Skip to main content

When a bunch of grains of sand make a heap?

Posted by fabriziogiudici on April 11, 2010 at 2:27 AM PDT

The spring is here at last, and yesterday I was walking on a beach. Often walking on sand beaches recall me the "paradox of the heap": you have a heap of sand, let's say made of a million of sand grains. Then you remove one, and get 999,999 sand grains. No doubt, it's still a heap. Now you repeat the process, and eventually you'll get with a single grain of sand in your hand. No doubt, a single grain of sand is no more a heap. This means that, somewhere in the process, what was initially a heap was no more a heap. This paradox, attributed to Eubulides of Miletus, an ancient greek philosopher, demonstrates that some concepts of the real life can't be precisely defined. There are many proposed solutions, such as fuzzy logics or defining things by consensus (e.g. setting an arbitrary amount of sand grains to be the conventional threshold for a heap) - this means that the problem is left with a big degree of uncertainty and disagreement.

When I got back home, yesterday night, I learned of James Gosling leaving Oracle and figured out the large echo that the news would have in the community. Also considering that James is not the first, but the n-th prominent engineer that didn't accept the new Oracle/Sun panorama, I'd say that we are facing with a paradox of the heap (with a slight variation: in case of people such as Tim Bray or Gosling or others we should give grains a weight). We know that the new Oracle won't be equal to the previous Sun (otherwise, it would fail in a few years and will be sold to IBM) and we have to accept it. Most of us expected that it will be a mix of the two original corporate cultures. A few prominent engineers flying away won't change things so much; but when the Sun heap won't be a heap any longer?

I think that there are mixed messages in the community now. The official message that Oracle gave us just after the acquisition was a very reassuring one, about the future of many products that the community is interested on. Further pieces of news and small announcements went in the same direction (for instance, I was pleased to learn that Glassfish will incorporate Coherence support, which is a piece of the upper-level suites by Oracle, that one might figure out are confined into the WebLogic family).

So, from a rational point of view I still have reasons to be fine. But, the paradox of the heap doesn't leave you with a satisfactory rational point of view. And I must take into account that Oracle is not communicating with the community since a few months now, and at this point this couldn't be but a deliberate communication policy and the impact on the relationship with the community could be huge. Sure, we'll have conferences and e.g. Jazoon comes in a couple of months, but it's really too long, so I say that we (as JUGs and whatever) should take some initiative and (re)start asking questions to Oracle.


Related Topics >>


Oracle is not Sun, don't

Oracle is not Sun, don't expect that type of engagement and relationship between Java community and Oracle. They talk to big companies not to communities. IMHO.