If you misspell the word, the webster server may offer a list of close matches as alternatives. If you know the prefix of the word, a list of potential matches may be obtained by entering a * after the prefix, e.g., alpha*.
This Web interface is rather simple-minded -- all but very short words are hyperlinked back to the dictionary. Because it doesn't know if the words are actually in the dictionary, these links may fail. In particular, prefix/suffix removal to find the root word is neither performed by the Web interface nor the Webster server, so the link for alphabetically doesn't work, since the webster server only has an entry for alphabetic, and it is not smart enough to find the root. The job of root extraction should be done in the Webster server, just as spell correction feedback is done now for alphebetic. Note that you can use cut-and-paste to extract the root and paste it into the search window: you just have to click-hold-drag instead of single-click.
The webster client program used internally by this interface is /afs/cs.cmu.edu/user/bsy/bin/webster, if you want to use it from outside of the WWW. The webster database server that it contacts is configurable via your ~/.webster file; see the man page for details. The webster client currently uses firstname.lastname@example.org as its webster server.
The webster client used to use one of several webster servers, depending on which responds first. Unfortunately, all other webster servers have been shut down. This hypertext Webster interface was written email@example.com