the onda

During my last job at HP, I got a chance to get up close and personal with a few big partnerships with other tech companies which had similar goals with respect to specific product roadmaps. In all of these cases, the point of the relationships was to work together on projects that would result in a bunch of software that would ultimately be better for both companies.

Now hold that thought for a moment.

Over the last few days I've been having a blast at Pycon, the main annual conference behind the Python programming language. It's fantastic to see how far it has come over the last two decades, almost entirely on the back of committed volunteers who burn nights and weekends maintaining and building a high quality software project which is so critical to many industries, and especially for the future of the open Web OS.

And yet, there are still a few warts on Python that could benefit from the focused effort of sponsored work. Of all of the big companies, Google has done the best job to date here— sponsoring projects like Unladden Swallow, a needed speed boost. There are other big companies that occasionally sponsor Python work, but not as many as you would think— especially not given how strategically important the language is to them.

There are two things that hit me this conference that might provide a solution. The first is that as serious as these warts might be, it would take relatively little manpower to remove them. Specifically, paying for something like 2 engineers for a sustained year of 100% work could go a long way towards solving speed, packaging, or concurrency issues in a really deep way. The second observation is that by the standards of big company budgets— and specifically as applied to the kinds of partnerships I mentioned above— this is an insignificant rounding error when it comes to dollars spent— even for really senior folks.

Instead of spending the money on needlessly flying executives back and forth to regurgitate the same 20 slides over the "partnership possibilities," I wonder whether there isn't the possibility or reallocating these Steaks&Strippers budgets to fund open source wart removal as a kind of joint venture where each side dedicates a couple of engineers to something that would benefit both companies and allow meaningful longterm partnerships to develop.

It would take more thinking to see exactly how this could be fit into something like removing the Python GIL (a impedance to really good concurrency and multicore support), but I'm sure it could be done. Each big company usually has its cash cow business model: Google and web advertising, Intel and chips, HP and ink, etc. And each of these cash cows often generates a whole host of "ecosystem partnerships" that companies go into to ensure the long-term continued growth of X, where X is whatever makes the cash register continue to sing.

In effect, newer projects like Android may provide the perfect test for these types of partnerships. Except of course that a mobile OS is pretty strategic to most big tech companies which is usually code for more Steaks&Strippers and less meaningful partnering.

Something to think about.