If we consider email addresses as links, and put them in our graph structures, there will be no outdegrees for this kind of links (a waste of space for adjacency matrix, if it is implemented). Also, they will not be indexed since we can't parse this type of urls. Which would mean, there is really not much gain counting them as links in part I and II since we can't retrieve them.
If we are really persistent, maybe we can associate the text between <a > and </a> as the keywords for this email, and enter these words into the index structure.
e.g.
<a href="mailto:..."> someone's email </a>
someone's email will now be the keywords for the email link, rather than the present page. But this is still not helpful, if the text is not the person's name, but, let's say, "contact me".
So, what is the official oppinion on indexing email addresses? |