Asterisms

The work and its context

Asterisms offers a sky whose patterns will not stay still. Points drift across a dark square, joining nearby points with luminous lines and releasing those connections as they separate. A triangle becomes an open chain; a crowded knot loosens into several fragments. The eye keeps proposing figures, but the drawing changes before any one interpretation can become final. ^{1 2 3 4}

Falko, also identified by Art Blocks as Pedro Falco, released the 100-work Factory series on 20 April 2021. It is project 47 on its Ethereum contract. The project description invites viewers to add stars, supply a name and save an image. Its subject is therefore the making of a pattern as much as the appearance of a star field. The Falko dossier gives the wider artist context. ⁵

The work begins moving without intervention. Interaction changes an already active system rather than supplying the movement it lacks. This distinction is visible in the four original renderers examined here: connections form and disappear while the viewer simply watches. A saved frame can preserve one arrangement, but cannot contain the experience of waiting for a shape to emerge. ^{1 2 3 4}

Conception and development

The official artist profile describes a practice concerned with data, real-time processes and interactive experiences. The project statement makes the viewer's role explicit by treating imagined stellar patterns as something people create. These sources support a connection between interaction and interpretation; they do not establish a detailed chronology of prototypes or earlier versions. No separate development interview was recovered for this essay. ⁵

The title's invitation becomes concrete through the interface. The initial label identifies the hash. Typing replaces that identifier with a viewer-supplied title. Holding the pointer adds points to the field. The code supplies movement and rules for connection, while the viewer can alter the population and choose a name for the result. The name does not direct the geometry: a word typed above the field does not instruct the program to draw that object. ⁶

That separation leaves room for interpretation. Naming a temporary group makes a private association public within the displayed image, but the moving points need not continue to resemble it. The work can outlast its own caption. This is an inference from the relationship between the inspected title mechanism and the changing forms, rather than a claim that every viewer will approach the invitation in the same way. ^{6 1}

How the work was constructed

The complete indexed program contains 6,293 characters and uses p5.js 1.0.0. A numeric value derived from the token hash seeds the library's random generator. Setup chooses an initial population between 120 and 199 points, along with size ranges, speed ranges, line characteristics and colours. The canvas is square and scales to the smaller viewport dimension. The requested frame rate is twenty frames per second; actual browser performance can differ. ^{5 6}

Each point receives a position and two velocity components. Those components advance its horizontal and vertical coordinates on successive frames. When a point crosses a boundary, the corresponding velocity changes sign. The program does not calculate gravitational attraction between stars. Its celestial appearance arises from moving points, proximity tests, glow and naming, rather than an astronomical model. ⁶

Connections depend on distance. The program examines pairs and draws when their separation is below a selected threshold, subject to a condition placing the second point within a central circle. That condition is ordered by point index, so it should not be simplified into a perfectly symmetric rule that treats both endpoints identically. The network is a particular drawing procedure, not a generic diagram of every nearby pair. ⁶

Drawing order introduces another qualification. After the point pass disables strokes, the line loop issues its first line command before setting the stroke and width for subsequent commands. A mathematical account of all eligible pairs would therefore not necessarily describe every line that appears. Several calculations also produce values that are not subsequently applied. Reading the instructions in sequence matters more than assigning visual effects to suggestive variable names. ⁶

The field is cleared and redrawn each frame. Screen blending and a shadow blur give bright marks their luminous appearance against a selected dark background. The blur changes in steps with viewport size. Points are drawn before their positions advance, which makes the displayed state a succession of samples from the moving system. The light is an effect of compositing and blur, not evidence of a physically simulated glow. ⁶

Additional branches place coloured bodies with names including Aldebaran, Mars, Moon and Sirio. Their locations are selected within the canvas rather than taken from an observed sky. One branch varies a body's opacity during drawing; another introduces shooting points. These names help establish a celestial vocabulary, but should not be treated as coordinates, astronomical classifications or a realistic relation of scale. This review does not calculate how often every branch occurs across the edition. ⁶

The shooting-point routine has a consequential side effect: it assigns new values to the global colour channels also used by the ordinary field. It can therefore affect more than the small objects it draws. Its removal test also mixes an updated horizontal coordinate with the original vertical argument. These are properties of the inspected implementation. A preservation effort should retain that behaviour before considering whether a later revision would be desirable. ⁶

Holding the mouse adds a point on each drawing cycle until the count passes the script's limit condition. Since the comparison includes equality, the nominal limit of 450 can admit a final point beyond it. The points inherit fresh speeds and sizes from the seeded sequence. Adding several together can make a bright, crowded region whose members then move apart. The held-work test produced such a cluster near the pointer. ^{6 7}

The mouse wheel changes a scale factor around the pointer. The added-point coordinates, however, are taken directly from the mouse rather than transformed back through that zoom. A description of interaction should not promise that placement remains geometrically identical under every scale. The title is drawn before this transformation, keeping it outside the zoomed star field. ⁶

Typing, Backspace and Delete edit the title. Enter calls saveCanvas with a PNG extension and a filename derived from the text. The p5 reference describes this operation as saving the current canvas image. The inspected artist program contains no transaction submission or persistent storage for the viewer's edits. Exporting a picture is therefore different from changing the token or preserving an editable session. ^{6 8}

The script requests Montserrat by name but does not load a font file itself. A future reproduction needs to check the available font as well as the library and token input. For this essay, the original renderers were observed in motion; the held work was also tested by adding points and replacing the title. The export routine was inspected in code, but no saved-file fidelity test is claimed. ^{6 7}

The mint

The collection page gives 20 April 2021 as the release date. The index records the first token on 15 April at 13:01:04 and activation on 19 April at 17:13:12. The second mint and scheduled start are recorded on 20 April at 17:15:38. Completion is recorded on 27 June at 22:12:15. These fields identify different events, with full timestamps retained in the evidence record. ⁵

The present record reports 100 invocations, a maximum of 100 and a price setting of 0.1 ETH. The price setting is not independently established here as the original public price. Nor does the current maximum alone demonstrate the announced edition at every earlier point. Artist, reserved and public mint classifications remain unresolved; no precise public sellout duration is inferred from the index timestamps. ⁵

The first four examined editions are not four stages in a single animation. Each has its own hash and mint record, while each rendering then develops through time. This distinction keeps edition history separate from the changing appearance of a work. A screenshot's visible configuration does not identify the moment when its token was minted. ^{6 1 2 3 4}

Copyright and licensing

The current artwork label is NFT License. The record supplies no terms link, and this research did not recover an operative text or version tied conclusively to the project. The label is preserved exactly and classified as unresolved. It should not be expanded into a Creative Commons grant or silently equated with another similarly named licence. ⁵

The p5.js 1.0.0 repository includes LGPL version 2.1. That dependency permission is separate from the artwork label and from the artist's program, for which no distinct reuse grant was established. AB5D's CC0 dedication applies to the original essay prose. Displaying the original renderer does not make the artwork or its code part of that dedication. ^{6 9}

Close reading of four outputs

The comparison uses editions 0, 1, 10 and 20, selected in advance. Their original live renderers were each inspected twice, with movement and changed connections visible between observations. The identifiers are 47000000, 47000001, 47000010 and 47000020 on Ethereum contract 0xa7d8d9ef8d8ce8992df33d8b8cf4aebabd5bd270. These are short observations of four works, not exhaustive accounts of their evolving states or a representative survey. ^{1 2 3 4}

Number 0 presents fine pale pink-white connections on an almost black olive field. In the first observation, a broad connected group occupies the upper middle while smaller fragments remain around it. Some are triangles; others are open chains or isolated points. A small yellow glow sits low in the field. In the next view, several triangular relations have shifted or opened, making the larger arrangement feel less like a finished emblem than a temporary negotiation between neighbours. ¹

Number 1 has brighter, cooler lines and comparatively sharp white points. A concentrated knot near the upper-left side of the central area anchors several longer chains. Elsewhere, single segments and angular fragments leave wide spaces of darkness. The second observation changes the knot's internal divisions while preserving the broad impression of a crowded centre of activity surrounded by sparse groups. Attention can follow a local shape or move outward to the intervals between shapes. ²

Number 10 sets dimmer pink connections against a dark teal field. Bright points remain legible where the thin lines almost disappear. A dense cluster left of centre breaks into a looser structure in the second observation, while a luminous white body below and to the right offers a comparatively stable visual reference. The difference between a bright point and a faint link becomes especially important: the eye can recognise the stars before it decides which network to follow. ³

Number 20 is more assertive in its points and pale yellow-green lines. Larger lights punctuate the field, and a busy lower region contains several connected polygons. Between observations, some of those polygons open and others acquire new diagonals. Their changing density gives the lower area a sense of activity without requiring a single object to travel across the whole image. The contrast with #10 shows how brightness and point size affect the apparent complexity of comparable connecting rules. ⁴

AB5D's #23 was separately inspected as the lead example. It has a dark purple field, white points and delicate pale connections. Holding the pointer added a bright group near the centre; typing replaced the identifier with a letter. Those reversible actions confirmed the basic interaction in the original renderer. The essay's embedded presentation begins again from the original work, rather than preserving that test state. ⁷

Significance and limitations

Asterisms makes recognition temporary. A viewer may discover a figure, name it and then watch the arrangement move away from the resemblance that prompted the name. The system's achievement lies in making that instability easy to see through a small vocabulary of points and links. Interaction adds another level: the person interpreting the pattern can also change the material from which the next pattern forms. ^{6 1 7}

A still image can document a moment of this process, but the moving field and the ability to intervene are central to the work examined here. Its preservation should retain both, while distinguishing the initial token-derived state from user additions and exported frames. That distinction allows the project to remain an experience of making and naming, rather than only a collection of star-like pictures. ^{6 8}

Unresolved research questions:

• Recover a first-person development account and any earlier prototypes.
• Establish launch-era price, supply announcements and mint classifications.
• Resolve the exact NFT License terms and any separate artist-code grant.
• Compare the indexed program with a chain retrieval pinned to a block.
• Test PNG export, font availability, zoomed point placement and touch behaviour across supported environments.
• Survey longer animation intervals without treating a momentary configuration as a fixed feature.