The reviewed record of science sign in
Pith

arxiv: 2606.19609 · v1 · pith:BYCMTGQN · submitted 2026-06-17 · cs.HC · cs.GR

Building Drift: Documenting On-Site Construction Adaptations Across Material Lifecycles

Reviewed by Pith2026-06-26 19:01 UTCgrok-4.3pith:BYCMTGQNopen to challenge →

classification cs.HC cs.GR
keywords building driftreclaimed materialson-site adaptationscircular economyconstruction documentation3D Gaussian Splattingmaterial reusevideo documentation
0
0 comments X

The pith

Building drift is the collective deviation from digital models caused by on-site adaptations when using reclaimed materials, captured in a five-category taxonomy and documented through video and 3D Gaussian Splatting.

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The paper establishes that reclaimed materials introduce unpredictability requiring improvisation, and that the resulting physical deviations from planned models, called building drift, can be systematically classified. A case study of a reclaimed timber pavilion yields five categories of adaptations that occur across building lifetimes. The authors then implement a tool to record these changes in ways that preserve material histories for later stakeholders. If correct, this approach makes the knowledge embedded in improvised construction available for assessment, continuation, and reuse instead of being lost after each project.

Core claim

Through the ReShelter case study of a reclaimed timber pavilion built in the forest, the authors define building drift as the collective deviation of physical state from the digital model arising from on-site adaptations. They classify this drift into five categories—Tending the Site, Foraging for Fit, Interpreting the Material, Marking Measurements, and Coordinating Across Communities—and introduce Pentimento, a documentation system that combines video with 3D Gaussian Splatting to represent the adaptations spatially, temporally, and semantically relative to the original model.

What carries the argument

The five-category taxonomy of building drift, operationalized by the Pentimento tool that links video documentation and 3D Gaussian Splatting to the designed model.

If this is right

  • Collaborators, evaluators, and future inheritors receive the information needed to continue, assess, and reuse materials.
  • Computational tools become feasible that support the on-site improvisation required for reclaimed-material construction.
  • More sustainable cycles of recovery, repair, and reuse become possible because material histories remain accessible.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • The same recording approach could be tested on projects that reuse steel, concrete, or other non-timber materials to check whether the five categories still cover the observed drift.
  • Embedding the Pentimento output directly into building information models might allow designers to anticipate common drift patterns in future projects.
  • Longitudinal use of the tool across multiple buildings could generate data on which adaptation types most affect later reuse decisions.

Load-bearing premise

The adaptations seen in one reclaimed-timber pavilion project are representative of the challenges that arise across different reclaimed materials and construction settings.

What would settle it

Documentation of a second reclaimed-material project in which none of the five adaptation categories occur or in which the recorded deviations cannot be related back to the original model would falsify the taxonomy.

Figures

Figures reproduced from arXiv: 2606.19609 by Amritansh Kwatra, Jan H\"uls, Martin Tamke, Mette Ramsgaard Thomsen, Ritik Batra, Steven J. Jackson, Thijs Roumen, Tom Svilans.

Figure 1
Figure 1. Figure 1: Pentimento aligns the as-built structure as Gaussian splats with the as-designed model using overlaid on-site adaptations and comments. Abstract In a circular economy for construction, reclaimed materials carry prior lives of use and go on to have post-lives in future buildings. Yet working with such materials introduces unpredictability that requires on-site improvisation, making their reuse challenging t… view at source ↗
Figure 2
Figure 2. Figure 2: Aligned on-site adaptation in Pentimento [PITH_FULL_IMAGE:figures/full_fig_p002_2.png] view at source ↗
Figure 3
Figure 3. Figure 3: Examples of reclaimed material sources and storage for ReShelter. [PITH_FULL_IMAGE:figures/full_fig_p005_3.png] view at source ↗
Figure 4
Figure 4. Figure 4: ReShelter design and components. ReShelter's construction proceeded in two phases as shown in [PITH_FULL_IMAGE:figures/full_fig_p006_4.png] view at source ↗
Figure 5
Figure 5. Figure 5: ReShelter construction phases. 3.2 Building Drift Taxonomy Tending the Site ( [PITH_FULL_IMAGE:figures/full_fig_p006_5.png] view at source ↗
Figure 6
Figure 6. Figure 6: Examples of Tending the Site. Foraging for Fit ( [PITH_FULL_IMAGE:figures/full_fig_p007_6.png] view at source ↗
Figure 7
Figure 7. Figure 7: Examples of Foraging for Fit. Interpreting the Material ( [PITH_FULL_IMAGE:figures/full_fig_p007_7.png] view at source ↗
Figure 8
Figure 8. Figure 8: Examples of Interpreting the Material. Marking Measurements ( [PITH_FULL_IMAGE:figures/full_fig_p008_8.png] view at source ↗
Figure 9
Figure 9. Figure 9: Examples of Marking Measurements. Coordinating Across Communities ( [PITH_FULL_IMAGE:figures/full_fig_p008_9.png] view at source ↗
Figure 10
Figure 10. Figure 10: Examples of Coordinating Across Communities. [PITH_FULL_IMAGE:figures/full_fig_p009_10.png] view at source ↗
Figure 11
Figure 11. Figure 11: Pentimento interface features. To make on-site captured data legible, Pentimento represents sources of building drift across three dimensions: spatially, through alignment of the Gaussian splat with the original model; temporally, through a chronological record of when and under what conditions adaptations were made; and semantically, through transcribed voice recordings categorized into the sources of bu… view at source ↗
Figure 12
Figure 12. Figure 12: Pentimento workflow. 4. Results 4.1 Spatially Pentimento’s spatial grounding enabled stakeholders to locate and interpret sources of drift relative to the original model. In Figure 13a, a builder took a measurement of a bolt and documented it in place, anchored to the designed geometry. An inheritor reusing that timber beam would see not just the measurement but how it was measured, information invisible … view at source ↗
Figure 13
Figure 13. Figure 13: Examples of overlaid adaptation splats in Pentimento. [PITH_FULL_IMAGE:figures/full_fig_p011_13.png] view at source ↗
Figure 14
Figure 14. Figure 14: Pentimento timeline with example sources of drift. [PITH_FULL_IMAGE:figures/full_fig_p012_14.png] view at source ↗
Figure 15
Figure 15. Figure 15: Search and filter features of Pentimento. [PITH_FULL_IMAGE:figures/full_fig_p013_15.png] view at source ↗
read the original abstract

In a circular economy for construction, reclaimed materials carry prior lives of use and go on to have post-lives in future buildings. Yet working with such materials introduces unpredictability that requires on-site improvisation, making their reuse challenging to document and scale across building lifetimes. Without documentation, the on-site adaptations that make construction with reclaimed materials possible leave collaborators, evaluators, and inheritors without the information they need to continue, assess, and reuse materials. We call the collective deviation of the physical state from the digital model through these adaptations "building drift." Through a case study, ReShelter, a reclaimed timber pavilion constructed in the forest, we develop a taxonomy for building drift that characterizes the collective deviation across building lifetimes: Tending the Site, Foraging for Fit, Interpreting the Material, Marking Measurements, and Coordinating Across Communities. To put our taxonomy for building drift into practice, we present Pentimento, a documentation tool that leverages video documentation and 3D Gaussian Splatting to spatially, temporally, and semantically represent on-site adaptations in relation to the designed model. Pentimento enables each stakeholder to navigate material histories in ways that reduce barriers to material reuse. Together, these contributions open pathways towards computational tools that support the on-site improvisation essential to construction with reclaimed materials, enabling more sustainable cycles of recovery, repair, and reuse.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Referee Report

1 major / 1 minor

Summary. The paper introduces the concept of 'building drift' to describe collective on-site adaptations when using reclaimed materials in construction. Through a single case study of the ReShelter reclaimed-timber pavilion, it develops a five-category taxonomy (Tending the Site, Foraging for Fit, Interpreting the Material, Marking Measurements, Coordinating Across Communities) and presents the Pentimento tool, which uses video documentation and 3D Gaussian Splatting to spatially, temporally, and semantically link adaptations to the original digital model, with the goal of supporting material reuse in circular economies.

Significance. If the taxonomy and tool prove effective beyond the single case, the work could contribute to HCI and sustainable construction research by providing a structured way to document improvisation with reclaimed materials, potentially lowering barriers to reuse. The integration of 3D Gaussian Splatting for multi-stakeholder navigation of material histories is a concrete technical contribution that merits credit as an applied demonstration.

major comments (1)
  1. [Abstract] Abstract: The central claim that the taxonomy 'characterizes the collective deviation across building lifetimes' rests on adaptations observed in the single ReShelter case study; no additional cases, cross-material comparisons, or member-checking are described to establish representativeness, making the generalization load-bearing for the stated scope.
minor comments (1)
  1. [Abstract] The abstract and introduction could more explicitly delimit the taxonomy as an initial, case-derived framework rather than a comprehensive characterization.

Simulated Author's Rebuttal

1 responses · 0 unresolved

We thank the referee for their review and for identifying the scope of the generalization in the abstract. We address the comment below.

read point-by-point responses
  1. Referee: [Abstract] Abstract: The central claim that the taxonomy 'characterizes the collective deviation across building lifetimes' rests on adaptations observed in the single ReShelter case study; no additional cases, cross-material comparisons, or member-checking are described to establish representativeness, making the generalization load-bearing for the stated scope.

    Authors: We agree that the taxonomy is developed from a single case study (ReShelter) and that the abstract phrasing implies broader characterization without additional validation. In revision we will reword the abstract to state that the taxonomy is derived from adaptations observed in the ReShelter reclaimed-timber pavilion and characterizes the deviations documented in that project. We will also add an explicit limitations paragraph noting the single-case origin, the absence of cross-material or multi-site validation, and the exploratory status of the taxonomy, while preserving the contribution as an initial structured account of on-site adaptations with reclaimed materials. revision: yes

Circularity Check

0 steps flagged

No circularity; purely descriptive qualitative case study

full rationale

The paper develops a taxonomy of building drift categories directly from observations in a single ReShelter case study and proposes a documentation tool. No equations, fitted parameters, predictions, or self-citations appear in the provided text. The central claim is an empirical characterization emerging from the case rather than a derivation that reduces to its own inputs by construction. The work is self-contained as a descriptive contribution without any of the enumerated circular patterns.

Axiom & Free-Parameter Ledger

0 free parameters · 1 axioms · 1 invented entities

The paper's contributions rest on domain assumptions about construction practices and introduces a new conceptual entity without additional free parameters or formal axioms.

axioms (1)
  • domain assumption Working with reclaimed materials introduces unpredictability that requires on-site improvisation.
    This is presented as the core problem in the abstract.
invented entities (1)
  • building drift no independent evidence
    purpose: To describe the collective deviation of the physical state from the digital model due to on-site adaptations.
    Newly defined term in the paper to frame the research.

pith-pipeline@v0.9.1-grok · 5803 in / 1284 out tokens · 52891 ms · 2026-06-26T19:01:13.981265+00:00 · methodology

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Reference graph

Works this paper leans on

15 extracted references · 1 canonical work pages

  1. [1]

    Yet buildings made from reclaimed materials, like other forms of human endeavor, are rarely executed exactly to plan (Suchman 1987; Klemp et al

    Introduction Imagining a circular economy of construction materials requires recovering, repairing, and reusing materials across multiple building lifetimes. Yet buildings made from reclaimed materials, like other forms of human endeavor, are rarely executed exactly to plan (Suchman 1987; Klemp et al. 2008). Unlike standardized and pre-processed materials...

  2. [2]

    and makes them selectively legible to different viewers. This work sits at the intersection of computational design, human-computer interaction, and science and technology studies, demonstrating how 3D Gaussian Splatting (3DGS), voice transcription, and LLM-powered categorization can be combined to document building drift and therefore reduce barriers to ...

  3. [3]

    blueprint-first

    and aligns them with the original CAD model (Kwatra et al. 2025). Stakeholders separated by space and time can navigate construction project states and adaptations spatially, temporally, and semantically to further support the recovery, repair, and reuse of reclaimed materials across building lifetimes. Through this work, we make two central contributions...

  4. [4]

    bind[s] their own pathways or lines of becoming into the texture of material flows comprising the lifeworld

    State of the Art 2.1 Computational Models for Reclaimed Materials Computational models for architecture assume material homogeneity. Popular Computer-Aided Design (CAD) software such as Rhino3D or Autodesk Fusion abstract materials away as uniform, homogeneous surfaces. Abstract models reflect an implicit assumption that, for example, a timber beam is alw...

  5. [5]

    but assumes a perfect correspondence between model and reality and therefore does not capture the deviations that occur. Documenting drift therefore requires Building Drift 5 novel computational representations to capture and share the gaps between the assumed digital model and the physical as-built state. 3DGS offers an alternative representation for doc...

  6. [6]

    and architectural heritage preservation (Yu et al. 2025). Capturing improvisations requires a video recording using any camera (Batra et al

  7. [7]

    Existing splatting workflows explored a spatially grounded reconstruction that can be aligned with the original 3D model (Kwatra et al

    without needing site connectivity (critical for remote construction sites like forests). Existing splatting workflows explored a spatially grounded reconstruction that can be aligned with the original 3D model (Kwatra et al. 2025). We extend this method of geometric alignment with Pentimento, integrating on-site voice-recorded adaptations spatially, tempo...

  8. [8]

    Construction is rarely an isolated act; site owners, designers, builders, and inhabitants each shape what gets built and where, often in ways that emerge once construction begins

    describes drift arising from navigating the different knowledge, roles, and expectations of the communities surrounding a construction project. Construction is rarely an isolated act; site owners, designers, builders, and inhabitants each shape what gets built and where, often in ways that emerge once construction begins. These social negotiations are as ...

  9. [9]

    on the physical structure at locations corresponding to marked squares in a glTF representation of the model (exported by CAD tools such as Rhino). The backend detects these markers and calculates a weighted transform to align the splat to the model based on the number of markers detected and the diversity of viewing angles across frames. For temporal rep...

  10. [10]

    push back

    Discussion 5.1 Documenting on Site Deploying Pentimento for ReShelter revealed that the same heterogeneity that makes building drift inevitable also makes it difficult to document. ArUco markers often fell off timber surfaces due to wood shavings and weather, requiring constant reattachment. Capturing complete splats requires a steady loop around the stru...

  11. [11]

    Prototyping New Scenarios for Fabricating Structural Components from Reclaimed Timber

    "Prototyping New Scenarios for Fabricating Structural Components from Reclaimed Timber." In Fabricate 2024: Creating Resourceful Futures, 114–121. UCL Press. Journal Articles Alizadehsalehi, Sepehr, Ahmad Hadavi and Joseph Chuenhuei Huang

  12. [12]

    (De)composing Craft: An Elementary Grammar for Sharing Expertise in Craft Workflows

    "(De)composing Craft: An Elementary Grammar for Sharing Expertise in Craft Workflows." arXiv preprint arXiv:2506.10891. Building Drift 17 Gordon, M. C. and R. Vargas Calvo

  13. [13]

    SplatOverflow: Asynchronous Hardware Troubleshooting

    "SplatOverflow: Asynchronous Hardware Troubleshooting." In Proceedings of the 2025 CHI Conference on Human Factors in Computing Systems. New York: ACM. 1–16. Peng, Xue Bin, Marcin Andrychowicz, Wojciech Zaremba and Pieter Abbeel

  14. [14]

    Sim-to-Real Transfer of Robotic Control with Dynamics Randomization

    "Sim-to-Real Transfer of Robotic Control with Dynamics Randomization." In 2018 IEEE International Conference on Robotics and Automation (ICRA). IEEE. 3803–3810. Radford, Alec, Jong Wook Kim, Tao Xu, Greg Brockman, Christine McLeavey and Ilya Sutskever

  15. [15]

    Durable Data Models for Circular Building Practices in Bio-based Construction: The History Stack

    "Durable Data Models for Circular Building Practices in Bio-based Construction: The History Stack." In EC3 Conference 2025, vol