No. 0501 min read
What mediation taught me about people (not that kind of mediation)
Fourteen years reconciling billing records at Sprint taught me to check what I know before assigning blame.
No. 0341 min read
Your parser rejected the record. Fine. Somebody still has to find out what happened to the usage, folks. An error directory is not a finished accounting job.
Your parser[1]Ref. 1parser en.wikipedia.org rejected a record, the pipeline kept moving, and you would like to call that handled.
The software did its job, but the usage is still sitting there waiting for somebody to figure out what the hell happens next. An error queue[2]Ref. 2error queue en.wikipedia.org is a reasonable place to put something that cannot proceed, as long as somebody remembers to go back to it.
I'd keep the original rejection time separate from the latest attempt time. Reset its age on every retry and the oldest problem becomes permanently new. Accounting is still fucked, but look at those fresh dates.
When reviewing the queue, I want a decision about the oldest unresolved item. It may have to stay open, with an owner and a next action. I'm glad that newer record retried successfully, but I'm still looking at the old bastard.
No. 0501 min read
Fourteen years reconciling billing records at Sprint taught me to check what I know before assigning blame.
No. 0361 min read
Corrupt how, folks?
No. 0331 min read
A bigger sequence number does not mean your missing file arrived, folks.
No. 0151 min read
Sprint's D3884 finished early and under budget with no customer service impact.