Input

Web forms

Since most modern web browsers support UTF-8, you can get your form input in UTF-8 format. If you're using HTML, set your HTML editor to save files in UTF-8 format and add this tag:
<meta http-equiv="Content-Type" content="text/html; charset=UTF-8" />
If you're generating your form with a scripting language add this header:
"Content-type: text/html; charset=utf-8"
In PHP you would write:
header("Content-type: text/html; charset=utf-8");
and add the meta tag in your generated HTML. Adding the charset information will return the form data in UTF-8 encoding. The browser will not return any information about the encoding of your form data however. If you send out your page in UTF-8, the form data will be posted back in UTF-8 (most of the time).

Now what if the user changes the page encoding or the browser doesn't support Unicode? (We actually have a client that still uses Internet Explorer 4 on and old Macintosh and the data he submits is in MacRoman encoding.) So we need to detect which charset we're dealing with.
Add a hidden field to you form with characters that have a different value in the character sets we want to check:
<input type="hidden" name="charset_check" value="&auml;&reg;"/>
When we get this value back, the browser has translated the HTML entities to characters and we can check their binary value. Here is a PHP example:

$test = $_POST['charset_check']; // our hidden field
if (bin2hex($test) == 'c3a4c2ae';) { // UTF-8
   $charset = 'utf8';
} elseif (bin2hex($test) == 'e4ae') { // ISO-8859-1 (=Latin-1) or Windows-1252
   $charset = 'ISO-8859-1';
} elseif (bin2hex($test) == '8aa8') { // MacRoman
   $charset = 'MacRoman';
} else {
   $charset = 'Unknown';
}

If you don't want to go through all this trouble, you can use a regular expression to check if the input is in UTF-8 or not. If it's not, warn the user. This regular expression can be found on the W3C website.

ASP (Microsoft's active server pages) is a different story. I am not an asp developer myself, but my friend Stijn Matthyssen of the company Weblogik is one. We spend a couple of evenings trying to apply the Unicode Workflow in asp. We gave up on the old asp soon, because it just didn't work.
Asp.net handles UTF-8 transparently. Input and output are UTF-8 by default. And however MS Access and MS SQL Server's native encoding is UTF-16, it is all taken care of by the .net framework. But because all posted form data is put into String objects by the .net framework before it reaches your code, you can't detect what the original encoding is. (I guess you could parse the raw query string, but if you're gonna do that, what's the sense of using a framework.)
If you're posting a form in an encoding that wasn't expected by the .net framework, all the String objects will be empty.

Flash

Macromedia Flash MX (Flash 6) and above uses UTF-8 as its native text format. Unless someone has put system.useCodepage = true; somewhere in the code. So data coming from a Flash form doesn't need any conversion.

CSV, Tab separated files

Most spreadsheets and databases have a charset export option. If not, open your file in a decent text editor such as UltraEdit or BBedit and convert it to UTF-8.

If users can upload CSV files, use W3C's regular expression to check for valid UTF-8 data. If it's not valid, assume your local charset, which is ISO-8859-1 (Latin-1) or Windows-1252 in English speaking countries and Western Europe.

Microsoft Excel up to version 7 (Excel 95) stores strings as 8bit values. Excel 97, 2000, XP stores strings in UTF-16 format. You can parse an excel sheet with the Spreadsheet::ParseExcel Perl Module and convert it to UTF-8 if needed. I don't know of any decent PHP classes that read Excel files. If you do, please let me know.