A paper from the Sloan Digital Sky Survey (SDSS), the Dark Energy Spectroscopic Instrument (DESI), or the Vera C. Rubin Observatory’s Legacy Survey of Space and Time (LSST) routinely carries dozens to hundreds of author names, most of whom never wrote a line of the analysis it reports. That looks superficially like the high-energy-physics megacollaboration pattern covered in CASRAI’s guide to ATLAS, CMS, and LIGO authorship — a standing, alphabetical, membership-based author list — but astronomy survey collaborations arrived at that outcome through a mechanically different route, and the differences matter to anyone trying to read a survey co-authorship credit accurately.
Why survey collaborations aren’t just “megacollaborations with telescopes”
ATLAS, CMS, and LIGO/Virgo qualify members for standing authorship primarily through documented service to the experiment — commissioning hardware, operating software, taking data-taking shifts — reviewed and tracked by the collaboration itself. Large astronomy surveys use a related but distinct mechanism layered on top: a formally defined Builder or Architect status tied specifically to having helped construct, fund, or commission the survey’s instrument or software infrastructure, which then confers a standing right to request co-authorship on the survey’s science papers — separately from, and in addition to, ordinary analysis-based authorship. That construction-linked credit category is the load-bearing difference this guide focuses on, along with two further mechanics HEP megacollaborations don’t share: a hybrid two-tier author-order convention (part contribution-ranked, part alphabetical) used by SDSS and DESI, and Rubin Observatory/LSST’s federated structure, where there is no single collaboration-wide author list at all.
SDSS: Architects, the two-group model, and an ombudsperson
SDSS’s current publication policy (the collaboration is now in its fifth phase, SDSS-V) sets out two distinct paths onto a science paper’s author list. Ordinary science authorship goes to anyone who made a “significant contribution to that specific research analysis.” Separately, Architects — a defined category for people who contributed substantial effort (on the order of a year of full-time work) to survey hardware, software, calibration, commissioning, management, or education/outreach, rather than to a specific science result — may, once approved, “add their name to the author list of any science paper that they have read and whose main content and results they broadly understand.” That is a materially different bar from science authorship: an Architect qualifies once, based on infrastructure work, and can then appear on papers they had no analytical role in, provided they’ve actually read and understood the specific paper.
Most SDSS science papers use what the policy calls a two-group model: an analysis group, ordered by the same kind of contribution-based negotiation any small-team paper would use, followed by other qualifying authors — including Architects who opt in — typically listed alphabetically. Large-scope publications, particularly SDSS data-release papers describing an entire data set rather than one analysis, instead use a single alphabetical author list covering the whole qualifying membership. Authorship requests go through a webform with a brief description of the contribution (Architects need only indicate interest); the paper’s first author and analysis-group members decide group membership, and can offer acknowledgment instead of authorship where the contribution doesn’t meet the bar. Disputes over analysis-group membership or order that can’t be resolved internally go to an Ombudsperson, whose decision is final. Every SDSS publication also carries a standard collaboration acknowledgment, maintained centrally by the Spokesperson, that is separate from and additional to the author list itself.
DESI: Builder status and the two-set author-order model
DESI’s Publication Policy (maintained by the collaboration’s Institutional Board and Publication Board) uses the term Builder for essentially the same idea SDSS calls Architect: a formally granted status, earned through contribution to constructing or commissioning the instrument and survey infrastructure, that carries a standing right to request co-authorship on any DESI paper — independent of involvement in the specific science analysis being published.
Outside of DESI’s flagship Key Papers, ordinary DESI papers use what the policy describes as a two-set model: a first set of authors — those most directly engaged in the specific analysis or technical development the paper reports — in an order the group agrees on among themselves, the same way authors on any small paper would negotiate order; followed by a second set, everyone else with a standing right to authorship (including Builders who opt in), listed alphabetically. A paper’s readiness for submission is reviewed by a Review Committee; in the rare case where the Review Committee and the authors can’t reach consensus, the Publication Board can call a formal vote, in which every author who has co-signed the paper by the end of its internal Collaborative Writing and Review period — Builders included — gets a vote. [REPORTED — drawn from search-indexed excerpts of DESI’s published policy PDF, which returned a 403 on direct fetch; the structural description above is consistent across the excerpts retrieved and treated as reliable, but hasn’t been independently confirmed against the full primary document this session.]
Rubin Observatory / LSST: no single author list — a federation of Science Collaborations
Rubin Observatory (the facility that conducts the Legacy Survey of Space and Time, LSST) doesn’t maintain one collaboration-wide author list the way ATLAS, SDSS, or DESI do. The Observatory’s own project publication policy governs technical papers and data-release papers written by Rubin staff, but science analysis of LSST data is instead carried out by a set of independently organized Science Collaborations (SCs) — thematic groups such as the Dark Energy Science Collaboration (DESC), each of which sets and governs its own membership rules, publication policy, and authorship criteria. The one rule Rubin Observatory imposes across all of them is eligibility: only holders of data rights (broadly, scientists and students affiliated with a qualifying US or Chilean institution, plus a published list of international data-rights holders) can join an SC in the first place — data rights is a distinct, separate axis from authorship qualification within whichever SC a researcher joins.
DESC’s own publication policy illustrates how the Builder concept carries through this federated structure: DESC Builders can co-author any DESC Key Paper, and any Standard Paper their contribution enabled; DESC Full Members can co-author any Key or Standard Paper they contributed significantly to. Rubin Observatory’s own construction-team Builders carry equivalent rights into DESC, scaled to their DESC membership tier — Rubin Builders who are also DESC Full Members get DESC Builder-level rights, while Rubin Builders who are ordinary DESC Members get ordinary DESC Member rights. The practical effect: a Rubin/LSST-data paper’s authorship has to be evaluated against whichever Science Collaboration published it, not against a single Rubin-wide policy, and its data-rights gate is a separate prerequisite layered underneath authorship criteria, not a substitute for them.
Survey authorship models side by side
| Collaboration | Standing-authorship mechanism | Author-order convention | Dispute/oversight body |
|---|---|---|---|
| SDSS | Architect status (infrastructure contribution) or per-paper science-analysis contribution | Two-group: analysis group by negotiated order, then alphabetical — single alphabetical list for data-release papers | Ombudsperson (final on disputes); acknowledgment maintained by Spokesperson |
| DESI | Builder status (construction/commissioning contribution) or per-paper analysis contribution | Two-set: analysis-engaged authors by negotiated order, then alphabetical for all other qualifying authors | Review Committee, escalating to a Publication Board vote of all co-signed authors |
| Rubin Observatory / LSST | No single Rubin-wide list; each Science Collaboration (e.g. DESC) sets its own Builder/Member tiers, gated by Rubin data-rights eligibility | Set independently per Science Collaboration | Governed independently per Science Collaboration |
| ATLAS / CMS / LIGO-Virgo (for comparison) | Standing membership via documented service/effort threshold — see the HEP megacollaboration guide | Single alphabetical list, collaboration-wide | Publications Committee (ATLAS); LSC governance (LIGO) |
Where individual credit actually lives, and why CRediT statements don’t appear here
As with the HEP megacollaborations, a Builder-status or alphabetical byline on an SDSS, DESI, or Rubin/LSST-derived paper carries little to no information about what a specific person actually did on that specific result — the ordinary reason a CRediT contribution statement exists in the first place. CRediT’s 14 roles are calibrated for teams of a handful to a few dozen contributors credited against one output; they aren’t practically published per-author against a survey’s full qualifying membership, and none of the three policies described above use one. Individual credit instead lives in the same kind of internal machinery documented for HEP collaborations: analysis-note authorship, internal review records, Builder/Architect qualification files, and conveners’ or working-group leads’ assessments — real, but not visible from outside the collaboration. See CASRAI’s CRediT & Authorship Attribution overview for the model that does work at ordinary team scale, and the group authorship dictionary entry for how journals and indexes handle a collaboration name standing in for an author list generally.
What this means for research administrators and hiring committees
A CV or NIH-style biosketch listing co-authorship on an SDSS data-release paper or a DESI Key Paper cannot be read the way a five-author paper’s byline is read, and reviewers outside astronomy routinely need to be told that explicitly. Two things are worth checking specifically, given the mechanics above: whether the credit in question reflects Builder/Architect status (construction and commissioning contribution) or analysis-group membership (contribution to that specific result) — the two are not interchangeable, and a strong record of one doesn’t imply the other — and, for Rubin/LSST-derived work, which Science Collaboration actually governed the paper, since authorship criteria and order conventions are set independently per SC rather than by Rubin Observatory as a whole. As with HEP megacollaborations, researchers from these fields typically supplement the raw author list with a separate personal-statement account of specific analyses led, hardware or software built, or working-group roles held — the practical equivalent of a CRediT statement for a paper too large to carry one.
Frequently asked questions
Is “Builder” or “Architect” status the same as being an author on every paper the survey publishes?
Not automatically. It grants a standing right to request co-authorship, which the individual (or, for some categories, their opting-in) still exercises paper by paper, and it typically requires having actually read and broadly understood the specific paper — SDSS’s policy states this explicitly for Architects. It is a qualification pathway, not blanket automatic authorship.
How is this different from the alphabetical author lists on ATLAS or CMS papers?
ATLAS and CMS qualify essentially all standing authors through one track — documented service/effort to the experiment — and then list the whole qualifying membership alphabetically on every paper. SDSS and DESI instead combine two tracks (per-paper analysis contribution and construction/commissioning-based Builder or Architect status) and, for most papers, order the byline in two tiers rather than a single flat alphabetical list. Rubin/LSST goes further still and has no single collaboration-wide list at all.
Does a survey collaboration’s author list appear the same way in citation databases as an ordinary multi-author paper?
Indexing and citation databases generally still list every named author, the same as any paper, though very large author counts can strain how some tools display or de-duplicate author records. What the databases can’t recover is the internal distinction this guide describes — Builder/Architect status versus analysis contribution — since that information isn’t encoded in the byline itself.
Can someone be a DESI Builder and also lead the analysis-group ordering on a specific paper?
Yes — the two tracks aren’t mutually exclusive. A Builder who is also substantially engaged in a given paper’s specific analysis would typically be placed in the paper’s first, contribution-ordered author set for that reason, rather than (or in addition to) relying on their Builder standing, which would place them in the alphabetical second set if that were their only basis for authorship.
Where can I read the actual policies rather than a summary?
SDSS publishes its current policy at sdss.org/collaboration/publication-policy; DESI’s is maintained on its collaboration wiki (DESI Publication Policy, approved by its Institutional Board); Rubin Observatory’s project-level and Science-Collaboration-level policies, including DESC’s, are linked from lsst.org’s LSST Data Policy page.
