Fernweh

The work and its context

Fernweh makes a woodland out of branching lines and accumulated leaves. Its wide field places the viewer close to the trunks, with crowns rising into and beyond the upper edge. There is usually no distant landmark to organize the scene. Instead, the eye moves between neighbouring stems, patches of foliage and openings onto a softly varied ground. In the five official previews examined here, the forest ranges from almost dissolved marks to a dense screen of dark branches. ¹²³⁴⁵

The project belongs to Art Blocks Factory and was released on 4 February 2022. Its current edition contains 121 works. The artist's description links the title to longing for unfamiliar places and frames the project as an attempt to connect recognizable landscape with generative procedure. That stated ambition is useful without requiring acceptance of every generalization the description makes about representational and abstract art. ⁶

Fernweh also has duration. The program builds three versions, named Original, Etched and Abstract, and provides both timed transitions and a mouse control. AB5D holds #40, whose live rendering and click transition were inspected separately from its official preview. The stills support close reading of particular recorded appearances; they do not exhaust what the originals can display. ⁶⁷⁵

Conception and development

The project statement gives a clear account of the desired encounter: procedural images that retain the familiarity and irregularity of a landscape. It associates that irregularity with the accidents of handmade painting. The code makes this ambition more concrete. Rather than drawing a photographically accurate tree, it combines recursively divided branches, reusable leaf shapes, bark marks and a layered background. Their relationships produce the recognizable scene. ⁶⁷

The program's opening comments are an unusually explicit record of technical debts. They credit Robert D'Arcy's OpenProcessing sketch for segmented-bark mathematics and identify an earlier tutorial behind that adaptation. They also name Kjetil Golid's watercolour techniques as inspiration for the background, and thank Alexandre Bizord for corrections, review and optimization. These are the artist's own credits in the inspected program. The cited OpenProcessing page could not be retrieved in this review, and the earlier tutorial's domain did not resolve, so their exact historical versions remain unverified. ⁷

The comments distinguish those contributions from features the artist describes as original, including recursive leafing, veins, connectors, palettes, bark perforations and multiple render treatments. This is a useful account of how the artist understood the work's construction. It should not be converted into a complete independent audit of originality or permissions. A public code credit establishes an acknowledged relationship; it does not resolve every question about the terms of reuse. ⁷

The oliveskin dossier places Fernweh within a broader practice. Here the most productive historical connection is between the stated desire for a recognizable landscape and the program's repeated acts of subdivision. The resulting trees remain legible even when their surfaces become misty or emphatically textured. Recognition survives substantial changes in finish. ⁷¹²

How the work was constructed

The inspected artist script is 19,534 bytes and uses p5.js 1.0.0. The complete program was read, and its retrieval query and digest are recorded with this essay. This is inspection of the indexed script, not a comparison against a contract retrieval at a pinned block. The working image space is 1,920 by 926 units; the display is scaled to fit the browser while preserving that broad proportion. ⁶⁷

A custom random generator takes the two halves of the token hash, initializes two streams, warms them up and alternates between them. Weighted tables select season, leaf shape and size, canopy density, tree-height range, bark treatment and transition timing. These choices are not all independent visual decorations. A small leaf repeated across many tips creates a different density from a large leaf, while the permitted tree heights determine how much branching occurs above the frame. ⁷

The background is made separately from the trees. A row of points is repeatedly subdivided, and its intermediate positions are displaced. Copies of that irregular boundary receive further perturbations, then are drawn as translucent filled regions extending downward. Many such layers produce a mottled field with soft changes of colour. The background's apparent wash comes from accumulated polygon layers rather than a simulation of water soaking through paper. ⁷

Trees begin near the lower edge at selected horizontal positions. Each trunk is built from smaller segments whose directions and lengths vary. At successive branch levels, length is reduced and width narrows. Longer branches split into two; shorter ones fan into additional directions before reaching tips. This creates a hierarchy of thick supports and increasingly fine divisions. It is a drawing procedure for tree-like structures, with no inspected model of species growth, light competition or wind. ⁷

The trunk segments receive shading through narrow strokes across their width. Bark patterns add another scale of detail: spots, short horizontal marks, narrow furrows or nested rings. The Wild option chooses among several treatments. At ordinary viewing distance these marks can merge into a sense of roughness. Closer inspection reveals them as repeated graphic elements whose density follows the changing width of the branch. ⁷

Leaves are prepared in a small graphics buffer and reused at terminal positions. The available forms include triangles, ellipses and a diamond-shaped construction, with seasonal choices for boundaries, veins and connectors. Their size changes how the crown is read. Tiny leaves become atmospheric granulation; larger ones remain discrete shapes even when packed into a continuous mass. The contrast between #1 and #10 makes this consequence particularly visible. ⁷²³

The three render modes alter the treatment of this vocabulary. Etched introduces an offset shadow; Abstract changes trunk visibility and shadow treatment; Original draws an additional inset leaf form. The program stores these versions in separate graphics buffers. It resets the custom generator for each version, coordinating many choices across the three. That coordination should not be described as proof of exactly identical branch geometry. ⁷

One branch-angle expression calls p5's own random function inside the custom random helper. No call seeding that separate function appears in the inspected artist script. The version-specific p5 source shows that an unseeded call uses the browser's random generator. This is a concrete limit on claims of complete hash determinism: the custom stream is repeatable, but it is not the only source of variation in the drawing. A systematic comparison across reloads remains to be done. ⁷⁸

Time is handled through elapsed milliseconds after loading. Once the versions have been generated, the program waits for a selected interval and blends between buffers during the final second. Intervals range from hours to days and longer periods. They are session timers, not an inspected lookup of the calendar date or a synchronized blockchain clock. Reloading starts a new execution, which matters when describing a work that can appear still during an ordinary visit. ⁷

A mouse press can advance the timer after generation has finished, allowing the viewer to move through the versions. The S key has a PNG-saving handler. The collection original was slow to prepare in the test browser, but #40 eventually rendered, and clicking changed its surface from an airy green grove to a darker, more strongly articulated state. That verifies a short interaction, not a full unattended day-long transition. No audio implementation was found. ⁷⁵

The mint

The index records the first token on 25 January 2022. Art Blocks gives 4 February as the release date; the second indexed token was minted that day at 17:02:06, shortly after the scheduled 17:00 start. Activation is recorded separately on 27 January. These fields identify different events and are retained separately in the evidence. The first token's date should not silently replace the public release date. ⁶

Completion is recorded on 14 May 2022, and the present invocation count and maximum both read 121. The current price setting is 0.13 ETH. These observations do not establish the originally announced maximum, opening price, any later reductions or the allocation of early tokens. In particular, a current completed status cannot reconstruct how the edition reached its present size. ⁶

The interval between the first public-facing date and the completion field is therefore reported as chronology rather than labelled an exact sellout duration. Launch announcements, changes to project configuration and transaction classifications would be needed for a fuller sale history. The project's present supply is secure as an index observation; the path to that supply is less fully documented here. ⁶

Copyright and licensing

The artwork label is recorded as CC BY-NC SA 4.0, with no terms URL in the project record. The corresponding Creative Commons Attribution-NonCommercial-ShareAlike 4.0 deed is linked separately. It describes attribution, noncommercial use and the same-licence condition for shared adaptations. The raw metadata spelling is retained in the source record rather than silently rewritten as a historical document. ⁶⁹

The p5.js 1.0.0 repository carries the LGPL 2.1 text. That library licence is distinct from the artwork label and from any separate permissions for the artist's complete script. The code credits discussed above are preserved as evidence, but this review did not establish a comprehensive artist-code grant or verify the exact terms of every acknowledged contribution. AB5D's CC0 dedication applies to the original essay prose. ⁷¹⁰

Close reading of five outputs

The visual sample consists of official previews for editions 0, 1, 10 and 20, selected as an early group, plus AB5D's #40. The latter's ownership was checked against all three linked AB5D wallets. Each preview is a recorded still from a multi-state work; the following comparisons do not claim that it is the only appearance of that token or that five examples represent the edition's statistical distribution. ⁶¹²³⁴⁵

In #0, tiny yellow marks gather into broken crowns against a purple and dusty-rose field. The trunks are faint, almost absorbed into the background. Their vertical traces become clearer only after the brighter foliage has established where to look. The woodland seems to arrive through a veil. This is a useful example of how little structural information is needed for a landscape to remain recognizable. ¹

The broad gaps in #0 are as important as its marks. Near the lower half, long stretches of colour remain open between widely separated trunks. Above, yellow clusters form interrupted arcs rather than a solid roof. The image offers a rhythm of appearance and disappearance: foliage asserts the trees, while the softened supports let them recede. Its space feels less like a measured grove than a remembered one. ¹

In #1, dark grey trunks and angular branches stand firmly against pale cyan and grey. Small purple leaves accumulate into fine, dense crowns. Intersections between branches are conspicuous, and trunks of different widths create layers of apparent distance. There is no need for a horizon line to establish depth; overlap and scale perform that work within the crowded foreground. ²

The purple foliage does not simply fill the spaces above the trunks. It catches in bands and clusters, leaving narrow windows onto the pale field. At the top edge, branches continue beyond the frame, suggesting that the view is a fragment of a larger stand. The lower trunks remain comparatively open and upright, giving the dense upper network a clear structural base. ²

In #10, large brown leaves gather into heavy horizontal crowns. Their individual outlines remain visible, making the canopy more tactile than the powdery purple of #1. Spotted dark trunks rise through a cool teal background. The warmer leaf masses appear to sit in front of that cool field, although the depth is constructed entirely through colour, overlap and relative scale. ³

In #20, gold foliage spreads across most of the upper field, with irregular openings resembling arches between the crowns. Long grey trunks carry that dense canopy over a quieter lower zone. A pale violet background and muted grey passages keep the gold from becoming uniformly brilliant. The eye follows openings between leaves as readily as it follows the branches that support them. ⁴

AB5D's #40 is more open. Fine green leaves hang from an extensive network of narrow branches, with pale grey, pink and cyan in the preview's ground. The crowns break into small floating islands rather than forming a continuous cover. Individual diagonals remain easy to follow from trunk to tip, making the recursive structure unusually legible without exposing it as a diagram. ⁵

The live #40 developed a related grove over a cooler blue-green wash. A tested click then brought stronger dark definition to foliage and bark. The comparison demonstrates why a preview is a useful record but an incomplete account of the encounter. Surface treatment changes the apparent weight of the trees, while their branching vocabulary continues to hold the image together. ⁵

Significance and limitations

Fernweh's interest lies in the relationship between procedural regularity and an immediately familiar subject. The trees share a repeatable grammar of subdivision, yet differences in canopy scale, bark, palette and finish make the five inspected records feel spatially distinct. A sparse, veiled stand and a densely articulated grove can emerge from the same family of drawing operations. ⁷¹²⁵

The multiple views add another kind of comparison. Instead of treating one finished screenshot as the whole work, the program allows a scene to acquire different degrees of surface presence. Its long timers also make the duration of a visit part of what the viewer encounters. That possibility is grounded in the code and the tested click interaction; uninterrupted transitions across the longest intervals remain unobserved. ⁷⁵

Further work should establish the original edition and pricing announcements, recover the acknowledged source versions and permissions, compare the indexed script with a pinned chain retrieval, and measure the effects of the separate unseeded random call. Those are specific preservation and historical questions. They leave room for a close account of what already works: a forest built through layered rules, whose apparent atmosphere changes with the way those rules become visible. ⁶⁷⁸