Actually, this is not a design flaw, and it is not because of internals or performance.
It comes simply from the fact that functions in Python are first-class objects, and not only a piece of code.

As soon as you think of it this way, then it completely makes sense: a function is an object being evaluated on its definition; default parameters are kind of "member data" and therefore their state may change from one call to the other - exactly as in any other object.

In any case, Effbot has a very nice explanation of the reasons for this behavior in Default Parameter Values in Python.
I found it very clear, and I really suggest reading it for a better knowledge of how function objects work.

Python 3.x makes a clear distinction between the types:

• str = '...' literals = a sequence of Unicode characters (Latin-1, UCS-2 or UCS-4, depending on the widest character in the string)
• bytes = b'...' literals = a sequence of octets (integers between 0 and 255)

If you're familiar with:

• Java or C#, think of str as String and bytes as byte[];
• SQL, think of str as NVARCHAR and bytes as BINARY or BLOB;
• Windows registry, think of str as REG_SZ and bytes as REG_BINARY.

If you're familiar with C(++), then forget everything you've learned about char and strings, because a character is not a byte. That idea is long obsolete.

You use str when you want to represent text.

print('שלום עולם')

You use bytes when you want to represent low-level binary data like structs.

NaN = struct.unpack('>d', b'\xff\xf8\x00\x00\x00\x00\x00\x00')[0]

You can encode a str to a bytes object.

>>> '\uFEFF'.encode('UTF-8')
b'\xef\xbb\xbf'

And you can decode a bytes into a str.

>>> b'\xE2\x82\xAC'.decode('UTF-8')
'€'

But you can't freely mix the two types.

>>> b'\xEF\xBB\xBF' + 'Text with a UTF-8 BOM'
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
TypeError: can't concat bytes to str

The b'...' notation is somewhat confusing in that it allows the bytes 0x01-0x7F to be specified with ASCII characters instead of hex numbers.

>>> b'A' == b'\x41'
True

But I must emphasize, a character is not a byte.

>>> 'A' == b'A'
False

In Python 2.x

Pre-3.0 versions of Python lacked this kind of distinction between text and binary data. Instead, there was:

• unicode = u'...' literals = sequence of Unicode characters = 3.x str
• str = '...' literals = sequences of confounded bytes/characters
  • Usually text, encoded in some unspecified encoding.
  • But also used to represent binary data like struct.pack output.

In order to ease the 2.x-to-3.x transition, the b'...' literal syntax was backported to Python 2.6, in order to allow distinguishing binary strings (which should be bytes in 3.x) from text strings (which should be str in 3.x). The b prefix does nothing in 2.x, but tells the 2to3 script not to convert it to a Unicode string in 3.x.

So yes, b'...' literals in Python have the same purpose that they do in PHP.

Also, just out of curiosity, are there more symbols than the b and u that do other things?

The r prefix creates a raw string (e.g., r'\t' is a backslash + t instead of a tab), and triple quotes '''...''' or """...""" allow multi-line string literals.