germanophones if they *had* to use these.


I completely agree with this way of thinking.

To summarize, an upcoming "international" SmartyPants could do this:


| Input    | Output                   |

| -------- | ------------------------ |

| "        | “smart” [customizable]   |

| '        | ‘smart’ [customizable]   |

| ``       | “                        |

| `        | ‘                        |

| ''       | ”                        |

| ,,       | „                        |

| <<       | «                        |

| >>       | »                        |


with some some presets for different languages.

Making `<<` work correctly alongside tags could be a problem however.


Michel Fortin
michel.fortin at michelf.com
http://www.michelf.com/