Unigrid: The National Park Service Publication System, Reconstructed

Why the National Park Service needed Unigrid

National Park Service visitor publications long predate Unigrid. As the National Park System expanded, its folders accumulated many formats, fonts, and folding methods. As their number grew, the publications became harder to produce and maintain. In 1977 Gleason sought a more efficient and cost-effective approach.²

The 1985 specification states the problem exactly. It identifies two primary tasks in folder planning. The first is organizing editorial and graphic components within a format. The second is determining how the folder will be printed: inks, paper, and sheet size.¹

Graphic order and production economy enter together. Format, paper, grid, fold, and type become known conditions before the particular park story is resolved. The designer can then give more attention to the specific publication: the territory a map requires, the sequence through which a history should unfold, the position of visitor information, or the relationship between an image and the text beside it.

Unigrid addressed institutional scale through standardization and left room for each publication to differ.

Vincent Gleason and the publications program

Vincent Gleason was Publications Chief when development of Unigrid began. National Park Service sources trace the program’s institutional beginning to Gleason: he sought a new publication system and engaged Vignelli’s expertise.²

Gleason had been thinking about the conditions of this work for years. In April 1964, as Chief of the Publications Branch, he proposed bringing publications, museum and exhibit work, audiovisual presentations, and motion pictures together in purpose-built facilities at Harpers Ferry. His proposal argued that closer contact among the creative branches would improve communication and produce more unified work for individual parks.³

That earlier proposal helps explain what Gleason was trying to solve: who worked together, where the work was made, which equipment and spaces supported it, and how different interpretive media could share a common purpose.

Vignelli’s recollection gives the other side of their meeting. In his 2011 oral history, he remembered giving a lecture in Washington with Gleason in the audience. Gleason approached him afterward and later came to New York to discuss the Park Service’s publication problems. Vignelli recalled that those discussions led to the system called Unigrid.⁴

Massimo Vignelli and the design of Unigrid

The 1985 specification identifies Massimo Vignelli as consulting designer and states that the system is based on his evaluative analysis of existing National Park Service folders.¹

Vignelli’s oral history describes a program wider than a page grid. He recalled establishing a grid and format, a direction for information, rules for images, cartography, writing, folders, books, and other publications. He also described regular visits to Harpers Ferry and Washington as Park Service designers began working within the system.⁴

In his account, Unigrid is a language practiced by many people. Vignelli did not claim authorship of every brochure. He described a system used by Park Service designers to create continuity across the program and respond to individual assignments.

The Vignelli Center identifies members of the Park Service design team as Melissa Cronyn, Nicholas Kirillof, Dennis McLaughlin, Linda Meyers, Philip Musselwhite, Mitchell Zetlin, and Vincent Gleason, Chief of the Division of Publications. Later histories give Dennis McLaughlin a specific design credit on the 1977 system. Together, these records place Vignelli’s consulting role within a wider Park Service design team.⁵⁶

Clara Barton National Historic Site, 1978

The National Park Service identifies the 1978 Clara Barton National Historic Site brochure as the first park brochure produced through Unigrid.²

The brochure is the first park-specific Unigrid publication. The black identification band, the relationship of title to field, the broad horizontal organization, and the integration of text and image can be read in an object made for a specific historic site rather than in the abstract language of a specification.

The timing matters too. Development began in 1977; by 1978 Unigrid had already left the drafting table and become something people could fold, carry, and use.

Evidence classes

This study uses four evidence classes.

A derived value is only as sound as the evidence it starts from. Exact arithmetic establishes the consequence of a premise; it cannot strengthen that premise as historical evidence. A numerical conversion introduced for comparison is still a conversion. A browser rule is contemporary even when it closely preserves a historical relationship.

Primary sources and provenance

The principal technical source is the 1985 reprint of Unigrid Design Specifications: National Park Service Informational Folder Program, issued by the National Park Service Division of Publications. The two-sided sheet documents formats, grid dimensions, title treatment, typography, colour, maps, pictorial methods, broadside logic, paper, folding, and production.¹

NPSHistory.com hosts the facsimile used here. It is an independent archive; the National Park Service document is the historical source. Citations name the issuing body, title, date, side, and internal section where possible.

The National Park Service’s Harpers Ferry Center history records the institutional chronology and the agency’s present account of Unigrid.² Vincent Gleason’s 1964 proposal records his earlier approach to the institutional conditions of publication and interpretive media.³ Massimo Vignelli’s 2011 Smithsonian oral history gives his first-person account of meeting Gleason and developing the system.⁴ The Vignelli Center holds publication records and names members of the Park Service design team.⁵ Meggs’ History of Graphic Design provides a secondary attribution record for Vignelli, Gleason, and Dennis McLaughlin.⁶

UniDescription independently describes the boxes in the historical grid illustration as twelve wide and seventeen deep.⁷ It corroborates a count directly visible in the primary drawing.

These sources perform different work, and each is used only for what it can establish.

Methodological context

The method draws on several neighbouring traditions: design history, generative design, cultural-heritage reconstruction, digital preservation, and software preservation. Several of these traditions treat designed artifacts as evidence of rules and document how reconstructions follow from their sources.

Simon treats designed systems as objects of systematic study, examined in relation to their purposes and environments.⁸ Stiny and Gips made the argument visual through shape grammars: classes of forms can be specified by generative rules rather than catalogued only as finished examples.⁹ This paper does not claim that Unigrid is a shape grammar in their formal sense. Their distinction allows finished artifacts to be studied as outcomes of a rule system.

The London Charter requires research sources to be identified and evaluated, uncertainty to be stated, and the analytical, deductive, interpretative, and creative decisions connecting sources to computational outcomes to be documented.¹⁰¹¹ Hugh Denard describes this requirement as a problem of intellectual transparency: a digital reconstruction should allow a reader to understand the character and limits of the knowledge claim it makes.¹² The four evidence classes used here apply that requirement to individual geometric relationships.

Digital-preservation research poses a related question: which properties must survive when an object changes technical environment? The InSPECT project developed methods for identifying “significant properties” whose preservation supports continued access, use, meaning, authenticity, and integrity across transformations.¹³ Christoph Becker has argued that such properties become meaningful within preservation systems and performances rather than existing as self-evident technical attributes.¹⁴ This reconstruction concentrates on relationships that can be established from the evidence and formalized for another medium. The larger historical system includes editorial judgment, cartography, image commissioning, print procurement, folding, and institutional practice.

Software introduces a further preservation layer. Di Cosmo and Zacchiroli argue for source code as a first-class preservation object because it contains human-readable technical knowledge and records the evolution of software systems.¹⁵ Here, source code and verification reports record the contemporary implementation. The historical specification records the evidence used to reconstruct selected relationships. The two records answer different questions, so the paper keeps them separate.

In this paper, executable preservation means encoding historical relationships so they can generate states, reject invalid ones, and be tested while their provenance remains visible. Here, the term applies only to the relationships encoded in the model.

The ten formats

Unigrid provides ten basic formats in two series. A-Series formats are one folded panel wide. B-Series formats are two folded panels wide. Each series includes depths of one, two, three, four, and six panels. A5 and B5 are omitted because those sizes cannot be cut economically from the two specified printing sheets.¹⁶

The folded panel measures 99 × 210 millimetres. From that panel structure, the open formats are:

A6 can be doubled to make A12 where an extra-length map is useful for a parkway, trail, canal, or another linear subject. The specification uses accordion folds for the format family.¹⁶

The name B6 belongs to the National Park Service’s internal format system. The specification states that its 420 × 594 millimetre B6 corresponds to DIN A2. It therefore differs from the much smaller paper size commonly called DIN or ISO B6.¹⁶ The family remains recognizable across very different publications.

The broadside as the organizing field

The fully opened sheet, or broadside, is the main field for layout. Fold lines and the panels they create remain part of the object, while the layout is organized across the larger sheet.¹⁷

A park folder is handled in motion. It is opened to find a map, turned to locate a piece of information, folded again, placed in a pocket, and opened later somewhere else. The specification responds by treating the folder more like a chart or map than a sequence of small book pages. Accordion folds allow either side to open quickly, and content remains oriented to the larger broadside.

Maps, photographs, drawings, titles, and text can therefore be arranged across the full sheet. The fold determines the physical object. The layout is organized at the scale of the broadside.

Grid and module

The grid is the common coordinate structure beneath the format family. The 1985 specification describes acetate overlays used to locate graphic components during layout and mechanical art. The grid remains constant as the formats change.¹⁸

The A- and B-Series use related modules. Both are 80 points high. The A-module is 6½ picas wide. The B-module is 7 picas wide. Spaces between module columns are one pica; spaces between module rows are 10 points. Since one pica equals 12 points, the B-module measures 84 × 80 points, separated by 12 × 10 point intervals.

Three units of measure appear in the specification. Trim sizes are stated in millimetres. Horizontal grid measurements are stated in picas. Vertical grid measurements are stated in points. The source explains that typographers and compositors work in the latter two units.¹⁹

The specification describes the grid as an unseen supporting structure. On the specification sheet itself, the reader is instructed to hold the paper to a strong light so the grid on one side can be seen beneath the prototype on the other. It is a wonderfully literal demonstration of the principle: the grid is present even when it disappears from view.

Typography

The 1985 specification names Helvetica as the house typeface for Unigrid. It specifies Helvetica for display type, body text, maps, tables, and captions, connecting the choice to offset printing, the available range of weights and sizes, and Helvetica’s large x-height.²⁰

The title bar carries the most explicit hierarchy. Park names shorter than twelve letters are set in 60-point Helvetica Medium, aligned with the first grid module. Longer park names use 42 point. In the B-Series, the site designation and location are set in 12/14 or 8/9 Helvetica Medium and align with the seventh module. Service and Departmental designations align with the tenth. Title type hangs 10 points below the top trim.

B-Series body text uses 8/10 or 9/10 Helvetica Regular or Medium in measures of 15 or 23 picas. A-Series measures are 14 or 21½ picas. Text is justified. Text banks align at their upper and lower edges, while the far-right column may end short. Paragraphs carry no indent and are separated by one open line.

Major display type uses 12, 18, or 24 point Helvetica Medium and sits 10 points above its related text. Captions use Helvetica Regular or Medium, roman or italic, at 7/7, 7/8, or 8/9. Preferred B-Series caption measures are 7 or 15 picas.²⁰

The National Park Service’s later historical account records Helvetica and Times Roman among the program’s original typefaces and Frutiger and NPS Rawlinson in later standards. The 1985 specification therefore documents one stage in a system that continued to develop.²

Titles, bars, and recurring horizontal relationships

Across both series, the cover title treatment is uniform. The title bar is 100 points high and black; titles reverse in white. The specification describes the bar as taking on the quality of a logotype for the folder program.²¹

Horizontal organization continues through bottom bars, divisional bars, margins, and fascia. Bottom and divisional bars are 25 points high. The specification connects their regularity to layout and maintenance: horizontal parts can be exchanged as publications are revised.

A 40-point horizontal space is maintained above text entries, and a bank of text carries at least a 20-point drop below. These relationships give the broadside a recurring horizontal rhythm that works with the module grid without requiring every element to fill a complete module.

Colour

The 1985 specification identifies 24 standard colours.

The document also identifies twelve preferred bright flat inks and eight muted flat inks. Colour is used across maps, scientific and historical illustration, safety information, photographs, and background fields.²²

The restricted palette gave the office a common production vocabulary and left considerable variation between brochures.

Maps and pictorial material

The specification calls for larger maps with a more uniform graphic treatment. It keeps maps free of unrelated text and decorative image clusters so cartographic information can be found and maintained more directly. Large maps may carry a 25-point border containing location keys where useful.²³

Images may sit within a toned band, join the cover assembly, or form an illustrated essay across the broadside. Images can occupy small areas or several large grid segments, depending on the hierarchy of the story.

Printing, paper, and folding

The production plan uses 70-pound dull-coated text stock in two press-sheet sizes: 635 × 965 millimetres and 965 × 1270 millimetres. The larger sheet accommodates all ten basic formats. The smaller accommodates every format except A4 and B4.²⁴

Both sheets run through the press on the short side. Paper grain therefore lies in the long dimension, and the format and fold system is planned with that grain direction in mind.

Format family, paper stock, grid, press sheets, inks, and folds belong to one production method. A coordinate on the designer’s overlay is connected, through the system, to a physical sheet that will pass through a press and later through someone’s hands.

For this reconstruction, that connection matters: the grid ties graphic decisions to material consequences in production.

Constraint and variation

Unigrid establishes recurring conditions before an individual folder is laid out. Format, module, title treatment, type hierarchy, paper, and production method are known in advance.

The specification calls this allocation of content the Design Platform. Park staff and the creative team determine the intended use of the folder, identify its essential components, estimate how much space each requires, make a rough dummy, and prepare text, maps, and images to that plan.²⁵

A large park may give half of its broadside to a map. A historic site may organize its sheet around archival material or a reconstructed scene. A linear landscape may choose an elongated A-format. Each publication is made within the same system of formats, grids, typography, and production rules.

Vignelli later described Park Service designers working within a common language while producing hundreds of individual publications.⁴ That is the useful tension in Unigrid: the system decides a great deal before the designer begins, but it does not decide the publication for them.

Established geometry

The module, intervals, title bar, and bar thicknesses come from the specification’s measurement and title sections.¹⁹²¹ The twelve-by-seventeen exposed matrix is counted from the published drawing and corroborated by UniDescription.⁷ Arithmetic is kept separate from those observations.

The exposed field

The B6 illustration contains twelve complete module columns and seventeen complete module rows fully exposed beneath the title region. The count is directly visible in the historical drawing. UniDescription independently describes the same illustration as boxes “stacked 12 wide and 17 deep.”⁷

The exposed field contains 204 complete modules:

Its dimensions in the grid’s native point system are:

The point dimensions are the result that matters here. Converting them to millimetres requires a point convention that the historical specification does not state. The document itself mixes millimetres for trim, picas for horizontal grid measurements, and points for vertical grid measurements without giving a conversion rule between those systems.¹⁹

For comparison, a 72-point-per-inch PostScript convention gives approximately 402.17 × 536.22 mm for the exposed field; a 72.27-point-per-inch traditional convention gives approximately 400.66 × 534.22 mm.²⁶ These converted values are contemporary comparisons, not additional historical measurements.

Vertical phase and the title-region row

The B-module is 80 points high and the interval between rows is 10 points. The vertical pitch is therefore:

In the source drawing, the first fully exposed row begins at the lower edge of the 100-point title region. Extending the established 90-point cadence upward by one pitch gives:

The preceding module occupies:

followed by the normal interval:

The title region therefore accommodates the sequence:

One complete 80-point module row lies within the title region, with 10 points above it and the normal 10-point row interval below it. The source does not state this in prose. The result follows from the specified title height, specified module and interval dimensions, and the phase of the grid shown in the drawing, so it remains a derived reconstruction.

Including that row, the lattice from its first module start through the end of the seventeenth exposed row contains eighteen complete module rows:

Across twelve columns, those eighteen complete row positions contain:

module positions. The number 216 describes a reconstructed extent of the lattice. The graphically exposed field remains 204 modules.

Lower-edge sensitivity

The established cadence begins another row before the lower trim. The exact amount of that row retained by the sheet depends on the point convention used to compare it with the 594 mm trim.

With sheet coordinates beginning at the upper trim, the eighteenth complete row ends at:

the following interval ends at:

and the next 80-point module begins there. A complete module would end at:

Two conventional point systems give different numerical trim positions.²⁶

Both conventions lead to the same geometric conclusion: the cadence reaches the start of another row at 1,630 points, and the 594 mm sheet cannot contain that 80-point row in full.

The specified lower bar is 25 points high.²¹ If its lower edge is registered to the lower trim as depicted, the amount of the projected row visible above the bar is approximately 28.8 points under the PostScript convention and 35.1 points under the traditional convention. That partition remains conditional on both the point convention and the depicted registration of the bar.

Together, Figures 13a–13c form the evidence plate for the reconstruction: source drawing, annotated count, reconstructed lattice.

Panels and modules

B6 carries twelve physical folded panels on each side: two across and six deep.¹⁶ Panels and modules describe different structures. The panels determine the physical fold pattern of the broadside. The twelve-column module field determines graphic position and span. Their shared number on one axis does not make them interchangeable.

Coordinate model

To keep sheet registration separate from local field addressing, the reconstruction uses two coordinate systems.

Sheet-space lattice. Sheet coordinates begin at the upper-left trim corner. The vertical lattice is registered from the phase established above:

where $r$ is a zero-based sheet-space lattice row index. The module occupies:

Under the reconstruction:

is the complete row within the title region;

are the seventeen complete rows fully exposed beneath it; and

begins at 1,630 points and is clipped by the lower trim.

Local exposed-field coordinates. Within the twelve-by-seventeen exposed field, let $i$ denote the zero-based column and $j$ the zero-based row:

The corresponding sheet-space row is:

An exposed-field element is:

where $w$ and $h$ are positive integer spans. Its dimensions are:

Valid local placements satisfy:

The reconstruction establishes horizontal module dimensions and relative positions within the twelve-column field. It does not establish an absolute horizontal field origin against the left trim. Horizontal coordinates are therefore local to the field unless further evidence establishes an $X_0$. Vertical sheet registration is reconstructed separately from the visible phase of the historical drawing.

Keeping the coordinate systems separate avoids treating the residual at the sheet edge as evidence of symmetrical historical margins.

From geometry to rule

The arithmetic turns evidence into explicit relationships. Modules and intervals become dimensions; positions and spans become integer coordinates and counts. Trim, title and divisional bars, folds, and the module field remain separate parts of the geometry.

The model then supports three operations:

1. Generation. The rules generate new rectangles.
2. Validation. The model rejects rectangles whose coordinates or spans exceed the declared field.
3. Transformation with provenance. A relationship can be implemented in another medium with its historical source and contemporary treatment recorded separately.

Transferable relationships

The software implements five selected relationships from the reconstruction:

fixed module dimensions; fixed intervals; integer coordinates and spans; a shared lattice; material capacity.

The unit can change while these relationships remain explicit. A typographic point, millimetre, and CSS pixel have different physical meanings. What transfers is the structure of the geometry: fixed dimensions, integer coordinates and spans, shared edges, and the number of complete modules the available material can contain.

A shared lattice can be stated formally. Let:

be the horizontal and vertical pitches. For a field origin $(X_0,Y_0)$, module start and end edges are:

A cell with integer coordinates $(i,j)$ and integer spans $(w,h)$ occupies:

Independently authored components therefore begin and end on edges generated by the same module-and-interval construction. Shared lattice names that common coordinate discipline.

Fixed size across historical formats

The historical specification states that type, maps, and illustrations keep their size as the publication format changes. Smaller formats carry a different amount of material rather than proportionally scaled versions of the same graphic components.²⁷

The digital implementation keeps the constituent geometry fixed while available space determines how much fits.

Material capacity

Let:

For available width $W$, the number of complete modules that physically fit is:

A field of $N$ complete columns occupies:

The residual width is:

When that field is centred:

The equations require no device categories. Width changes alter the number of complete modules the material can contain while module dimensions and intervals remain fixed.

Authored composition

An authored composition declares a finite horizontal coordinate field of $A$ columns. Cells may carry:

for integer position and span. The active column count is:

At full authored capacity, explicit coordinates can be honoured as one two-dimensional arrangement. Below that capacity, the implementation releases explicit positions together and continues in source order unless another authored behaviour has been declared. Releasing them together prevents collisions and unintended reading order.

The authored ceiling and coordinate-release behaviour are adapted implementation rules.

Fluid composition

A fluid composition declares no authored horizontal coordinate field. Cells carry span and source order, and the field uses complete modules permitted by available width:

There is no coordinate threshold to release. Authored and fluid composition use the same module geometry with different placement rules.

Coordinates and spans remain whole

A cell spanning $w$ columns occupies:

A cell spanning $h$ rows occupies:

Position and span are integer relationships. Responsive behaviour does not multiply them by a viewport ratio. If a cell is wider than the active field, the implementation may expose overflow, release its authored position, or apply another explicit rule. Shrinking the module or proportionally remapping its span would alter the declared geometry.

What changes on screen

Paper is finite in both axes. A B6 sheet has a fixed physical width and height. A browser field has variable inline width while the document can continue in the block direction. Runtime measurement, source order, scrolling, reflow, overflow, and continuing rows become part of the material condition.

Those are properties of the browser implementation, so this paper classifies them as adapted.

Paper resolves these five relationships against a bounded sheet. Screen resolves them against available width and continuing document height.

Numerical values are implementation parameters

The historical B-module measures 84 × 80 points with 12 × 10 point intervals. Those numbers belong to the print specification.¹⁹

The KhannaOffice web implementation uses the same numerical parameters in a different unit system: 84 × 80 CSS-pixel modules with 12 × 10 pixel intervals.²⁸ These are adapted implementation parameters. The shared numerical values are deliberate. Typographic points and CSS pixels retain different physical meanings.

A twelve-column digital field therefore occupies:

The implementation preserves the numerical module and interval relationships while available width determines how many complete modules fit.

The implementation also distinguishes a finite authored model from a fluid model with no authored coordinate ceiling. Twelve columns describe the historical B6 field and can also describe a particular authored digital composition; they do not impose a universal ceiling on screen layout.

The field widths therefore correspond numerically: 1,140 pt historically / 1,140 px digitally. The units retain their separate physical meanings.

Representation

A software implementation can represent the transferable geometry with four primary parameters:

module width  = m_x
module height = m_y
column gap    = g_x
row gap       = g_y

A cell is represented as:

E = (x, y, w, h)

for authored compositions, or by span and source order for fluid compositions.

The representation is intentionally small. It only needs to hold the relationships that determine position, span, alignment, and capacity.

Validation

For a finite field of $n_x$ columns and $n_y$ rows, an authored element is valid when:

An implementation can add further rules for overlap, duplicate identifiers, incomplete coordinates, or source-order requirements.

The specification's dimensions define which states are valid.

React component model

The public package is @khannaoffice/unigrid. The software record cited in this paper is release 1.0.0 at commit 931c228ccaf7eff72b7af0f565bc5e88d5b8810f.²⁸ A composition can be written directly in modules:

import {
  Unigrid,
  Cell,
} from "@khannaoffice/unigrid";

<Unigrid authoredColumns={6} initialColumns={6}>
  <Cell x={0} y={0} w={4} h={2}>
    Body
  </Cell>

  <Cell x={4} y={0} w={2} h={1}>
    Record
  </Cell>
</Unigrid>

Unigrid establishes the field and its geometry. Cell expresses position and span.

The package also exposes pure geometry, placement, validation, capacity, and audit functions. These functions let the geometry be inspected independently of React rendering.

Historical verification

Specified values are read from the 1985 specification. Graphically specified relationships are counted from the source drawing. Derived values are recalculated from identified inputs. Inferred relationships remain labelled as inference.

The B6 reconstruction is presented as a single sequence of evidence:

1. the untouched historical drawing;
2. the annotated twelve-column by seventeen-row exposed count;
3. the derived 1,140 × 1,520 point exposed field;
4. the 90-point cadence extended through the title region;
5. the lower-edge sensitivity analysis.

Figures 13a–13c present the source drawing, counted field, and reconstructed lattice used by the calculations.

Software verification

The software evidence cited here comes from release 1.0.0 at a fixed commit.

Release @khannaoffice/unigrid 1.0.0 is identified by commit:

931c228ccaf7eff72b7af0f565bc5e88d5b8810f

The checked-in deterministic report verification/local-verification.json was generated on 12 August 2026 at 03:12:30 UTC on Node v26.3.1. It records allChecksPassed: true, a valid twelve-state material audit, and capacities from one through twelve columns. Every state reports 84 × 80 pixel modules with 12 × 10 pixel intervals; the twelve-column state resolves to 1,140 pixels.²⁹

Separate checked-in Playwright reports cover React 18.3.1 and React 19.2.6 across Chromium, Firefox, and WebKit. Each report records 123 expected checks, zero skipped, zero unexpected, and zero flaky.³⁰ Those results apply to the release files and runtime versions named in the reports.

Release 1.0.0 predates the corrected B6 row registration. Its reports verify the published screen geometry and browser behaviour, not the revised historical reconstruction.

Historical scope

Unigrid is larger than the B6 geometry reconstructed here.

The historical system also includes editorial practice, cartography, pictorial commissioning, colour, print procurement, maintenance, fold planning, and decades of institutional development. The executable model concentrates on modular and material relationships.

The 1985 specification is one documented state of a long-running program. Later National Park Service practice changed typography, imagery, maps, and production while retaining a recognizable family structure.²

Attribution

Institutional leadership, consulting design, staff design, later maintenance, historical reconstruction, and software implementation remain separate forms of authorship.

National Park Service sources establish Gleason’s initiating institutional role and Vignelli’s involvement. The 1985 specification identifies Vignelli as consulting designer. His oral history describes both his own contribution and the work of Park Service designers using the system. The Vignelli Center records the names of the wider design team.

KhannaOffice’s reconstruction and software are later work and do not settle unresolved facts in the original record.

Geometric scope

The twelve-column by seventeen-row exposed field is graphically established. The title-region row is a derived reconstruction from specified dimensions and visible phase. The lower continuation is also derived: both point conventions tested here place its start before the lower trim, while its exact division at the bottom bar depends on the convention and the depicted registration of that bar.

Absolute horizontal registration remains unresolved. The source establishes the module dimensions and relative twelve-column field, but not an $X_0$ against trim from which symmetrical margins could be inferred.

Digital scope

The web implementation adds CSS units, authored and fluid composition, material-capacity rules, source-order behaviour, scrolling, overflow, runtime measurement, and continuing document height. These are contemporary adaptations.