2 files changed, 68 insertions, 0 deletions
diff --git a/notes/background.txt b/notes/background.txt
new file mode 100644
index 0000000..9449861
--- /dev/null
+++ b/notes/background.txt
@@ -0,0 +1,10 @@
+
+## Libraries
+
+- [tablib](http://docs.python-tablib.org/en/latest/)
+- [records](https://github.com/kennethreitz-archive/records)
+
+## File Formats
+
+- column stores like parquet, arrow
+- data packages
diff --git a/notes/plan.txt b/notes/plan.txt
new file mode 100644
index 0000000..8f09fc0
--- /dev/null
+++ b/notes/plan.txt
@@ -0,0 +1,58 @@
+
+x write basic exmaple from TSV file
+x pass-through basic thing in pipeline
+x pretty printer (using column writing, term color)
+- convert
+    to/from TSV
+    to/from JSON
+- example apps
+    ls
+    stat
+    df, mount, something like that
+    trivial web server (or other thing that logs)
+- example datasets (eg, for benchmarking): compare TSV, AFT, JSON
+    million-line CDX
+    log file
+- manpage (?)
+- build .deb and installable
+- helper library
+    => R/W trails/wrappers
+    => header struct
+    => iterate rows (from input)
+    => pretty-print output based on tty status
+    => validation/check modes
+    => stream mode/helper for subprocesses
+- tests
+- compare with xsv command (?)
+- reimplement basic commands
+    cut (accept field names)
+    cat (combining files with compatible headers)
+    head, tail
+    wc (count rows, records)
+    format (accept field names)
+    grep/match/filter by column value?
+    paste
+    uniq (by column)
+    sort (by column)
+    join
+
+- extended commands
+    parallel (with column names)
+    shuf
+    comm
+    expand/unexpand
+    nl
+    seq
+
+ideas:
+- python stuff
+- C stuff
+- log log format integration
+- rust serde integration
+- aft-header: pretty-prints header as rows
+- aft-single: pretty-print single row (first?) as rows
+- aft-format (or printf?) "this {col1} to that {col2}" "some other column"
+- aft2json, json2aft
+- aft2html
+- aft-stats: sum, mean, stddev, min, max
+- aft-sql