Wait is a verdict, not an absence of one

Of the three verdicts this journal hands out, "wait" gets the least respect. "Worth adopting" starts projects. "Marketing" gets quoted in Slack. "Wait" reads to some readers as the journal declining to do its job, and we would like to state, for the record, that it is the job.

A platform's first six months in general availability are its least representative. Launch pricing has not met its first renewal cycle. The operational sharp edges have not found enough production victims to be documented. The migration tooling is written for the demo schema, which is never your schema. Reviewing a launch and reviewing a product are different assignments; the launch is simply what is available on launch week.

The Lindy effect is the folk version of this argument: what has survived tends to keep surviving, and what is new has not yet shown you which failure it carries. We would put it less mystically. Every platform we have reviewed carried at least one defect that only surfaced under a real workload at real scale, and the defects surface on the early adopters' schedule, not yours. Waiting outsources your discovery costs to teams in more of a hurry.

The objection arrives on cue: wait too long and you pay the laggard tax. True. Which is why "wait" in these pages always carries a date or a condition. Wait for the first renewal pricing. Wait for the second major version. Wait until the migration guide mentions a schema that looks like yours. A wait with a trigger is a plan; a wait without one is a nap.

No column this week, no chart, one sentence of standing policy instead: when we say wait, we mean the product is interesting and the evidence is young. That is not indecision. That is the review.