Linear-Column Representation of Scientific Tables: A Method for Cross-Format Semantic Stability of Structured Information

This article systematizes and formalizes Linear-Column Representation of scientific tables — an author-developed method that has been repeatedly applied in practice for representing predominantly text-rich, multi-parameter scientific information. The initial problem concerns the partial dependence of the semantics of a classical two-dimensional table on the spatial arrangement of rows, columns, headers, and cells. During transfer among text, Word, PDF, mobile display, and subsequent text extraction, individual values may remain intact while their relationship to a specific object or field becomes less explicit. Comparison with existing approaches shows that table linearization, serialization, field-value representation, vertical restructuring of rows, and responsive table layouts have substantial prior history. Therefore, the independent scientific value of the proposed method is not attributed to any of these operations in isolation. The method is based on transforming each logical row of an original table into a semantically autonomous information module with a stable structure of named fields. The relation Object–Field–Value is proposed as the minimum unit for controlling semantic preservation. The article introduces the concepts of semantic address, cross-format semantic stability, logical reversibility of representation, controlled structural redundancy, comparative representational efficiency, and practical cross-format validation. A system of indicators is proposed for measuring preservation, loss, misassignment, reconstructability, and redundancy of structured information. A Candidate Theory of Cross-Format Semantic Stability is formulated. It proposes that locally explicit representation of essential semantic relations should increase the probability of preserving those relations under transformations that modify or disrupt the global geometry of the original representation. The method has already been repeatedly used in the actual preparation cycle of scientific publications and is therefore treated as practically validated in the author’s workflow. However, the hypothesis that it has higher comparative efficiency than conventional tables and the closest established forms of linearization requires dedicated controlled testing. Keywords Linear-Column Representation; scientific tables; structured information; cross-format semantic stability; semantic autonomy; semantic address; Object–Field–Value; linearization; serialization; Word; PDF; mobile representation; scientific communication; structure reconstruction; comparative efficiency; controlled redundancy; practical validation.