more

PyCon2012/picloud.rst

@@ -47,3 +47,4 @@ The Journey to Give Every Scientist a Supercomputer
 - OpenAFS for environment storage
 - Getting data to the cloud is a feature they have
 - Over 100,000,000 jobs processed to date
+- Pay by the millisecond, not the hour

PyCon2012/unicode.rst

@@ -0,0 +1,47 @@
+=====
+Pragmatic Unicode - or - How do I stop the pain?
+=====
+
+- Entire presentation made with unicode
+- 256 symbols is not enough for the world to communicate using text
+- Started with encodings for 1-byte, then 2-byte
+- Now we are using unicode
+  - Assigns characters to code points (integers)
+  - 1.1M code points
+  - 110K assigned
+- Pile of poo character was covered
+- Python can address characters by their unicode name
+- str stores bytes
+- unicode stores 'code points'
+- Unicode encode method turns code points into bytes
+- str decode turns bytes into code points
+- Encode and decode can't always work (if ordinal is out of range)
+  - Error handling: my_unicode.encode("ascii", "replace")
+    Or xml/html character replace
+
+- Python implicitly converts your 'assumed' ascii data when concatenating
+
+There are two things: Bytes and Unicode
+You have to know what you're dealing with
+
+- str in python 2 is a byte string, str in python 3 is a unicode string
+- Python 3 will never implicitly convert bytes into unicode
+
+Mixing bytes and unicode is always **PAIN**
+You are forced to keep them straight in Python 3
+
+In python 3, th data you get back from a read operation depends on how you open it (r vs rb)
+
+locale.getpreferredencoding() will get the default for read
+
+stdin and stdout are pre-opened filed and must be handled or wrapped
+
+"There is no string"
+
+Fact of life: Sometimes you are told the wrong encoding. It sucks.
+
+I/O is always bytes.
+
+Test!
+
+http://bit.ly/unipain