Freeplan

The work and its context

Freeplan presents architectural drawings whose components seem to have acquired a measure of independence. Floors overlap, walls stand apart, a circular pool interrupts an orthogonal arrangement, and a path may continue beyond the sheet. The familiar furniture of a plan remains: grids, section marks, a north sign, a scale bar and a numbered title block. Yet those conventions give authority to spaces that do not resolve into a conventional sequence of enclosed rooms. ^{1 2 3 4 5}

The work is project 245 on its Art Blocks contract, credited to xnmtrc. The current edition contains 128 minted outputs. Each is a particular arrangement of shared elements, not a different view of one underlying building. The five originals examined here range from a pale grey rendered drawing to a blue field with white lines, a sepia sheet and a blue-pen treatment. Their representational differences change how substantial or provisional the proposed spaces feel. ^{6 1 2 3 4 5}

The strongest tension lies between the drawing's apparent precision and its open spatial proposition. A scale bar invites measurement; a detached wall invites speculation about what it might divide. Neither response cancels the other. The image can be read as an abstract composition, but its architectural signs keep prompting the viewer to imagine passage, shelter and occupation. These are interpretations of the drawings, not claims that the program produces construction-ready buildings. ^{7 5}

Conception and development

In a retrospective account, xnmtrc connects the project to early modernist attempts to free space from the conventional box. The artist describes experimenting with a system that would assign components and volumes algorithmically, and gives a three-month span for conception, coding and minting. Later, selected outputs were developed into three-dimensional models. This account establishes an architectural investigation behind the graphic vocabulary; it does not make every generated plan a tested building proposal. ⁸

The choice of plan view is productive because it withholds as much as it describes. Height is largely absent, while the placement of walls and floors can be judged immediately. Shadows add a suggestion of projection without supplying a complete elevation. The viewer is invited to move mentally from marks to possible space, then back to the marks when an arrangement resists easy visualisation. That oscillation is especially clear where a wall extends beyond a tiled floor or where an isolated bar seems to belong to no closed room. ^{1 5}

The title block gives each sheet the appearance of a discrete proposal. Its number changes with the token, while the common label and scale graphic establish a family. This is a form of serial authorship in which the shared drawing language matters as much as any single composition. Comparing sheets makes the system's choices visible: not simply whether a pool exists, but what the pool does to the hierarchy of circles, bars and tiled fields around it. ^{7 1 2}

How the work was constructed

The complete indexed script uses p5.js 1.0.0. It makes a square canvas from the smaller window dimension and divides that square into twenty modules. Stroke weights and most dimensions derive from the resulting unit. The drawing is produced in one pass: setup calls noLoop, and the drawing function assembles the layers. No ongoing animation, sound or interaction mechanism appears in the inspected program. ^{6 7}

A seeded generator supplies the decisions, but its initial parsing deserves a qualification. The code places the number 16 inside the substring call rather than passing it as the radix to parseInt. The seed fragments are therefore not converted in the usual explicit hexadecimal manner. Decimal prefixes can be read only as far as the first incompatible character, and nonnumeric starts become NaN before bitwise operations coerce them. This is a code-level observation, not an audit of collisions or a claim that the whole edition repeats. The visibly different inspected outputs should not be explained by assuming a standard full-hex conversion that the code does not perform. ⁷

Other initial choices establish wall count, thickness, orientation tendency, component conditions and a seven-style palette system. The orientation threshold is obtained from a random-decimal method that ignores the arguments supplied to it, so the apparent intended narrow interval is not enforced by that call. The threshold can favour one direction more strongly. This helps frame the program's available behaviour, without claiming to have traced every choice in a particular output. ⁷

The layer order begins with paper-like backgrounds and optional context curves, then adds a path and two floor rectangles. The backgrounds can be plain, dotted, gridded or marked with small crosses. Context curves provide a topographic convention rather than measured terrain. The floors have their own subdivisions, whose intervals need not match one another. Overlapping tile fields therefore produce local changes of density even before the walls are drawn. ⁷

Between three and seven walls are generated. Their lengths come from a sequence of choices among short, medium and longer ranges; the routine is not a single uniform draw over all lengths. Coordinates are gathered from banded ranges, checked for duplicate values and shuffled separately for the two axes. This constrains placement, but it should not be described as a comprehensive architectural collision or feasibility solver. The inspected procedure contains no general structural analysis. ⁷

All wall shadows are drawn before the walls themselves. The walls are rectangular bars with selected fills, possible diagonal hatching and optional extensions at their ends. The shadows use repeated blurred, slightly displaced shapes. Their common direction is related to the sign choices that also inform the north annotation. This coordination makes the sheet feel coherent, even though the shadows are graphic devices rather than a detailed environmental lighting calculation. ⁷

Optional columns can be square, circular or cross-shaped and can occupy a small repeated array. Their marks are much smaller than the walls, so the viewer must distinguish them from the paper grid. A roof can be absent, indicated by a dotted rectangle or represented as a filled plane. The order matters: a roof arrives after the plan and can alter how the underlying arrangement reads. Section lines add another kind of information, with dashed runs and either lettered or hooked ends. ⁷

The circular pool is built from concentric forms, radial rim divisions and a softly graded interior. Trees use different circular or radial drawing procedures and may number from zero to five. A metadata label naming a tree type does not establish that a visible tree is present: the count is a separate decision. AB5D's #7, for example, has a tree-style label but no trees visible in the inspected original. Reading feature names without looking would produce an inaccurate description. ^{7 5}

Finally, the program writes the edition number, Freeplan title, segmented scale bar and circular north indicator. These annotations make the image resemble a professional drawing, but their presence is not proof of a viable circulation plan, load-bearing structure or specified material assembly. The work uses the language of architectural exactness to open a field of possible relationships. That distinction preserves both the specificity of its reference and the freedom of its pictorial result. ⁷

The mint

The first indexed token is dated 15 January 2022. Activation is recorded on 17 January, and the scheduled opening is 20 January at 17:00. The second token appears at 16:59:39 that day, 21 seconds before the schedule. These values describe different events and should remain separately named. An early token is not sufficient evidence of an earlier ordinary public sale. ⁶

The completion field is 21 January at 19:49:26, or 26 hours, 49 minutes and 26 seconds after the scheduled time. This is an interval between recorded fields, not an independently classified public sellout duration. The artist's later account recalls roughly half the works selling within a few hours and completion close to the release day. That rounded recollection can coexist with the index's more precise dates without being presented as a stopwatch measurement. ^{6 8}

The same retrospective account identifies an original maximum of 128, providing historical support beyond the current index's 128 invocations and maximum. The current price setting is 0.15 ETH, but the opening price and amounts paid by individual recipients were not independently recovered. Administrative and artist mints have not been separated from public transactions in this review. ^{8 6}

Copyright and licensing

The current artwork designation is CC BY-NC 4.0. The project record supplies no terms link, so the named Creative Commons deed is cited directly. It permits sharing and adaptation under attribution and noncommercial conditions, including indication of changes. This is the observed artwork label; a full history of earlier terms has not been reconstructed. ^{6 9}

No separate comprehensive reuse grant for the artist's script was established in the inspected material. Its implementation is discussed here without republishing it. The p5.js 1.0.0 dependency has its own LGPL 2.1 licence, which does not establish an independent licence for the artist's program. AB5D's CC0 dedication covers the original essay prose, with those other rights kept distinct. ^{7 10}

Close reading of five outputs

The comparison uses #0, #1, #10 and #20 as an early-edition group, together with AB5D's #7. Their full identifiers are 245000000, 245000001, 245000010, 245000020 and 245000007 on the project's Ethereum contract. All five original generator pages were opened and inspected. The collection example was selected through a fresh check of all linked AB5D wallets. These observations support a close comparison, not a frequency survey of all 128 works. ^{1 2 3 4 5}

In #7, a pale grid extends across the sheet while overlapping grey floor planes occupy the lower centre. The larger floor is densely divided by verticals; a smaller field behind it has a finer lattice. Five wall bars introduce a different weight of mark. A tall vertical at upper left rises into an otherwise open area, while a short horizontal sits above the floors and another vertical stands detached on the right. The long horizontal across the lower field and a final vertical near its corner imply divisions without closing them. ⁵

The empty upper-right region matters as much as the occupied lower-left. It allows the detached bars to seem deliberate rather than merely unfinished. Shadows raise the walls above the tiled fields, while their end ticks keep asserting that this is a drawing. The viewer switches between imagining solid pieces in space and reading precise line conventions on paper. Because floor edges and wall positions do not coincide, neither system can be treated as a simple outline of the other. ⁵

Edition #0 introduces a round pool near the lower right of a grey tiled floor. Its fine radial rim and pale blue centre interrupt the network of horizontal and vertical lines. A broad path extends off the right side, suggesting continuation beyond the sheet. A dashed section line passes through the floor and pool, but supplies no accompanying sectional view. It makes the image feel part of a larger documentation set while leaving that larger set to the imagination. ¹

The wall bars in #0 remain separated: some align with floor boundaries, others stand away from them. The pool's circular edge crosses that organisation rather than politely occupying a fully enclosed compartment. This makes it a particularly clear example of the work's spatial openness. The drawing offers recognisable elements of occupation, yet resists the familiar expectation that every element must be assigned to an already bounded room. ¹

Edition #1 replaces the pale field with deep blue and uses fine light marks. Two tiled rectangles overlap in the upper middle. A long hatched horizontal wall crosses them, a shorter bar sits above, and a detached vertical appears at lower right. A circular pool rests below the main floor area, while faint shrub-like crowns occupy the upper zone. The palette reduces the apparent weight of the walls: they seem traced or luminous rather than grey solid pieces. The drawing's lightness changes the same basic vocabulary into a more provisional proposition. ²

The sepia #10 is unusually layered. Six horizontal wall bars occupy different heights, including one long central run and a detached lower bar. Tile grids, a rightward path, a dotted roof boundary, small square columns and lettered section lines compete for attention. Tree bursts soften the upper edge. The largely horizontal walls do not produce a row of rooms; instead they suggest a sequence of screens encountered across overlapping fields. The sheet's technical detail intensifies its openness rather than resolving it. ³

In #20, blue lines on a pale grid articulate a stepped arrangement from upper left to lower right. Two floor areas carry different tile scales, seven wall bars alternate direction, and six cross-shaped columns make a separate small array. Their stronger blue distinguishes them from the quieter background grid. Two overlapping shrub crowns gather near the upper centre-left. The result depends on a hierarchy of line strengths: without it, floor divisions, columns, wall extensions and paper guides would become difficult to separate. ⁴

Significance and limitations

Freeplan's achievement is to preserve the intelligibility of architectural signs while loosening their expected assembly. A wall remains recognisable as a wall, yet its position no longer guarantees a room. A floor offers a field of occupation without containing every component. The title block and scale maintain the promise of exactness while the plan itself keeps that promise open to interpretation. The five inspected works make this tension visible in different representational styles. ^{1 2 3 4 5}

The code also deserves to be understood in its actual form. Its parsing and threshold details qualify easy claims about randomness, while its drawing order explains the convincing hierarchy of paper, floor, shadow and wall. Further work could compare the indexed program with a pinned on-chain retrieval, investigate the effect of seed parsing across the whole edition and examine the artist's later spatial models. The present reading treats the outputs as precise graphic propositions whose value lies in the relationships they invite the viewer to imagine. ⁷