Especially when adding “historic” publications, the authors often are no longer working or even alive. Thus they do not have a homepage. To circumvent the mandatory field for the homepage URL, it is necessary to enter bogus like “http://not-available.com”. I am sure that you wouldn’t want such clutter in your database. Therfore, the homepage should be optional.
It might also be worthwhile to think about binding the author homepage to the affiliation of the author. Sometimes, people are affiliated to different institutions (with different homepages for them) and selectively publish their papers with one of their affiliations respectively.

Submitted on 5 October 2010 at 18:24

This is a valid concern. For Alan Turing, I added a link to his wikipedia page :) But not all authors have one, I suppose. How do you propose to implement identification, i.e. distinction of different authors with the same name? I suppose adding a list of affiliations should be sufficient.

Binding homepage to affiliation is also a good idea. However, homepages don’t necessarily have a very long lifetime.

In summary: homepages are not useful for identifying historic authors, and they are not a durable means of identification. But they do serve to identify currently active authors.