Can Google be used to measure the popularity of something? Maybe. That is the approach of the famous TIOBE Programming Community Index. ( See http://www.tiobe.com/tpci.htm).
As the maintainer of an open source project, it is something I look at. I am happy to that ehcache now returns more results than any other Java cache with 348,000 results.
So, is is the most popular? Who knows. But downloads and google suggest it is.
(I have been very quite lately. This is my first post in 7 weeks. Blame it all on a new job. I do have some pent up blogs though…)
I have just become a new member of JSR-107, the caching API. This one has been around for a long time and has gotten bogged down. I am shortly going to start an ehcache implementation of JSR-107. I have already done a first pass over ehcache and added extra features and done some restructuring to make that job a lot easier.
The key benefit I can see in JSR-107 is that, if it becomes the standard, like JDBC, then you can write to it and have an almost zero cost of switching to a different caching provider. At present, users of Hibernate and Spring have much the same benefit, based on the cache plugin approach taken by those tools. But there are many caching users who use cache APIs directly. They will get the same benefit.
Hopefully this will also encourage cache implementations to provide a lowest common denominator of functionality.