|top|: Oblitum/interception

In modern distributed systems, two opposing forces are at play: the need to retain data for debugging, and the need to forget data for privacy/compliance. The oblitum/interception pattern (Latin oblitum = “forgotten” + English interception ) provides a structured way to intercept sensitive data before it is stored, then deliberately erase or anonymize it after a defined purpose.

# Python example using a decorator/interceptor class OblitumInterceptor: def __init__(self, rules): self.rules = rules # e.g., "credit_card": "redact", "ssn": "hash" def intercept(self, data: dict) -> dict: for field, action in self.rules.items(): if field in data: if action == "redact": data[field] = "[REDACTED]" elif action == "hash": data[field] = hashlib.sha256(data[field].encode()).hexdigest() elif action == "tokenize": data[field] = self.tokenize(data[field]) # Tag metadata for oblitum layer data["_oblitum_expiry"] = time.time() + 3600 # 1 hour TTL return data A background job or database trigger that enforces expiration. oblitum/interception