True Fallacies of Distributed Computing
There’s a wikipedia article about the fallacies of distributed computing.
Having pretty tight control of a computing space allows people to maintain the illusion that data “exists.” But if we tell the truth, forces outside the control of your program disrupt data on a regular basis (buffer overruns, memory errors, seg faults, drive failures, power surges, solar flares?, etc.).
When we expand computing outside the bounds of a single computer these disruptions become so frequent that it we perform extremely awkward gyrations to preserve the illusion of data integrity (checksums, cryptographic hashes, signatures, consensus, leader election strategies for resolving inevitable differences, etc.)
The problem that data
The world is not stateful, it is interactive process. We resolve the quantum superpositions of probabilities only by interacting with something, and it is the AGENCY involved in that process of interaction that determines/reports/creates information or data. To divorce that data from the agency which produced it and act like it “exists” as some kind of truth is to create a fiction which we confuse with reality.
All data comes into existence through a process of authorship by some agent.
- The sales figures are reported by someone (they may have reasons to inflate them).
- The temperature is measured by some device (it may not be well placed or calibrated) -