Sunday, September 16, 2012

Scrum-of-scrums for the King-of-Kings


I recently found myself reading the December 2011 National Geographic. In particular, I was drawn to an article “The Bible of King James”. My lovely wife had recently sent me an interesting historical book from her trip to England about the Winchester Bible. The NatGeo article was interesting from both a religious and an historic perspective, but there was another thing that caught my eye as a sometime software professional. 

They went about their work in a precise and orderly way. Each member of the six subcommittees, on his own, translated an entire section of the Bible. He then brought that translation to a meeting of his subcommittee, where the different versions produced by each translator were compared and one was settled on. That version was then submitted to a general revising committee for the whole Bible, which met in Stationers' Hall in London. Here the revising scholars had the suggested versions read aloud—no text visible—while holding on their laps copies of previous translations in English and other languages. The ear and the mind were the only editorial tools. They wanted the Bible to sound right. If it didn't at first hearing, a spirited editorial discussion—extraordinarily, mostly in Latin and partly in Greek—followed. A revising committee presented a final version to two bishops, then to the Archbishop of Canterbury, and then, notionally at least, to the King.

For some reason, this approach struck me as very “agile” with each subcommittee being a Scrum and the successive rollups turning into a scrum of scrums. Now, it is true that we don’t have the luxury of having a group of developers each building their own version of a feature and then comparing them in the Scrum and taking the best of each – but can you imagine how optimal that would be if we did have that approach?

I recalled a story about Remedy working on a variety of play approaches for an Alan Wake sequel. They went so far as to code the most interesting to a playable level of quality and then put them on consoles around their offices. Then they watched which ideas were most popular for their staff to walk up and play. One of the ideas emerged as such a clear crowd favorite that they cleaned it up and finished it as “Alan Wake: American Nightmare”. I couldn’t find the exact link that described this background, but this interview alludes to it.

I just can’t shake the notion of “competitive collaboration” and the potential advantages that duplicating the King James process  could have for creating great software. Ahhh…but the cost…the cost. No idea what to do about that. :-)