The purpose of this standard is to help make the task of implementing an algorithm in software a quick one - all the way from finding some source code to full integration and testing.
Many algorithms exist out there. A fair proportion exist on the internet but trying to find them is often a very frustrating process. Say you want the source code for the Luhn algorithm written in C that works on Visual Studio 6. An internet search for "Luhn VC6 C source code" is probably your best bet. Unfortunately I think it would probably take a while to find something half workable. "VC6" and "C" stand for many things! In my opinion, publishers just need to go the proverbial "extra mile."
A part of the extra mile this standard encourages publishers to go is to use tags for languages and IDEs (see below for a comprehensive list of tag names) so that the search is narrowed to pages that a) adopt the standard and b) are for the desired language/IDE. Your internet search now becomes "Luhn langu-c ide-vc-6."
A further part of the extra mile (and arguably the most important part in fact) is to provide some test cases. Someone may not even by needing to download your code because they already have their own implementation. It is comforting, however, to know that they are arriving at the same answers. The example program available to download should use a class for the algorithm section itself, using a wrapper that feeds the test cases (read from an XML file) through it. Some people may just be seeking the answer to a particular test case. They can then edit this XML file and get the answer they need.
Below is a reference for the standard and here's an example page made for my implementation of the Luhn algorithm.