Research bibliography
The published research, codes, standards, and books behind the directive library. We do not use any underlying research datasets in our app — see the licensing principle for the full statement.
Read depth: 4 abstract verified · 7 primary source we own
International Residential Code for One- and Two-Family Dwellings (IRC), 2024 Edition
International Code Council (2024). "International Residential Code for One- and Two-Family Dwellings (IRC), 2024 Edition". ICC. https://codes.iccsafe.org/content/IRC2024.
Abstract
The International Residential Code (IRC) is the model building code for one- and two-family dwellings adopted (with amendments) by most US states and counties. It governs life-safety, structural, plumbing, electrical, and mechanical requirements for residential construction. The 2024 edition refines egress, accessibility, and energy provisions over the 2021 edition.
Why we read it: Every directive tagged type='code' in our library cites a specific IRC section. The IRC is the single most consulted source in the directive library. The paywalled distribution means the user-facing /library detail page surfaces a 'Buy from ICC' link rather than reproducing the section text.
Licensing: The IRC is a copyrighted publication of the ICC. We cite section numbers (R304.1, R310.2.1, etc.) and paraphrase the requirements in our own words. We do not reproduce IRC text beyond fair-use citation. The directive's user-facing UI links to ICC's online code library where the user can purchase access to the full source text.
4 findings extracted.
Bathroom Planning Guidelines with Access Standards
National Kitchen & Bath Association (2024). "Bathroom Planning Guidelines with Access Standards". NKBA. https://nkba.org/professional-resources/kitchen-bath-planning-guidelines/.
Abstract
The NKBA Bathroom Planning Guidelines are 27 numbered design recommendations covering doorway clearances, fixture placement, vanity centerlines, shower/tub access, ventilation, lighting, and storage for residential bathrooms. The current edition aligns with ANSI A117.1 accessibility standards.
Why we read it: Every directive in our library that touches the bathroom — entry doorway, toilet centerline + front clearance, vanity-to-door clearance, tub-wall enclosure, shower clearance — cites the corresponding NKBA bathroom guideline.
Licensing: The NKBA guidelines are copyrighted by NKBA. We cite by guideline number and paraphrase the dimensional requirements; we don't reproduce the text. The directive user-facing UI surfaces a 'Buy from NKBA' link to the source publication.
3 findings extracted.
Kitchen Planning Guidelines with Access Standards
National Kitchen & Bath Association (2024). "Kitchen Planning Guidelines with Access Standards". NKBA. https://nkba.org/professional-resources/kitchen-bath-planning-guidelines/.
Abstract
The NKBA Kitchen Planning Guidelines are a set of 31 numbered design recommendations covering work zones, clearances, lighting, ventilation, storage, and accessibility for residential kitchens. The current edition incorporates ANSI A117.1 accessibility standards and is widely adopted as the de-facto standard for professional kitchen design in North America.
Why we read it: Every directive in our library tagged type='standard' that touches the kitchen — work-aisle widths, range landing area, refrigerator landing, work triangle, island clearance — cites a specific NKBA guideline number. Together with the IRC, the NKBA guidelines are the second-most-consulted source in the library.
Licensing: The NKBA guidelines are copyrighted by NKBA. We cite by guideline number and paraphrase the dimensional requirements; we don't reproduce the text. The directive user-facing UI surfaces a 'Buy from NKBA' link to the source publication.
4 findings extracted.
FloorPlanCAD: A Large-Scale CAD Drawing Dataset for Panoptic Symbol Spotting
Zhiwen Fan, Lingjie Zhu, Honghua Li, et al. (2021). "FloorPlanCAD: A Large-Scale CAD Drawing Dataset for Panoptic Symbol Spotting". ICCV 2021 (arXiv cs.CV). https://arxiv.org/abs/2105.07147.
Abstract
Access to large and diverse computer-aided design (CAD) drawings is critical for developing symbol spotting algorithms. In this paper, we present FloorPlanCAD, a large-scale real-world CAD drawing dataset containing over 10,000 floor plans, ranging from residential to commercial buildings. CAD drawings in the dataset are all represented as vector graphics, which enable us to provide line-grained annotations of 30 object categories. Equipped by such annotations, we introduce the task of panoptic symbol spotting, which requires to spot not only instances of countable things, but also the semantic of uncountable stuff. Aiming to solve this task, we propose a novel method by combining Graph Convolutional Networks (GCNs) with Convolutional Neural Networks (CNNs), which captures both non-Euclidean and Euclidean features and can be trained end-to-end. The proposed CNN-GCN method achieved state-of-the-art (SOTA) performance on the task of semantic symbol spotting, and help us build a baseline network for the panoptic symbol spotting task.
Why we read it: FloorPlanCAD's 30-category symbol vocabulary is a touchstone for our renderer's fixture set. The paper's separation of 'countable things' (toilets, sinks, ranges) and 'uncountable stuff' (walls, hallway runs) maps cleanly to our PositionedRoom + Fixture distinction in the spec. We use it as a reference for which symbols are universally recognized in residential CAD; we don't reproduce any of the actual drawings.
Licensing: FloorPlanCAD is released under CC BY-NC. We do NOT download or use the dataset, and we do not reproduce any of its CAD drawings or annotations. We cite the paper for its taxonomy and the panoptic-symbol-spotting framing only.
1 finding extracted.
Graph2Plan: Learning Floorplan Generation from Layout Graphs
Ruizhen Hu, Zeyu Huang, Yuhan Tang, et al. (2020). "Graph2Plan: Learning Floorplan Generation from Layout Graphs". ACM Transactions on Graphics (SIGGRAPH 2020). https://arxiv.org/abs/2004.13204.
Abstract
We introduce a learning framework for automated floorplan generation which combines generative modeling using deep neural networks and user-in-the-loop designs to enable human users to provide sparse design constraints. Such constraints are represented by a layout graph. The core component of our learning framework is a deep neural network, Graph2Plan, which converts a layout graph, along with a building boundary, into a floorplan that fulfills both the layout and boundary constraints. Given an input building boundary, we allow a user to specify room counts and other layout constraints, which are used to retrieve a set of floorplans, with their associated layout graphs, from a database. For each retrieved layout graph, along with the input boundary, Graph2Plan first generates a corresponding raster floorplan image, and then a refined set of boxes representing the rooms. Graph2Plan is trained on RPLAN, a large-scale dataset consisting of 80K annotated floorplans. The network is mainly based on convolutional processing over both the layout graph, via a graph neural network (GNN), and the input building boundary, as well as the raster floorplan images, via conventional image convolution.
Why we read it: Graph2Plan formalizes the 'layout graph' as the input to floor-plan generation: nodes are rooms, edges are required adjacencies. That maps 1:1 to our AdjacencyConstraint + RoomRequirement model in lib/layout/types.ts — independent confirmation that the graph-shaped problem statement is the right one. The retrieve-and-adjust paradigm (find similar plans in a database, refine to fit constraints) is a useful frame for understanding why our parametric template library works while pure generative approaches overfit.
Licensing: RPLAN, the underlying dataset, is released for research use only and is incompatible with our commercial license. We do NOT download or use RPLAN. We cite Graph2Plan for its problem framing (layout graph + boundary → floor plan) and its retrieve-and-adjust methodology only.
2 findings extracted.
House-GAN: Relational Generative Adversarial Networks for Graph-constrained House Layout Generation
Nelson Nauata, Kai-Hung Chang, Chin-Yi Cheng, Greg Mori, and Yasutaka Furukawa (2020). "House-GAN: Relational Generative Adversarial Networks for Graph-constrained House Layout Generation". ECCV 2020 (arXiv cs.CV). https://arxiv.org/abs/2003.06988.
Abstract
This paper proposes a novel graph-constrained generative adversarial network, whose generator and discriminator are built upon relational architecture. The main idea is to encode the constraint into the graph structure of its relational networks. We have demonstrated the proposed architecture for a new house layout generation problem, whose task is to take an architectural constraint as a graph (i.e., the number and types of rooms with their spatial adjacency) and produce a set of axis-aligned bounding boxes of rooms. We measure the quality of generated house layouts with the three metrics: the realism, the diversity, and the compatibility with the input graph constraint. Our qualitative and quantitative evaluations over 117,000 real floorplan images demonstrate that the proposed approach outperforms existing methods and baselines. We will publicly share all our code and data.
Why we read it: House-GAN crystallizes three quality metrics for evaluating generated layouts — realism, diversity, and graph-constraint compatibility. Those exact axes inform our solver's evaluation methodology: hard constraints map to compatibility (does the layout satisfy what the brief asks for?), the soft-constraint score corresponds to realism, and the persona-weight system addresses diversity (different architects produce different solutions for the same brief). The paper's framing helps us articulate why a single objective function isn't enough.
Licensing: The 117K-image LIFULL HOMES dataset is released only to academic researchers via a strict institutional access process. We do NOT have access to or use LIFULL HOMES. We cite House-GAN for its evaluation metrics framework only.
2 findings extracted.
CubiCasa5K: A Dataset and an Improved Multi-Task Model for Floorplan Image Analysis
Ahti Kalervo, Juha Ylioinas, Markus Häikiö, Antti Karhu, and Juho Kannala (2019). "CubiCasa5K: A Dataset and an Improved Multi-Task Model for Floorplan Image Analysis". arXiv (cs.CV). https://arxiv.org/abs/1904.01920.
Abstract
Better understanding and modelling of building interiors and the emergence of more impressive AR/VR technology has brought up the need for automatic parsing of floorplan images. However, there is a clear lack of representative datasets to investigate the problem further. To address this shortcoming, this paper presents a novel image dataset called CubiCasa5K, a large-scale floorplan image dataset containing 5000 samples annotated into over 80 floorplan object categories. The dataset annotations are performed in a dense and versatile manner by using polygons for separating the different objects. Diverging from the classical approaches based on strong heuristics and low-level pixel operations, we present a method relying on an improved multi-task convolutional neural network. By releasing the novel dataset and our implementations, this study significantly boosts the research on automatic floorplan image analysis as it provides a richer set of tools for investigating the problem in a more comprehensive manner.
Why we read it: CubiCasa5K's room-type taxonomy and annotation conventions inform the way our directive library categorizes residential rooms. The 80+ object categories the paper enumerates are a useful comparison set for our own RoomType + FixtureKind enums. Because the dataset is sourced from Finnish floor plans, it also gives us a calibration point for regional variation — Nordic plans tend toward different bathroom + sauna conventions than US plans, which our directive applicability fields should be able to express.
Licensing: CubiCasa5K is released under CC BY-NC. We do NOT download or use the dataset. We cite findings reported in the paper text only — taxonomy notes, annotation conventions, and the regional-variation observation. No CubiCasa images, polygons, or annotations are reproduced anywhere in the codebase.
2 findings extracted.
Architectural Graphic Standards
Charles G. Ramsey, Harold R. Sleeper, and American Institute of Architects (2017). "Architectural Graphic Standards". Wiley (12th edition, AIA).
Abstract
Architectural Graphic Standards is a comprehensive reference covering construction drawing conventions, building components, and human-scale dimensions across every building type. The 12th edition (2017) is the AIA-published current revision. The residential chapters codify the rules of thumb most US architects use when sizing bedrooms, hallways, kitchens, and bathrooms.
Why we read it: Every directive in our library tagged type='rule_of_thumb' draws from professional consensus that this book formalizes. Bedroom sizes, hallway widths, kitchen work zones, bathroom clearances — when our directive says 'industry consensus' as the source organization, this is the reference. We don't reproduce drawings or tables.
Licensing: The book is copyrighted by AIA and Wiley. We cite chapters and paraphrase the rules of thumb; we don't reproduce drawings, dimension tables, or section text. Link to the publisher for purchase.
2 findings extracted.
The Not So Big House: A Blueprint for the Way We Really Live
Sarah Susanka and Kira Obolensky (2008). "The Not So Big House: A Blueprint for the Way We Really Live". Taunton Press.
Abstract
Susanka argues that American houses became larger but less livable over the late 20th century, with formal living rooms and dining rooms going unused while families crowd into kitchens. The book proposes a vocabulary of small-house patterns — pocket offices, away rooms, layered ceilings, alcoves — that achieve the function of larger houses in less square footage. Widely cited in mid-market residential design.
Why we read it: Susanka's circulation-budget heuristics (hallways at 6-14% of total sqft) and her argument that 'every square foot should serve double duty' are referenced by several rules-of-thumb directives in our library. Her work is also the archetypal statement of why size hierarchy matters — primary bedroom > secondary bedrooms is satisfying because it expresses social structure visually.
Licensing: The book is copyrighted by the authors and Taunton Press. We cite the patterns and paraphrase principles; we don't reproduce the text. The user-facing UI links to the publisher's listing where the user can purchase the book.
1 finding extracted.
Residential Structural Design Guide
U.S. Department of Housing and Urban Development (2000). "Residential Structural Design Guide". HUD / NAHB Research Center. https://www.huduser.gov/portal/publications/manuhsng/residentialstructuralguide.html.
Abstract
The HUD Residential Structural Design Guide is a free, open-access reference covering structural systems, foundations, framing, and enclosure for one- and two-family dwellings. Companion HUD publications cover universal design and Fair Housing Act accessibility requirements.
Why we read it: HUD's universal-design recommendations (32-inch interior doorways, 36-inch entry clear widths, level-changes-as-ramps) inform our accessibility-tagged directives. The structural design guide doesn't directly drive layout directives but is the authoritative source we cite when accessibility-conscious directives are referenced.
Licensing: HUD publications are works of the US federal government — generally in the public domain in the US. We cite directly and may quote freely.
1 finding extracted.
A Pattern Language: Towns, Buildings, Construction
Christopher Alexander, Sara Ishikawa, and Murray Silverstein (1977). "A Pattern Language: Towns, Buildings, Construction". Oxford University Press. https://patternlanguage.com.
Abstract
A Pattern Language is a 1977 architectural treatise describing 253 design patterns at scales from city to room. Each pattern names a recurring problem and a 'core solution' that shows up across cultures. The residential patterns (Pattern 127 Intimacy Gradient, Pattern 129 Common Areas at the Heart, Pattern 130 Entrance Room, Pattern 188 Bed Alcove, etc.) are the most-cited prescriptive language in the residential design literature.
Why we read it: Every directive in our library tagged type='pattern' draws from Alexander's catalog. Patterns are the closest thing residential design has to first principles: testable claims about how rooms feel based on how they're related to each other. Our solver's soft constraints — intimacy gradient, kitchen-near-dining, primary-suite-isolation — are pattern-based and cite Alexander where applicable.
Licensing: The book is copyrighted by the authors. We cite by pattern number and paraphrase the principles in our own words. The user-facing UI links to patternlanguage.com where the user can purchase the book or read selected patterns.
4 findings extracted.
New papers, books, or standards added to the library go through the admin update queue.