bcantrill

No. What Oracle does is Wrong (capital "W") in that they have no regard for the importance of a social contract; no company ever outgrows its responsibilities to society.

Someone asked:

What's the most beautiful piece of code you've seen over the years. (Code you wrote your self excluded)

Unfortunately they forgot a question mark, so the comment was nuked, but here's my reply:

Oh, exciting question -- and one that I reserve the right to answer more than once! ;) One piece of code that I absolutely love is this little bit of auto-scaling in the kernel memory allocator. It's very tight and elegant -- a really clever technique that I haven't seen used nearly enough. This is written by Jeff Bonwick, who later went on to (along with Matt Ahrens) invent ZFS. Jeff was my mentor at Sun, and a personal inspiration -- and I think that that little snippet of code embodies Bonwick.

I think I would rather fight one horse-sized Sun Microsystems than one duck-sized Oracle -- let alone a hundred. Never underestimate the depravity of a sociopathic duck!

Bonwick had an enormous influence on me in that he got me to change my expectations of myself: prior to working at Sun, I viewed defects as inevitable and that software was like much else that we engineered (namely, fated to be plagued by unknowable and transient errors). But in working with Jeff -- who came to software from math, not from engineering -- I came to appreciate that the better analogue for software is a mathematical proof, and that the beauty in software is that we fashion a proof that functions as a machine. As a result, we (and and we in software alone) can create machines that are perfect. This was tremendously inspiring to me, and has had a profound effect on my career and in the pleasure I take in the work.

In terms of accomplishing the same at Joyent, I certainly aspire to inspire software engineers in the same way. I think sometimes I succeed and sometimes I fail -- but I think the successes have outnumbered the failures, and I continue to be inspired by the team that we have built here.

One of the concerns that I have is that stacks are becoming ever-more complicated (accelerated by open source, which has allowed us to live the dream -- and the nightmare -- of componentization) but that our education isn't keeping pace. So permit me to be the anti-Peter Thiel: budding software engineers need to complete their formal education, first and foremost. I know that this isn't necessarily a popular opinion, but formal computer science education is essential, and I have seen (and continue to see) a marked difference between those who have completed it and those who haven't. An essential part of that formal education: learning what you don't know. If you don't come out of a formal education humbled by it all (and astonished that a software system can so much as boot, let alone run, let alone run well), then you missed the point.