This is the brighter branch of a dystopian choose-your-own-adventure essay on systemic failures to carry through sizeable projects in "enterprise" environments.
Its alternative write-up is found at: On unfinished software migrations.
Suppose that we have a problem.
Suppose that our entire business communicates through some handcrafted marvel of software architecture built by a few prolific souls who left the company 10 years ago, in a language that today's scholars could only describe as Pig Latin.
Or suppose that our data centre runs on a fleet of systems so ancient that
their glibc can't be used to compile or run any sort of ssh software using
algorithms that modern tools would agree to work with.
Or suppose that the 700 modules, servers, libraries and what-have-yous that
comprise our solution all use the same general bundle of configuration
files, a happy, hairy, slimy chimera distributed by a Rube Goldberg
machine of cron jobs
that only Tom Bombadil can tame.
And they're cross-referencing one another. And they're stashed across several
source repositories. Hosted at different providers.
Or suppose that it'd be cool to use that shiny "new" multiline Java
String.
In summary, suppose that we have a serious problem that calls for
all-hands-on-deck, cross-team, cross-discipline, focus-of-the-semester
remediation.
The new year hits, and we're even graced by the presence of some board member's face in the fully remote kick-off meeting: the show is officially on the road!
Q1, spirits are high, advisory groups aplenty, meetings eagerly planned and subsequently attended: the well-oiled Dreadnought of progress is set in motion.
Q2, meetings were planned recurrently to begin with; resources are temporarily diverted, as is the custom; the battering ram of progress is still certainly heading somewhere, as evidenced by the risk incurred, in leaning against it for and early afternoon nap, of slumping off in its trail.
Q3, every other worker (in the idiomatic sense) has their 4-week vacation lined up and every other worker (in the literal sense) is consequently overrun with unprecedented and unexpected workload. The rudimentary vehicle of progress is still rumoured to be crawling. The stakeholders are nowhere to be seen but, what stakes, really? The "place order" button is working, after all; life is beautiful and neither effort nor indolence is rewarded nor punished.
Before the end of the year, our retroactively baseless, worn-out jalopy of progress did indeed ram itself into a wall, there's no denying that, but its battering prowess may have been greatly exaggerated. It now shares Tom Riddle's predicament in that it shall not any more be named. It withered and went away without as much as a whimper. It was everybody's, therefore it was nobody's; the corpse remains unclaimed and unidentified but, comrade, is it really dead if the records don't show that we decided to kill it?
It was the spring of hope, it was the winter of
despair.
As an effort, it was, as always, resounding, and
you may join K. in inquiring of the authoritative
parties about what
constitutes, indisputably, one of the endeavours of all times.
As far as history shall be concerned, the solution is all but established, the time has passed and the resources were spent in accordance: in our book, it's a done task; but to the keen mind, it's one more shipwreck elevated to the status of cultural landmark, a new hurdle toughening the sell regarding possibly addressing the original issue "once more".
Suppose that we had a problem... Now the problem has its own meta problem.