commit a bunch of meta files

2 files changed, 49 insertions, 0 deletions

diff --git a/notes/process.md b/notes/process.md
new file mode 100644
index 0000000..bd3a516
--- /dev/null
+++ b/notes/process.md
@@ -0,0 +1,34 @@
+
+This is an ordered list of testable steps to get to a minimal dat client in
+just about any language.
+
+sleep
+  read/write headers
+  read/write individual elements as raw bytes
+  read/write contiguous batches
+
+registers
+  read elements by index
+  verify signature by index (not leaves)
+  append elements
+
+drive
+  data register to a single file
+  single file to a data register
+  print metadata tree ("ls")
+  create metadata tree for a directory
+  directory to metadata+data registers
+  registers to directory
+
+sync
+  send/receive messages to a known host
+  pull register from a known host
+  wrapper command
+
+
+
+
+Shortcuts:
+  key/value store instead of SLEEP files
+  pull/read/to-file only
+  don't verify any hashes (!)
diff --git a/notes/spec_todo.txt b/notes/spec_todo.txt
new file mode 100644
index 0000000..b361f70
--- /dev/null
+++ b/notes/spec_todo.txt
@@ -0,0 +1,15 @@
+
+Things that were underspecified in the spec/whitepaper:
+
+- endianness of magic (big-endian)
+- uppercase/lowercase of algorithm name (inconsistent in one place)
+- what if there is only a single chunk/entry in a register tree? then a leaf
+  acts as a root?
+
+"You can use the byteOffset property in the Stat meta- data object to seek into
+the right position in the content for the start of this chunk." => unnecessary,
+node.js specific?
+
+Clarify: appending to tree SLEEP results in writes into middle of file (for
+root nodes). This seems not-great for performance (can't bulk-write). Better to
+cache or work in RAM then batch commit?