Convert a plain Unicode text to an SMS compatible

If I send the SMS without any encoding, as the user types it in any language (áéíóú ç ñ), as most of the characters aren't in the GSM-7 encoding, it will look like "aéiou c ñ" in the recipients mobile phone. I've been finding out how has the text to be like and found it has to be encoded into UTF-16 encoding (Unicode) looking like "00E100E900ED00F300FA002000E7002000F1" (áéíóiú´ç ñ). Now that I know in what encoding I must encode that plain text the user types, what I need is a script to do that.

I'd prefer the script in PHP.

An example of all above:

If I write "Mañana iré a la tintorería" in Spanish and send it as it is, without any encoding, I'll receive in my selfphone only the characters and symbols which are in the GSM-7 encoding (Mañana ire a la tintoreria), but that's not the proper way. Now I write the same and click on Send, then my script converts the plain text into a SMS valid UTF-16 (00E100E900ED...) so I'll receive a SMS with the same text I previously wrote (Mañana iré a la tintorería).

In a shorter and more directly way:

I need a script to convert the plain text with any characters and symbols that a user types into a SMS valid UFT-16 encoding which looks like "00E100E900ED...". Prefered written in PHP.

Thanks!

the browser send that encoding which was specified by the server in the HTTP response haeder and/or the pages met information in the head section and/or the form's charset attribute
If your application gets the wrong encoding, then the page used by the browser most likely did not enforce the correct encoding (that one expected by your php script).

> I need a script to convert
for now I assume that this will not be necessary, I'll wait your next comment first ...

Hello, ahoffmann,

Thank you for you help but that's not what I need. I've been testing all what you say and more things, but it never worked fine. It only works when I use the Clickatell's converter to convert the text I type to a SMS valid Unicode that, after being finding out, I found it's UTF-16 encoding. It's a PHP page.

What I need is a PHP script that converts that text into a UTF-16 encoding, like "00E100E900ED...". The script takes the original text in a variable and converts it into a Unicode encoding (UTF-16 not UTF-8) into another variable.

Once the text has been converted I write it in the $text variable in the script I posted first and arrives perfect, but if instead of the encoded one I write the original text, when it arrives, it will only show the GSM-7 characters. Therefore I need a script to encode it, like Clickatell's one (see the picture attached above).

Thanks!

I still assume that you miss something, anyway ...

> I need is a PHP script that converts that text into a UTF-16 encoding, like "00E100E900ED...".

Is that text in ASCII or do you get it as binary data stream?

It's in Unicode, but not in Unicode encoding. I don't understand this very well; I mean, I don't know in what a text is, so I'm gonna explain to you as I understand. Just the plain text in any language, like Spanish, Chinese, Russian, Greek... any language which has characters and symbols that aren't inside GSM-7.

For example:

Convert Spanish signs like "áéíóú" into UTF-16 or UCS-2 (the same) that must look like  "00E100E900ED...", so that special characters will be shown in the recipient's mobile.

I know what you mean, but you need to understand what a "language" is for humans and what it is for the computer (in particular the programs like a browser).
The same applies to the "encoding".
With "language" and "encoding" I mean those terms you used in your description.

I am aware about different characters (not languages or encoding, but literal characters as used by humans like á â à å ä). They all have their own reüresentation in Unicode and UTF-8 (which is not your question).

But you wrote about special characters "00E100E900ED..."
wher emy question is: do you mean 00E1 (and following) as binary data (Unicopde for example), or do you mean these characters as literals. As I understand your description you're talking about the literal representation of these characters.

So please make yourself used to these differences and try to explain what exactly you have and what you want (the later is Unicode in binary representation, I guess).

OK, ahoffmann,

As I have probably been wrong all this time, or partially, I'm going to start over again without technical terms (if possible): (anyway, I know it's possible because Clickatell's converter does it, but when I asked for the script or even for help they said they couldn't give it)

I want to offer an SMS service in my Social Network which will be in many languages and internationally, so any people from anywhere will be able to use it. Therefore, as the GSM-7 doesn't have Chinese, some Spanish, or Arabic characters, for example, that text will have to be converted for when the recipient receives it to show those characters which aren't inside GSM-7. Don't ask me how or what codification or encoding, because I don't know, that's what I'm asking, how to do all that. The only thing I know is the final result MUST LOOK LIKE "00E100E900ED...", as the Clickatell's converter converts it; and after being finding out, I realized that secuence of numbers and letters is UCS-2 or UTF-16 (the same).

An example:

I'm a Chinese (or Spanish, that I'm really Spanish)  user who wants to send an SMS to his friend in Chinese via my service, but GSM-7 will show that characters as spaces or something that won't be the characters because they aren't in GSM-7, as you know. Now my script takes part, and, I don't know how, HAS TO MAKE THAT CHINESE TEXT LOOK LIKE "00E100E900ED...".  Leave apart all what I said in the first paragraphs, just take in mind the bold line above.

I hope you understand this time, and will give me the solution I need.

Thanks!

¿do you mean you want to convert something like ñ or ë to chinese characters?
For that you need a chinese character set which has these characters, I doubt that that exist (but I don't know for sure).

> final result MUST LOOK LIKE "00E100E900ED..."
hmm, it looks like that, what is wrong?

May be this is what you want: http://us2.php.net/manual/en/function.unicode-decode.php

No, I didn't mean that. I meant convert Chinese characters or any language which has characters that aren't inside GSM-7 to something that has to look like "00E100E900ED...".

Hi loggindevelop,

I had exactly the same problem but I'm afraid it's not completely solved for me yet.  I used the scripts here:

http://hsivonen.iki.fi/php-utf8/

There's a function in the package that takes a UTF-8 encoded string and turns it into an array of Unicode decimal values.  For Clickatell you need to do a little bit of PHP to turn it into the right format for the URL string.  You have to convert the returned decimal values to hex, then pad the string to 4 characters, and build your single long string.  I did all this as follows:

foreach($uni_array as $this_val) {
        $text.=str_pad(dechex($this_val),4,"0",STR_PAD_LEFT);
}

Where $uni_array is the array returned from the utf8ToUnicode function.

This worked really well for some messages but annoyingly fails for certain word/character combinations in Russian.  I've not yet been able to solve where the problem lies but it may not happen to you so hopefully this is useful.

Jon

Hi, shellgrip,

Really thank you for your so friendly answer! I've been trying but I couldn't make it work because I don't understand what that script exactly does, so I wonder if you could give or post your already made and working script because I need it and couldn't make it work.

Thank you!