It is generally presumed that an RSS or Atom feed is the counterpart of a blog, news site, etc., or is aggregated from other feeds that are. The current Atom format specification even goes so far as to require each entry to have a link pointing to an "alternate representation" of the entry. Digests certainly began as subordinate entities to web pages, but will they always remain so? I doubt it.

Sure it makes sense to have a web page with the same data as a feed. But are there reasons for having just a feed with no web page? If there are, it will be done. It may even be done if there aren't--not everybody is a genius. Let's start by examining some reasons for having both, and see if we can find ways around them.

You can advertise on a website and use a feed to drive people there.

True, and a good model. However, advertisements within digests are becoming more common all the time. If response rates within digests end up becoming higher than response rates within web pages, we all know where the (smart) advertising money will go and where it won't. Whether that will turn out to be the case is an open question. But the fact remains that you don't need a website to get advertising revenue.

You can control the appearance of a webpage more than you can a digest.

True, and good if you want to spend time on web design. Skipping the web page can save you time though. And digests don't have to be completely unstyled. Also, some people prefer reading unstyled content. People with poor design skills may find that using an unstyled digest saves them from their own mistakes by preventing them from making their content difficult to read.

A lot more people have web browsers than feed readers.

Yep. Even though sites like Yahoo! are presenting digests to their customers in their web browsers, this will be a good argument for a while.

It's easier to track traffic on a web page than a digest.

True. I occasionally try to come up with a good solution to that issue, and along with a lot of other people, will continue to do so. There are mediocre solutions like so called "web bugs", but nothing really fool proof yet.

Why not do both?

A good question with a few mediocre answers: as mentioned above, to save on web design time. To save on bandwidth requirements (which may or may not work--as infrastructure improves, this will probably be true). 'Cause you don't feel like doing both.

We'll it's difficult to come up with a good argument against doing both, but I'm not going to call someone who only wants to publish a digest a fool. And I imagine digest-only publishing will only become more common as time goes by.