summaryrefslogtreecommitdiff
diff options
context:
space:
mode:
authorjoey <joey@0fa5a96a-9a0e-0410-b3b2-a0fd24251071>2006-09-28 06:56:52 +0000
committerjoey <joey@0fa5a96a-9a0e-0410-b3b2-a0fd24251071>2006-09-28 06:56:52 +0000
commitaf05e7badc4e6524500c81489a5298c04e6dd502 (patch)
tree4680b807d6e5eeba0c30fd0f9a738789b40645b9
parent65c8cf753896ba4415feb46739fb4000ecada9ad (diff)
web commit by EthanGlasserCamp: another entry on my wish list
-rw-r--r--doc/todo/toplevel_index.mdwn22
1 files changed, 22 insertions, 0 deletions
diff --git a/doc/todo/toplevel_index.mdwn b/doc/todo/toplevel_index.mdwn
index 90137f4b2..d766db2b4 100644
--- a/doc/todo/toplevel_index.mdwn
+++ b/doc/todo/toplevel_index.mdwn
@@ -19,3 +19,25 @@ Some inconsistences around the toplevel [[index]] page:
if (! $skip) {
$path.="../";
+---
+
+> I would like to suggest another tack, namely a bigger, better special case.
+> The basic idea is that all indices of the form foo/bar/index get the wiki path foo/bar.
+> This makes some things more elegant:
+>
+> * All files having to do with foo/bar are in the foo/bar directory, rather
+> than the (admittedly minor) wart of having the index be in foo/.
+> * This sort of addresses [[bugs/broken_parentlinks]] in that example/ is
+> guaranteed to be a valid path. (There might be no index there, though.)
+> * This is more in line with standard HTML practice, as far as I understand it,
+> namely that linking to a/b means a/b/index.html rather than a/b.html.
+>
+> This would change the inline plugin in strange ways -- I think if foo/index.html
+> contains \[[inline "* and !*/Discussion"]], it should skip inlining foo/index.html
+> explicitly, but would inline index pages in child directories
+> foo/bar/baz/index.html as bar/baz.
+>
+> It always bothers me that foo/bar/ files need a foo/bar.html in front of them,
+> rather than a foo/bar/index.html, as is (to my mind) traditional.
+>
+> Ethan \ No newline at end of file