diff --git a/README.md b/README.md
index e64e00b..a0de5be 100644
--- a/README.md
+++ b/README.md
@@ -55,6 +55,7 @@ Screenshots:
 ![byvoid](https://github.com/ppwwyyxx/wechat-dump/raw/master/screenshots/byvoid.jpg)
 
 ### TODO
++ Export all messages to libchat, and render messages from libchat
 + Search by uid/username
 + Skip existing files when copying android resources
 + Fix rare unhandled types: > 10000 and < 0
diff --git a/libchat/README.md b/libchat/README.md
new file mode 100644
index 0000000..a38a3e6
--- /dev/null
+++ b/libchat/README.md
@@ -0,0 +1,19 @@
+
+# A Universal Chat History Management
+
+## Fields
+
+Table Message:
++ source(enum): can be 0:wechat, ...
++ time(time)
++ sender: id of the sender. id follows the name convention of source
++ chatroom: id of the chatroom (can be a private chat with a sender, or a group chat)
++ text: text content. emojis are encoded according to the source
++ image: binary of the image.
++ sound: binary of an audio
++ extra_data: other possible data specific to the source.
+
+Table Wechat:
++ avatars
++ nicknames
++ emojis
diff --git a/plot_num_msg_by_time.py b/plot_num_msg_by_time.py
index 92ec33f..1e17c7b 100755
--- a/plot_num_msg_by_time.py
+++ b/plot_num_msg_by_time.py
@@ -1,7 +1,7 @@
 #!/usr/bin/env python2
 # -*- coding: UTF-8 -*-
 # File: plot_num_msg_by_time.py
-# Date: Thu Dec 25 00:35:57 2014 +0800
+# Date: Tue Feb 03 21:12:30 2015 +0800
 # Author: Yuxin Wu <ppwwyyxxc@gmail.com>
 
 from lib.parser import WeChatDBParser
@@ -17,7 +17,7 @@ if len(sys.argv) != 3:
 
 db_file = sys.argv[1]
 name = ensure_unicode(sys.argv[2])
-every_k_days = 3
+every_k_days = 2
 
 parser = WeChatDBParser(db_file)
 msgs = parser.msgs_by_talker[name]
@@ -26,7 +26,7 @@ start_time = times[0]
 diffs = [(x - start_time).days for x in times]
 max_day = diffs[-1]
 
-width = 30
+width = 20
 numbers = range((max_day / width + 1) * width + 1)[::width]
 labels = [(start_time + timedelta(x)).strftime("%m/%d") for x in numbers]
 plt.xticks(numbers, labels)