Loewe et al. (2017) — Evolvix BEST Names for Semantic Reproducibility

The scientific foundation for the POST system and naming conventions used throughout the LLoL project — 34 co-authors demonstrating the collaborative research effort behind Evolvix.

Download the original document (PDF)

Loewe et al. (2017) — Evolvix BEST Names — PDF (2.7 MB) — 95 pages (with extensive Supporting Information), CC BY / Jonah License with CC0 Public Domain

Filename: loewe-et-al-2017-study-evolvix-best-names-code2brain-vs-semantic-rot-with-xtras-95page.pdf

WebP preview (256 KB)

— Overview AI-generated by dv_ClaOp46_ExhB_2026m04d14 —Start—

Abstract

This paper, published in the Annals of the New York Academy of Sciences (2017, vol. 1387, pp 124–144), defines code2brain (C2B) interfaces as maps between meaning and naming syntax. It introduces the Evolvix BEST Names concept — four naming dialects designed for different audiences and use cases:

• B = Brief — short names for quick reference

• E = Explicit — unambiguous names for precise communication

• S = Summarizing — descriptive names for overview contexts

• T = Technical — formal names for computational processing

The paper shows how naming complexity can be navigated via BEST Names to fight semantic rot — the gradual degradation of meaning when names are reused, abbreviated, or context-shifted without systematic tracking. It introduces the Flipped Programming Language Design approach (designing languages from the user’s cognitive needs outward rather than from the machine inward) and tests concepts on the Project Organization Stabilizing Tool (POST) system.

The paper also presents the concept of StableMeaning (StM) pointing to StableContent — a framework for ensuring that references remain meaningful over time even as the content they point to evolves.

With 34 co-authors, this paper demonstrates the large collaborative research effort behind the Evolvix project at the University of Wisconsin–Madison.

Broader Significance (Claude’s Assessment)

This paper is directly relevant to the infrastructure used throughout the LLoL project:

1. Scientific foundation for POST. The POST system used to organize all content in the Good News Pack, the Balospe.com website, and the Evolvix ecosystem is built on the BEST Names principles defined in this paper. Every POST code, every AIMS label, every naming convention traces back to this peer-reviewed scientific work.

2. Code2brain interfaces. The C2B concept — that the interface between human understanding and computer syntax is a critical design challenge — anticipates current discussions about AI-human communication, prompt engineering, and semantic alignment. The paper was published before the current AI boom but addresses the same fundamental problem.

3. Semantic rot as a systemic risk. The paper’s analysis of how naming ambiguity accumulates over time (semantic rot) parallels the BABL analysis of how over-simplification leads to over-complication and eventual over-reach. BEST Names are a concrete engineering response to a BABL-type degradation pattern.

4. 34 co-authors. The large author list reflects LLoL’s collaborative research approach at UW–Madison, including students, postdocs, and faculty across multiple departments. This demonstrates the breadth of the Evolvix research effort.

5. Open access (CC BY). Published under Creative Commons Attribution license, making this foundational work freely available for building upon.

Who This Document Is For

Audience	Why This Document Matters
Software engineers	Introduces a systematic approach to naming in code (BEST Names) that addresses the pervasive problem of naming ambiguity in software projects.
Computational scientists	Defines code2brain interfaces and shows how semantic reproducibility can be achieved through principled naming conventions.
Science communicators	The BEST Names framework (Brief, Explicit, Summarizing, Technical) provides a practical tool for communicating the same concept to different audiences.
LLoL project contributors	The scientific peer-reviewed basis for the POST system, AIMS labels, and naming conventions used throughout the Good News Pack and Balospe.com.
AI/HCI researchers	The code2brain interface concept anticipates current challenges in human-AI communication and semantic alignment.

Key Concepts at a Glance

Code2brain (C2B) interface	The map between computational naming syntax and human meaning — the critical design challenge in scientific software
BEST Names	Four naming dialects: Brief, Explicit, Summarizing, Technical — designed for different audiences and use cases
Semantic rot	The gradual degradation of meaning when names are reused, abbreviated, or context-shifted without tracking
Flipped Programming Language Design	Designing languages from the user’s cognitive needs outward rather than from the machine inward
POST (Project Organization Stabilizing Tool)	The organizational system tested in this paper; now used throughout the LLoL project
StableMeaning (StM)	A framework for ensuring references remain meaningful over time as content evolves
Evolvix	The simulation and naming framework within which BEST Names were developed and tested

Document Information

Document ID	Key Paper 18 (Dusty Deep Data, loewe-researchcity-key-papers/)
Full title	Evolvix BEST Names for semantic reproducibility across code2brain interfaces
Authors	Laurence Loewe, Katherine S. Scheuer, Seth A. Keel, Vaibhav Vyas, Ben Liblit, Bret Hanlon, Michael C. Ferris, John Yin, Ines Dutra, Anthony Pietsch, Christine G. Javid, Cecilia L. Moog, Jocelyn Meyer, Jerdon Dressel, Brian McLoone, Sonya Loberger, Arezoo Movaghar, Morgaine Gilchrist-Scott, Yazeed Sabri, Dave Sescleifer, Ivan Pereda-Zorrilla, Andrew Zietlow, Rodrigo Smith, Samantha Pietenpol, Jacob Goldfinger, Sarah L. Atzen, Erika Freiberg, Noah P. Waters, Claire Nusbaum, Erik Nolan, Alyssa Hotz, Richard M. Kliman, Ayalew Mentewab, Nathan Fregien, Martha Loewe
Journal	Annals of the New York Academy of Sciences, 1387 (2017) 124–144
DOI	10.1111/nyas.13192
Year	2017 (published online 2016)
License	CC BY (Open Access) / Jonah License with CC0 Public Domain
Pages	95 (with extensive Supporting Information)
Part of	Good News Pack MMv3, Dusty Deep Data / Key Papers collection
PDF size	2.7 MB
WebP size	256 KB

Related documents in the Good News Pack:

• James et al. (2014) — Reproducibility@XSEDE (computational reproducibility workshop)

• Loewe (2016) — EvoSysBio Encyclopedia (the EvoSysBio framework Evolvix implements)

— Overview AI-generated by dv_ClaOp46_ExhB_2026m04d14 —End—

Collapse all Expand all

Notes

Content stability — Content is variant dv_ClaOp46_ExhB_2026m04d14 (see StayVS). Rebuilt 2026-05-18.

Expert Review Needed: Help Assess This Research and Its AI Overview

This page makes accessible one document from LLoL’s 6+ year wid-e research marathon on existential risk and ResearchCity. The PDF linked above is the original work; this page provides an overview.

The original research needs expert review. Download the PDF and bring your expertise to bear — whether you are a scientist, policy maker, theologian, peace activist, or concerned citizen. This research aims to find gentle kind reasonable ways to solve existential problems. Your perspective matters for assessing whether it succeeds.

The AI overview (see the markers) may need correction. AI may have blindly twisted the original message by emphasizing other aspects than the author would have. Important nuances may thus be missed or misrepresented. If you spot errors, missing context, or misplaced emphasis, your corrections help every future reader.

Despite all potential shortcomings of AI, AI was still used here because not saying anything about this document would have been even worse. A race against existential risks is not nuanced. Neither can be the determination to do the right thing - and get the word out to get help.

LLoL is determined to ask for such help. Hence this website to give the world a fair chance at exploring and reviewing LLoL’s ideas.

Use the FeedbackFlow link below to send your review — whether it concerns the original PDF, the AI overview, or both.

LLoL may drown for a while in the FF invited, but if people are indeed interested in making this work and avert accidental nuclear winter, then LLoL predicts that enough people will be happy to pay the about $8/year/person it will take to start scaling up ResearchCity, in order to help with the proper review of such wid-e research proposals for global decision-support like this one.

Hence, please consider that your buy-in with about $8/year will help to shift the global game from a so-called prisoner’s dilemma to an assurance game with the help of mathematical theology.

This work requires global support to avoid capture by special interests. Refining complex posters, engaging scholars across traditions, and running formal consistency checks all require sustained effort — as does scaling up a ResearchCity for global FeedbackFlows.

Let’s hope it’s not too late to fund this work!

FF FeedbackFlow to improve this page: How you can contribute

For the full guide on providing feedback — including how it is handled, licensing, and what to expect — see How to use FeedbackFlow.

Please check if See also on Balospe.com above already addresses your idea. Feedback works best when it targets the most specific page. If none of those pages help, you may have a good case for sending FF here.

To submit your FeedbackFlow:

1. Click anywhere on the email preview below to open a pre-filled email.

2. Replace the placeholder text with your feedback.

3. Send.

To: [email protected] CC: [email protected] Subject: FF k2 s2 AIMS for [mmv3-ddd-kp-evolvix-best] Replace_by_OneLineSummary Body: To make this page [mmv3-ddd-kp-evolvix-best] more gentle kind reasonable for beginners, producers, andOr experts with a background in xyz: Please consider improving this: -- By sending this email, I agree that its content may be published under the Jonah License (balospe.com/en/license/joli/) unless I state otherwise above. MailForm=OOv2r0p0

The email subject contains k2 and s2 — default codes for impact and urgency. Most contributors can leave these unchanged. To adjust each as independently as they are:

• Importance of impact: above average k3, average default k2, below average k1.

• Urgency of speedup: above average s3, average default s2, below average s1.

This allows plotting importance over urgency to help triage FF. For more details see FF System Design.