Text to speech can be implemented in its simplest form using Microsoft's Component Object Model (COM) connecting to the Speech API (SAPI). The attached Python code shows you how to do this.

# to bring text to speech capability to your Windows computer
# install SAPI5Speech (XP and Vista should have it already)
# http://www.nolad.com/vt/redist/SAPI5SpeechInstaller.msi
#
# Windows XP has Sam, SAPI5Voice adds Mary and Mike
# http://www.nolad.com/vt/redist/SAPI5VoiceInstaller.msi
#
# the Python for Windows extensions should be installed
# for COM to work eg.  pywin32-210.win32-py2.5.exe
# from http://sourceforge.net/projects/pywin32/
#
# tested with Python25 on a Windows XP machine by vegaseat

import win32com.client

voices = {
'Sam' : 'Microsoft Sam',
'Mary' : 'Microsoft Mary',
'Mike' : 'Microsoft Mike'
}

# choose voice from the voices dictionary
voice = 'Sam'
# range 0(low) - 100(loud)
volume = 100
# range -10(slow) - 10(fast)
rate = -1

# some text to speak
text = """\
It is said, that if you line up all the cars in the world end to end, 
someone would be stupid enough and try to pass them.
"""

# initialize COM components of MS Speech API
# COM is Microsoft's Component Object Model
# (COM is also used by Peter Parente's pyTTS)
speak = win32com.client.Dispatch('Sapi.SpVoice')
# assign a voice
speak.Voice = speak.GetVoices('Name='+voices[voice]).Item(0)
speak.Rate = rate
speak.Volume = volume
# now speak out the text
speak.Speak(text)
The article starter has earned a lot of community kudos, and such articles offer a bounty for quality replies.