"Sherlock provides similar advantages simply because it provides a more efficient interface to get information via HTTP... To [book a flight on an airline site] you also have to contend with banner ads, cheap travel offers, a link your to your frequent flier account, etc."
That's true, but those ads and such are part of the business model of the service provider. Many of the Watson and Sherlock 3 tools use licensed datafeeds, and presumably the service will find a business justification in such ad-less use, but some of the other Sherlock plugins are more like skins over existing sites, and scrape markup to get data to re-display.
This is the question I haven't seen discussed much: what are the legalities and ethics of using someone else's data feed in your own page or application?
For instance, if a blogger has a sponsor yet still supports an RSS feed, then I'd assume that the blogger and sponsor both find the additional ad-less distribution useful. But suppose an aggregator displays a competitor's ad on their feed of this same content?
The separation of data and presentation is useful, and I have a feeling we-all can reconcile the business models involved, but I think we have a lot of work to do still in deciding how digital content can be re-used...?