A Relational Model of Data for Large Shared Data Banks
Communications of the ACM, 1970 · paper · Computer science
Proposes a data model based on n-ary relations, with a normal form and a universal data sublanguage, so users are shielded from how data is stored. It also discusses operations on relations and their use for redundancy and consistency.
Checked: PDF from engineering.upenn.edu, fetched 4 Oct 2026. Page numbers are those of that PDF.
6 findings
Reference or label Section 1.5 reads 'instead of only n + 1 with direct n-ary notation as described in Section 1.2', but n-ary relations are introduced in Section 1.3 ('A Relational View of Data'). Section 1.2 is 'Data Dependencies in Present Systems'.
- p. 6 Section 1.5, paragraph on LEAP
n-ary notation as described in Section 1.2
- p. 6 Section 1.5, nested binary relations paragraph
direct n-ary notation as described in Section 1.2
- p. 6 Section 1.5, paragraph on nested binary relations
with direct n-ary notation as described in Section 1.2
- p. 1 Section 1.2 heading
1.2. DATA DEPENDENCIES IN PRESENT SYSTEMS
Judge's note: The page 1 image shows the heading '1.2. DATA DEPENDENCIES IN PRESENT SYSTEMS'. The page 3 image shows the heading '1.3. A RELATIONAL VIEW OF DATA', which defines relations, domains and degree n. The p.6 crop confirms that the printed pointer reads 'Section 1.2'. The intended target appears to be 1.3.
Second check: another model agreed
Report a problem with this finding
Reference or label Structure 4, file F, segment PART lists 'part # / part description / quantity-on-hand / quantity-on-order' and omits 'part name'. The text says part name is recorded for each part, and Structures 1, 2, 3 and 5 all include it.
- p. 2 Structure 4, file F, segment PART
part description
- p. 2 Structure 4, segment PART, file F
part # part description quantity-on-hand quantity-on-order
- p. 2 Section 1.2.3 text before Structures 1-5
the part number, part name, part description, quantity-on-hand, and quantity-on-order are recorded
Judge's note: The page 2 image confirms Structure 4's PART segment has no part name. No other segment of Structure 4 carries it. The 'alpha' query on page 3 prints part name, and the text says any of the structures may hold the information mentioned.
Second check: another model agreed
Report a problem with this finding
Reference or label Summary 2.4 reads 'only a few of the more important properties of the data sublanguage in Section 1.4 are mentioned', but the data sublanguage is discussed in Section 1.5 ('Some Linguistic Aspects'). Section 1.4 is 'Normal Form'.
- p. 11 Section 2.4 Summary
the data sublanguage in Section 1.4 are mentioned
- p. 11 2.4 Summary, 'Many questions are raised' paragraph
properties of the data sublanguage in Section 1.4 are mentioned
- p. 5 Heading of linguistic section
1.5. SOME LINGUISTIC ASPECTS
Judge's note: The p.5 image shows the heading '1.5. SOME LINGUISTIC ASPECTS' followed by the sublanguage discussion. The text layer gives Section 1.4 as 'NORMAL FORM'. Outside the abstract, 'sublanguage' occurs only in Section 1.5 (text lines 1230 and 1259) and in this summary. The p.11 crop confirms 'Section 1.4' is printed.
Second check: another model agreed
Report a problem with this finding
Reference or label Fig. 10 heads R·S as '(project supplier)'. Since R·S = π13(R*S) and R*S is (supplier part project), the columns are (supplier, project).
- p. 9 Figure 10 header
R. S (project supplier)
Judge's note: Fig. 6 heads R*S as (supplier part project), and π13 keeps columns 1 and 3 in that order. The paper says column order is significant. The row values happen to be symmetric, so only the label is affected.
Second check: another model agreed
Report a problem with this finding
Reference or label Fig. 11 heads T as '(project supplier)'. T = π13(U) with U (supplier part project) from Fig. 7, so the columns are (supplier, project).
- p. 9 Figure 11 header
T (project supplier)
Judge's note: π13 of Fig. 7 gives {(1,2),(2,1)} with supplier first. This is the same kind of labelling slip as Fig. 10, but in a separate figure, so it is kept as its own cluster.
Second check: another model agreed
Report a problem with this finding
Wording The definition of restriction requires 'the maximal subset R' of R such that πL(R') = πM(S)'. In Fig. 13, S contains (c, B) but no row of R has p = c. So π23(R') = {aA, bB} cannot equal π12(S) = {aA, cB, bB}; the R' shown fits only a subset condition.
- p. 9 Section 2.1.5 definition and Figure 13
is the maximal subset R’ of R such that
Only the authors can say what is right here.
Judge's note: From the page 9 image: R = 1aA, 2aA, 2aB, 2bA, 2bB; S = aA, cB, bB; R' = 1aA, 2aA, 2bB. R' is exactly the rows of R whose (p, j) pair appears in S, which is a containment condition, not equality. Either the definition or the example needs adjusting; only the author can say which.
Second check: another model agreed
Report a problem with this finding
7 minor (typos and the like; most checked by the judge only)
Wording In Figure 8, relation T (j s) is printed with four rows 'd 1 / d 2 / e 2 / e 2', so the tuple (e, 2) appears twice. Rows of a relation are meant to be distinct, and projecting the Fig. 9 joins onto (j, s) gives only {d1, d2, e2}.
- p. 8 Figure 8, relation T (j s)
e 2
- p. 8 Paragraph after Figure 9, end of Section 2.1.3
the points 2 = a; y = d; x = 2 have this property
Judge's note: The page 8 image shows T as d1, d2, e2, e2. I recomputed the projections: π31(U) from Fig. 9 U = {(d,1),(e,2),(d,2)}, and π31(U') gives the same three pairs. The projections onto R and S match Figure 8 exactly, so the fourth T row is a duplicate. One reviewer also said the text has 'z = a ... x = 2'. That is an OCR artefact: the image prints 'x = a; y = d; z = 2'. Only the duplicated row is confirmed.
Second check: another model agreed
Report a problem with this finding
Typo The cited author's surname is printed 'Levein' in both the text ('Levein and Maron [2]') and reference 2 ('LEVEIN, R. E.'); the usual spelling is believed to be 'Levien'.
- p. 1 Section 1.1
Levein and Maron [2]
- p. 11 Reference 2
LEVEIN, R. E., AND MARON, M. E.
Only the authors can say what is right here.
Judge's note: Rests on outside knowledge: that the author of the Comm. ACM 10, 11 (1967) paper spells his name 'Levien'. Both page images show 'Levein', spelled the same way in text and reference list, so only the authors or the source can settle it.
Second check: none (judge only)
Report a problem with this finding
Typo Misspelling 'made to exceute a reference to a nonexistent file'; it should read 'execute'.
- p. 3 Section 1.2.3, paragraph after Structures
made to exceute a reference to a nonexistent
Judge's note: The page 3 image shows 'exceute'. The same sentence later spells 'execute' correctly.
Second check: none (judge only)
Report a problem with this finding
Typo Footnote 4 reads 'M. E. Sanko of IBM, San Jose, independently recognized...', but the Acknowledgment reads 'M. E. Senko of the IBM San Jose Research Laboratory'.
- p. 5 Footnote 4
M. E. Sanko of IBM, San Jose
- p. 11 Acknowledgment
M. E. Senko of the IBM San Jose
- p. 11 Acknowledgment
M. E. Senko of the IBM San Jose Re-
- p. 5 Footnote 4
M. E. Sanko of IBM
Judge's note: Both page images show the two spellings: 'Sanko' in footnote 4 on p.5 (journal p.381) and 'Senko' in the Acknowledgment on p.11. The initials and the affiliation are the same, so this is evidently one person spelled two ways.
Second check: none (judge only)
Report a problem with this finding
Typo Relation T (j s) in Fig. 8 prints four rows 'd 1, d 2, e 2, e 2'. The row 'e 2' is duplicated, contrary to property (3), 'All rows are distinct'.
- p. 8 Figure 8, relation T (j s)
FIG. 8. Binary relations with a plurality of cyclic 3-joins
Judge's note: The crop shows the duplicated row. Projecting U and U' of Fig. 9 onto (j, s) gives only {d1, d2, e2}, so the fourth row looks like a typesetting duplicate. The example is otherwise unaffected.
Second check: none (judge only)
Report a problem with this finding
Typo The running footer on page 386 reads 'Communications of the AMC' instead of 'ACM'.
- p. 10 Page footer
Communications of the AMC
Judge's note: The crop confirms 'AMC'. Every other page footer reads ACM.
Second check: none (judge only)
Report a problem with this finding
Typo Reference 8 reads 'IDS Reference Manual GE 625/635, GE Inform. Sys. Div., Pheonix, Ariz.'; the city is misspelled 'Pheonix'.
- p. 11 Reference 8
Pheonix, Ariz.
Only the authors can say what is right here.
Judge's note: Rests on outside knowledge: the city in Arizona is spelled 'Phoenix'. The p.11 crop shows 'Pheonix' printed in reference 8.
Second check: none (judge only)