Abstract identifiers, intertextual reference and a computational basis for recordkeeping

This paper presents three proposals concerning the structure and maintenance of formal, inter–referential, digitally stored texts: (1) include abstract atomic identifiers in texts, (2) identify these identifiers with references to text objects, and (3) keep among the texts records of computationally substantiated claims about those texts. We use "formal" in a narrow sense approximating computer–checkable; we are informed by informal symbolic practices used in mathematical text and program source text, which we hope to enhance and exploit explicitly. The basic management problem is how to alter texts rather freely without ruining the bases for claims depending upon them; this becomes an issue of accounting for various dependencies between texts.