The developer documentation conundrumData
Documentation is an essential requirement of any enterprise solution. It describes for code-based solutions: specifications for systems, procedures or workflow, algorithm contexts, use case scenarios or other related metadata.
Even given its importance for business continuity, documentation has always had a bad rap in the developer community. Its mere mention instantly causes an ominous dark cloud to descend on those tasked with the endeavour.
This reticence, drag on productivity and increased workload is known amongst managers when quoting for Indeed, this aversion-productivity-price triptych was probable cause why “Working software over comprehensive documentation” is one of the original values in the Agile Manifesto.
Fortunately, some companies are tackling this conundrum and have started offering products (some with nascent AI capabilities) to automate this process. In general, these products generate documentation for technical and increasingly business contexts for enterprise systems and applications. They serve to:
- mitigate manual input
- provide a wide range of output formats
- facilitate on-demand documentation re-generation so content remains current.
Perhaps, with the exponential growth of AI and compute capacity that these software AI offerings will finally create a rear-guard action that will finally crack the documentation enigma.
Indeed, a much-awaited Turing test will be when software is able to generate “comprehensive documentation” without human intervention. When that time arrives software will really be “working”, and the second value of the Agile Manifesto will become irrelevant. For all developers today still dug in in the documentation trenches – that time can’t arrive soon enough.