In the process of translating About the PerlMonks FAQ in French, I'm stuck in being unable to have node title embedding html accented chars.

The problem comes when nodes are interlinked via [id://nodeid]: The default behaviour during the rendering to html is to generate the link according to the destination node title (as specified in What shortcuts can I use for linking to other information?). Using the functionality is straight forward when language is english, but when speaking another language the pipe syntax ([id://nodeid|title]) is mandatory to keep the correct meaning, even when interlinking nodes of same language.

That mean that I couldn't make a French PerlMonks FAQ entry point without "overloading" all title nodes (very error-prone practice). On the other hand, if throwing away accented chars may keeps node title meaningful for frenchies, I doubt this would be the case for other languages (I won't think about japanese...).

I'm aware that this kind of feature, could be a nightmare to provide as all the binding and searching stuff of Everything is impacted (noticeably, it should have a fallback allowing accented title chars being matched with unaccented search chars).

NB: Looking at my alter-ego FAQ translator (monsieur_champs) writings, it seems that he hadn't been as safest as me letting the browser decide what to do with extended chars, but the purpose was to have the FAQ translated, no?

____
HTH, Dominique
My two favorites:
If the only tool you have is a hammer, you will see every problem as a nail. --Abraham Maslow
Bien faire, et le faire savoir...

  • Comment on Special & Accented chars in nodes titles ==> [à la française]

Replies are listed 'Best First'.
Re: Special & Accented chars in nodes titles ==> [à la française] (!ents)
by tye (Sage) on Jun 28, 2004 at 15:51 UTC

    Titles are Latin-1 text (and not HTML). Do you need any non-Latin-1 characters for French? Link descriptions [that come after the pipe (|) in links] are HTML.

    So don't use HTML entities in titles. If you have a hard time typing accented characters on your keyboard, then you can type HTML entities someplace other than the title and then cut'n'paste the rendered characters into the title.

    A worse problem would be if your browser tried to send UTF-8 to PerlMonks instead of Latin-1, but that doesn't appear to be the case.

    Note how the title for my reply displays correctly but the title of your node displays as HTML entities.

    Update: Another problem could be if you have your browser set to override the content encoding that PerlMonks sends out with every page we serve. Our pages are served in Latin-1, but if you tell your browser to use Latin-2 or something despite what the web site tells it to do, then some accented characters will be displayed as the wrong character.

    - tye        

      Do you need any non-Latin-1 characters for French?

      Hopefully, non-Latin-1 chars are not required for French. Anyway, this is nice to have PerlMonks/Everything encoding behaviour precised somewhere (I may add a note to the SDC Master Plan Wiki about it)...

      Thanks

      ____
      HTH, Dominique
      My two favorites:
      If the only tool you have is a hammer, you will see every problem as a nail. --Abraham Maslow
      Bien faire, et le faire savoir...

        Do you need any non-Latin-1 characters for French?
        Hopefully, non-Latin-1 chars are not required for French.

        Well, there is Œ / œ ... but there does not seem to be a consensus on whether they are truely required for French. ;-)

        print "Just another Perl ${\(trickster and hacker)},"
        The Sidhekin proves Sidhe did it!

        nice to have PerlMonks/Everything encoding behaviour

        This is specific to PerlMonks. I don't know what other Everything installations use these days, but PerlMonks used to interpret titles as HTML until I fixed it because it was causing problems and had the potential for even more abuses.

        - tye        

Re: Special & Accented chars in nodes titles ==> [à la française]
by castaway (Parson) on Jun 29, 2004 at 06:36 UTC
    Just a note cos Im not sure if anyone has noticed it: Part of the problem is that the Preview page shows a node title containing HTML entities just fine, thus leading the poster to believe that it will also look fine afterwards, which it doesnt.

    (Apologies if this has already been fixed)

    C.