Description

Utf8ToGsm.to_gsm provides functionality to convert UTF-8 characters (in a string) to their GSM equivalents for sending SMS messages via SMPP.

Examples

require 'utf8_to_gsm'
Utf8ToGsm.to_gsm('Convert to GSM: !@#$%^&*()')
=> "Convert to GSM: !\x00#\x02%\e\x14&*()" 

Usage

Provide Utf8ToGsm.to_gsm a UTF-8 string that you would like to convert into a GSM-compatible string.

Utf8ToGsm will go through each character in the string:

• If the character has an exact GSM equivalent, it will be used.
• Otherwise, the UTF-8 character is transliterated to ASCII.
• If no suitable character(s) is available in ASCII, a replacement symbol (question mark: ?) will be used.
• Once transliterated to ASCII, the character(s) will be converted to its GSM equivalent. (All ASCII characters are represented in GSM.)

Implementation

Any given UTF-8 character(s) that does not exist in the GSM alphabet is transliterated with the help of unidecoder to ASCII.

unidecoder is used so that Utf8ToGsm can work with Ruby 1.8.7. Much of the functionality of unidecoder is provided by Ruby 1.9.2. However... the need at the time of writing was Ruby 1.8.7.

Motivation

• Utf8ToGsm may be useful for people who need to send SMS messages via SMPP directly to an SMSC using the GSM-7 encoding ("Default SMSC Alphabet"), data_coding = 0x00.
• Transliteration used by this library is meant to provide the best possible ASCII replacement that is available for the given UTF-8 characters. It may be helpful to review the readme from unicoder.
• Clearly, transliteration is not ideal. However, the GSM-7 default alphabet ("Default SMSC Alphabet") only allows a total of 127 characters, and so a very limited character repertoire is available.
• It is presumed that providing the closest possible replacement is better than providing nothing at all.
• For example, if a user tries to send an SMS message via SMPP containing the character "À", there is a problem. "À" does not exist in the GSM-7 default alphabet. Sending "A" as a replacement instead of "?" is probably more helpful to the recipient.
• For a truly accurate representation, UTF-16 or UCS-2 should generally be used for transmitting the payload of an SMPP PDU to the SMSC when non-GSM characters are being communicated. However, not all telcos or SMSC's support UTF-16/UCS-16.
• Theoretically, GSM locking shift tables and GSM single shift tables should be usable to represent characters outside of the GSM-7 default alphabet. However, it seems that telco support for this (especially via SMPP) is very limited.

Authors

• Justin Kulesza (kulesza@atomicobject.com)

© 2011 Atomic Object

More Atomic Object open source projects