diff --git a/.gitattributes b/.gitattributes
index 555cb8e3e1e5909bad260d8f933cfdcd5193bc58..51415c4463ff34bf469853785e28cc167c0f7e67 100644
--- a/.gitattributes
+++ b/.gitattributes
@@ -6,6 +6,7 @@
 *.xcf filter=lfs diff=lfs merge=lfs -text
 *.bbl filter=lfs diff=lfs merge=lfs -text
 console_toolkit/console_toolkit_2.pdf filter=lfs diff=lfs merge=lfs -text
+*.html diff
 * !text !filter !merge !diff
 *.pdf !text !filter !merge !diff
 *.png !text !filter !merge !diff
diff --git a/pdf/PDF.pdf b/pdf/PDF.pdf
index 03f08cd9504070a5b38da33745aae3b3977b055b..d2518ea7a6f34b1f95a303979ea627dbd3a02f0f 100644
Binary files a/pdf/PDF.pdf and b/pdf/PDF.pdf differ
diff --git a/pdf/pdf.html b/pdf/pdf.html
index f6a6f1fdbeb76fc11cc17cc08a059cc17ceb87c9..ead404a9828f4e20ae1dd84741a1223132b7a8f0 100644
--- a/pdf/pdf.html
+++ b/pdf/pdf.html
@@ -498,6 +498,26 @@ In short: Read out tables, patch them together differently, pray everything work
 
 ---
 
+# Putting it all together
+
+TFF preprocessing:
+- parse TFF
+- transform glyphs & widths into suitable structure 
+
+TFF subset generation:
+- wait until PDF fully specified
+- collect all text within PDF file
+- use set of chars to create TTF subset
+- save TFF subset, remembering the order of the glyphs
+
+PDF text writing:
+- define encoding (pro-tip: use UTF-8)
+- reencode all text within PDF to this encoding
+- generate PostScript CMap knowing encoding / TFF glyph order
+- generate reverse PostScript CMap (why?)
+
+---
+
 # Bonus: How to extend ttf?
 
 Steps: