Unnaturally Agile

Posted 1 month, 2 weeks ago | Originally written on 16 Jan 2024

Inasmuch as the Internet has become a treasure trove of information, there is one form of information that abhors its radical transparency: curated content. The best it offers in its place is crowd-curated content or social knowledge such as on StackOverflow and its myriad sibling sites. Even with social knowledge there is a price to pay: it only applies in verifiable domains e.g. computing, mathematics, and fails miserably where a diversity of 'valid' opinions prevail. High quality, curated content such as books, scientific journals, periodicals, music, film and many other similar media are almost invariably hidden behind paywalls. It is encouraging to observe that prices have been steadily falling (iTunes made it possible to buy music by the title rather than by album; you can now rent movies on Google Movies for as little as £3 and for $500 O'Reilly's annual subscription one can access "more than 35K book titles plus 30K+ hours of video, proven learning paths, case studies, interactive tutorials, audio books, and videos from [their] global conferences and 250 of the world's best publishers") and some financing models have even been reversed (think of open access publishing where authors pay to publish) primarily where volume consumption can be leveraged. While YouTube is bursting with useful videos of virtually any subject, streaming services are significantly more bingeworthy but require a subscription.

What most of these types of media share in common is the demanding and painstaking process of curating content. Take film: the vast majority of the value creation of film is done in preparing the script; no film maker in his right mind will begin any part of production, even pre-production, without a locked script. This can take years with many scripts not seeing the light of day and others only going into full production years after the initial concepts were banged about. The same is true in academic circles: it takes several years before a research topic becomes a full academic paper and the writing process itself can hobble along for months before it is considered sufficiently presentable. Even in the physical engineering world, way before project kickoff, it can take teams of engineers many months verifying the feasiblity of the project before any work can be seen.

But for whatever reason the world sees software as being categorically different. So much so that the notion of agile development is casually accepted without much criticism.

Agile claims that valuable software can be developed with virtually no planning or forethought and can be 'continuously delivered' in steady streams of progress as though prior work is perfectly coupled to current work. Contrary to every other form of media, software can somehow magically appear and evolve effortlessly. I hope that one day the ludicrousness of this idea will become more obvious that people will look back and wonder, "how did we get away with this for so long?".

I think we are living in a world in which we have been so successful at factoring out the machine that the physical limits are 'safely' ignored. It's like the child of billionnaire who has become so accustomed to getting practically anything he can think of that his capacity to work is so ablated and it is almost certain that if he were to have any offspring then he cannot feasibly teach them anything useful. There is a price to pay for disregarding the machine. On the flip side, machines have become so powerful that there seems no limit to what we can lob at the machine still get amazing results. In the meantime, libraries have become bloated but the growing number of numerate computer users is oblivious to the layers of junk that is included in trivial libraries.

In a word: agile is unnatural.