[pmwiki-users] Markup for Compose?

Pico pmwiki at ben-amotz.com
Wed Sep 13 11:47:35 CDT 2006

Stirling Westrup <sti <at> pooq.com> writes:

> I'm doing a survey of opinions here. Lately I've been thinking of
> writing a recipe that allows for character composition and definition of
> new character entities as one goes. I'm not sure what a reasonable
> markup would be. I'm tempted to use a backslash, but I fear it would
> interfere with existing markup. The idea would be to allow someone (like
> me) who occasionally writes in another language, but who doesn't have a
> keyboard that supports it, to insert accented and special characters in
> a simple way without cutting and pasting from a character map.
> I'm thinking of something like this:
>   caf\e' => café
>   d\u` => dù
> I considered just making the various ASCII accent-like characters into
> combining chars, but it could cause mayhem:
>   don't == dońt or donť, depending

In my view there are two issues: (1) simplifying entry and (2) facilitating

Taking the example of words with accent marks, such as café, we could (as Sandy
points out), just use the PmWiki supported HTML special characters, like this:

caf&eacute;  or

1. Entry
Ok, that's an option, but its not simple to enter: you need to remember codes to
be able to enter them, and recognize them in order to be able to decipher the
markup in an editing screen; you have to enter many extra characters; and the
sequence is not intuitive.

2. Searches
What happens when you try to search for "cafe" will the search miss pages where
the markup contains "caf&eacute;"  or "caf&#233;" or even café?  Yes, according
to a quick test I just tried on PmWiki.org, a page with all three examples is
missed by a search for "cafe."

In thinking about alternative ways to enter accented characters, it would be
great if the markup could contain the plain ascii text to facilitate searches.
For example, instead of:

  caf\e' => café

What if the markup was more like:

Or even:

A search for "cafe" should return pages with either of these formats (I just
tested this as well on PmWiki.org).

Of course, these simple examples raise their own issues about what string
substitutions are being requested and the extent to which they should be applied
 (for example, what if the word has more than one instance of the letter "e" and
only one is accented).  


More information about the pmwiki-users mailing list