Focus on source XML manipulation, introduce indexes

Instead of XHTML or Atom, so resulting files get generated out of
custom XML structures now.

This was needed to get indexes that won't depend on Atom (and to avoid
the duplication).

Now indexes can be included with something like the following:

<notes xmlns="" number="5"
       src="../build/notes-by-modification-date.xml" />

1 files changed, 19 insertions, 10 deletions

diff --git a/tools/README b/tools/README
index 43dfea4..bbd251d 100644
--- a/tools/README
+++ b/tools/README
@@ -7,8 +7,8 @@ works for now, and is pretty simple.
 
 Editing:
 
-- skeleton.xml is the primary skeleton to start with, for *.xhtml
-  files in ~/homepage/src/.
+- skeleton.xml is the primary skeleton to start with, for *.xhtml files
+  in ~/homepage/src/.
 
 This stage is mostly needed to get unified styles, metadata
 properties, and other bits in case if I'll change them. The same can
@@ -19,21 +19,30 @@ Editing is done with the emacs html-mode, with
 <https://github.com/defanor/html-wysiwyg>.
 
 
+Indexing:
+
+- xml-notes-dump.xsl (with some help from publish.sh) is used to
+  create notes-dump.xml in the build directory.
+- xml-notes-sort.xsl sorts that to create
+  notes-by-publication-date.xml and notes-by-modification-date.xml.
+- xml-notes-limit.xsl is used to extract the most recent notes.
+
+
 HTML:
 
-- xml-to-html.sh translates source files into actual (X)HTML in
-  ~/homepage/ by applying xml-to-html.xsl to them.
+- xml-to-html.xsl is used to translate source files into actual (X)HTML
+  in ~/homepage/.
+
+Indexes can be included into those using the <notes> element (so it
+should run after those are built, if they are used).
 
 
 Atom feed:
 
-- html-to-atom-dump.xsl dumps HTML files into a single atom feed.
-- atom-sort.xsl sorts those by modification date.
-- atom-limit.xsl limits them, taking the newest entries.
-- html-to-atom.sh finds source files, and pipes them through the
-  XSLTs.
+- xml-notes-to-atom.xsl translates a dump of notes (such as
+  notes-by-modification-date.xml) into an atom feed.
 
 
 Publishing:
 
-- publish.sh runs the above two and uploads the files with rsync.
+- publish.sh runs all of the above and uploads the files with rsync.