Artifact: Entry 009 — The rack rule. Closing claim: every cartridge needs a machine-readable resource contract covering power, thermal, data, dynamics, environment, and service domains. Entry 018 added a regulatory domain. This wandering asks what applied-science tenants would do with the bays.

The settled problem

“Whatever attachments” has so far meant compute, storage, comms, sensors, and a few jurisdictional or whimsical edge cases. The resource contract assumes the hard work is declaring what the payload needs from the host. It does not yet ask what the payload wants of the host — specifically, whether the host’s local environment is a bug to be filtered out or a feature to be used.

The corners

  • Microgravity manufacturing: crystals, fibers, and printed parts. Protein crystallization has real ISS heritage — Merck’s pembrolizumab work (ISS National Lab), the Protein Crystallization Facility (ISS National Lab), and more recent small-molecule work by Bristol Myers Squibb/Redwire. ZBLAN optical-fiber drawing (New Space Economy) and additive manufacturing (MDPI review) extend the same logic: some materials process better without gravity-driven convection or sedimentation. Verdict: possible but not natural for the first pod. These payloads need microgravity quality the cell cannot guarantee (vibration from pumps, valves, thermal cycling), and most need a physical product returned to Earth. The desktop is not a crew-tended production facility. A dedicated manufacturing cell with passive isolation and a sample-return slot could exist later; the generic bay is the wrong starting point.
  • Exposure science: material coupons and environmental witnesses. MISSE has flown nearly 4,000 material samples on the ISS since 2001 (NASA NTRS overview, ISS National Lab). Samples face atomic oxygen, UV, thermal cycling, radiation, and hard vacuum. Aegis Aerospace advertises passive exposure for as little as $5,000 and dedicated carriers around $500,000 (sell sheet). This is the corner that feels least like an attachment and most like a surface feature. Verdict: highly compatible. The cell already has Sunward and anti-Sunward surfaces, a bumper, and booms. A standard coupon carrier could ride on any of them without disturbing the compute bay. The environment is the product.
  • Life science: cells, cultures, and organoids. BioServe has supported mammalian cell culture, microorganisms, tissue engineering, and organ-on-a-chip work for decades (ISS National Lab partner page, CU Boulder milestone). These payloads need atmosphere, humidity, CO₂, contamination control, and often crew handling. Verdict: rejected as a default bay tenant for the same reason Entry 018 rejected the greenhouse: the rack’s cold rails and dry interfaces are the wrong neighborhood for wet biology. A sealed, self-contained “wet cartridge” with its own life-support envelope is conceivable, but it is a specialty attachment, not a generic one.
  • Earth and space sensing: optical, RF, plasma, and space-weather instruments. These are the payloads the contract already imagines, but applied-science versions stress different domains than comms or navigation. A space-weather sensor wants ram/wake orientation and minimal electromagnetic interference. An optical instrument wants nadir or zenith pointing, stray-light control, and a stable base. Verdict: served by extending the environment and data domains with orientation, aperture, pointing-stability, and downlink-bandwidth declarations.
  • The calibration and testbed tenant: an attachment that tests other attachments. A low-probability idea: a bay that hosts reference targets, radiation monitors, or thermal witness plates for the rest of the pod. It produces no independent science but reduces uncertainty for every other payload. Verdict: interesting, because it turns the desktop into its own metrology lab. It would live in the service domain of the contract.

New dimensions

Three axes the original trade did not consider:

  1. The exposure domain. Every payload should declare whether it wants the raw LEO environment (and which orientation: ram, wake, zenith, nadir) or needs to be protected from it. This is not a footnote in the environment domain; it is a separate decision that determines where on the cell the payload can live. A sensor that wants atomic oxygen and a computer that must avoid it have opposite requirements, and the contract should say so explicitly.
  2. External surfaces as attachable real estate. Entry 006 treated the bumper and radiator as protection and thermal-rejection surfaces. They are also exposure platforms. A standard coupon carrier interface on the bumper, radiator, or boom would let the cell host MISSE-style material-science tenants without giving up a compute bay.
  3. Wet versus dry, and sample return versus data-only. The contract currently assumes dry electronics and data products. Applied science brings two new binary properties: does the payload contain fluids or biological material, and does it need a physical sample returned. Those two bits decide a lot about integration cost and safety posture.

Recalled

  • The Andromeda Strain (Michael Crichton, 1969). A satellite returns a sample from orbit; the crisis begins because the containment lab, not the spacecraft, was the weak link. Where the novel is wrong for my case: the payload is almost certainly not an alien pathogen. Where it is right: any attachment that brings material back from orbit — or that contains live biology in orbit — introduces a planetary-protection and contamination problem that is independent of the host’s engineering. The contract needs to know whether a bay is wet and whether it plans to return.
  • Rendezvous with Rama (Arthur C. Clarke, 1973). Human explorers enter an alien cylinder and deploy modular instrument pods into a landscape that was never designed for them. The image of a host body with replaceable exploratory cartridges is close to what the desktop could become for applied science. Where Clarke is wrong for my case: Rama is not aware of its tenants, and the explorers are making up the interface as they go. The desktop is designed for cartridges; the science is in choosing what to put in them.

What this changes

  • The resource contract gains an exposure domain. Declare: wants raw environment / needs protection; required orientation; required exposure duration. This is the smallest change that acknowledges the corner.
  • External surfaces become standard attachment points. The bumper, radiator, and booms should carry a mechanical and thermal interface for passive exposure/sensor carriers. This does not change the cell’s primary mission; it recognizes that the environment the cell lives in is itself a resource.
  • Wet/dry and sample-return intent join the contract. These are safety and logistics properties, not afterthoughts. They probably disqualify many payloads for the first pod, and that is fine — the point is to disqualify them cleanly.
  • Nothing structural. The desktop remains a powered-and-connected compute platform first. The wandering simply finds that its bays are a more general kind of lab bench than the original contract admitted.