I think there are probably a class of REST webservices that could not happen.
Let me explain…
I just realized the other day that Google search is still using 1999 era table-based layout HTML. I asked myself, the obvious question: WTF?
Then I had a moment when things suddenly became clear to me– Google is a prime target for scraping and therefore the increased bandwidth and ugliness they get from using table-based layouts is probably offset by the scraping they prevent through obfuscation. If their markup were cleaner and more meaningful, they would be much easier to scrape.*
Then another revelation came.
In the process of developing microformats, what we’re really doing is making the Web more scrape-able. By establishing some guidelines we make it so that if we want to scrape a certain kind of data, we should be able to get it from all sites that adhere to the microformat with minimal difficulty.
This led to another revelation.
I think there are probably a class of REST webservices that could disappear (or not appear, but I’ll get to that in a second).
I think webservices which are query-only could probably be accomplished by a well planned url-scheme and clean, meaningful markup, which in my mind includes judicious use of microformats.
Of course, with each page request there would be some wasted bandwidth, so high traffic webservices like Google, Amazon, EBay or Alexa probably wouldn’t work in this manner, but a smaller website could probably get away with saying “just scrape us” and avoid having to expose their data both in XHTML and some other homegrown XML format.
Additionally, with clean, valid, meaningful XHTML, what we’re talking about is no longer scraping, but parsing or consuming. I think this is a “difference in degree that results in a difference in kind.”
* After discussing this with Matt and Niall, I realized that Google probably isn’t maintaining the crusty layout to prevent scraping. Rumor has it that Google’s layout system is ‘inflexible’ and Niall pointed out that there are some cryptic looking html comments in the Google source that facilitate the scraping. Still, I think my conclusion holds and even if I’m wrong about Google, I put this out there as my thought process.