1) Log into an irb session, and enter your S3 login details.

require 'rubygems'
require 'aws/s3'

  AWS::S3::Base.establish_connection!(
    :access_key_id     => 'REPLACE_ME',
    :secret_access_key => 'REPLACE_ME'
  )

output:
=> #<AWS::S3::Connection:0xb75e0594 @http=#<Net::HTTP s3.amazonaws.com:80 open=false>, @secret_access_key="", @options={:server=>"s3.amazonaws.com", :access_key_id=>"", :port=>80, :secret_access_key=>"", :persistent=>true}, @access_key_id="19S45GYAGWK8DC2B8VG2">

2) Browse the existing buckets.

AWS::S3::Service.buckets

output:
=> [#<AWS::S3::Bucket:0xb75cc850 @object_cache=[], @attributes={"name"=>"ogg.twitteraudio.com", "creation_date"=>Sat Apr 26 10:40:16 UTC 2008}>, #<AWS::S3::Bucket:0xb75cc83c @object_cache=[], @attributes={"name"=>"t1000", "creation_date"=>Fri Apr 25 21:35:21 UTC 2008}>, #<AWS::S3::Bucket:0xb75cc814 @object_cache=[], @attributes={"name"=>"t2000", "creation_date"=>Fri Apr 25 21:53:15 UTC 2008}>]

3) Browse the buckets in a programmatical way.

AWS::S3::Service.buckets.each {|b| puts b.name}

output:
ogg.twitteraudio.com
t1000
t2000


4) Add a new bucket called t3000.

AWS::S3::Bucket.create('t3000')

output:
=> true

5) Observe adding the bucket again doesn't cause an error.

AWS::S3::Bucket.create('t3000')

output:
=> true

6) View the buckets again.

AWS::S3::Service.buckets

output:
=> [#<AWS::S3::Bucket:0xb75cc850 @object_cache=[], @attributes={"name"=>"ogg.twitteraudio.com", "creation_date"=>Sat Apr 26 10:40:16 UTC 2008}>, #<AWS::S3::Bucket:0xb75cc83c @object_cache=[], @attributes={"name"=>"t1000", "creation_date"=>Fri Apr 25 21:35:21 UTC 2008}>, #<AWS::S3::Bucket:0xb75cc814 @object_cache=[], @attributes={"name"=>"t2000", "creation_date"=>Fri Apr 25 21:53:15 UTC 2008}>]

Note: You would expect t3000 to be in there however it didn't appear possibly because of the bucket permissions.

7) Let's then look for bucket t3000.

t3000 = AWS::S3::Bucket.find('t3000')

output:
=> #<AWS::S3::Bucket:0xb76df724 @object_cache=[], @attributes={"prefix"=>nil, "name"=>"t3000", "marker"=>nil, "max_keys"=>1000, "is_truncated"=>false, "xmlns"=>"http://s3.amazonaws.com/doc/2006-03-01/"}>

8) Now that we've found the bucket let's upload a text file called works.txt.

file = "works.txt"

output:
=> "works.txt"

AWS::S3::S3Object.store(file, open(file), 't3000', :access => :public_read)

output:
=> #<AWS::S3::S3Object::Response:0x-608926458 200 OK>

9) Setting the file access to :public_read allows us to view the file from the http location http://t3000.s3.amazonaws.com/works.txt

References:
http://amazon.rubyforge.org/
upload_to_s3 - Ruby S3 upload client [dzone.com]

*update: 14:30 30 April 2008 *
I didn't use Bucket.objects(:reload) which is the reason why the bucket t3000 didn't show up with the statement Service.buckets

Reference: spatten design - Amazon S3, Ruby and Rails slides [spattendesign.com]

This ruby code builds upon the code S3 upload client for Ruby [dzone.com] by reading the account details and file details from XML files.

#!/usr/bin/env ruby
#file : rubys3file2upload.rb

require 'rubygems'
require 'aws/s3'
require 'rexml/document'
include REXML

class S3FileUpload

  def initialize(xml_accounts_file, xml_upload_file)
    initialize_account(xml_accounts_file)
    main(xml_upload_file)
  end

  def initialize_account(accounts_file)

    file = File.new(accounts_file)
    doc = Document.new(file)

    account = doc.root.elements['records/access']
    h = Hash.new
    h[:access_key_id] = account.elements['key_id'].text.to_s
    h[:secret_access_key] = account.elements['secret'].text.to_s
    AWS::S3::Base.establish_connection!(h)

  end

  def main(xml_upload_file)
    file = File.new(xml_upload_file)
    doc = Document.new(file)
puts doc
    doc.root.elements.each('records/file') do |f|
      local_file = f.elements['local'].text.to_s
      bucket = f.elements['bucket'].text.to_s
      mime_type = f.elements['mime_type'].text.to_s
      upload(local_file, mime_type, bucket)
    end
  end

  def upload(local_file, mime_type, bucket)

    base_name = File.basename(local_file)

    puts "Uploading #{local_file} as '#{base_name}' to '#{bucket}'"

    AWS::S3::S3Object.store(
      base_name,
      File.open(local_file),
      bucket,
      :content_type => mime_type,
      :access => :public_read
    )

    puts "Uploaded!"
  end

end

if __FILE__ == $0
  s3fu = S3FileUpload.new('s3accounts.xml', 's3files2upload.xml')
end

file: s3accounts.xml

<s3accounts>
  <summary/>
  <records>
    <access id="100"><name>REPLACE_ME</name><key_id>REPLACE_ME</key_id><secret>REPLACE_ME</secret></access>
  </records>
</s3accounts>

file: s3files2upload.xml

<s3files2upload>
  <summary/>
  <records>
    <file>
      <local>autocomplete.html</local>
      <bucket>t2000</bucket>
      <mime_type>text/html</mime_type>
    </file>
    <file>
      <local>autocomplete2.html</local>
      <bucket>t2000</bucket>
      <mime_type>text/html</mime_type>
    </file>
    <file>
      <local>autocomplete3.html</local>
      <bucket>t2000</bucket>
      <mime_type>text/html</mime_type>
    </file>
  </records>
</s3files2upload>

Note: The file given the correct permission could be read from http://t2000.s3.amazonaws.com/autocomplete.html

Reference: http://amazon.rubyforge.org/