Sources: the pool, the flow and the scores
This is the pool of the architecture-design review: how screening narrowed it, then one row per source (kind, route found, quality, use). Method and strings are in protocol; terms are in frame.
Flow
| Stage | Formal | Grey |
|---|---|---|
| Returned by the strings | 1,804 (Semantic Scholar only) | not recorded |
| Screened | 101 | 244 |
| Included as new | 62 | 30 |
| Carried from the repository | 3 | 0 |
| In the review | 65 | 30 |
- Formal:
- 1,804 records were returned by 31 strings (Semantic Scholar); arXiv counts were not obtained (HTTP 429).
- 101 titles were screened (the top hits of strings R1 to R17); strings F1 to F14 gave counts only.
- 62 works were included as new, found through leads, reviews and title lookups. Among them are 13 added in the second pass (socio-technical congruence, hotspots and architectural roots, architects' reasoning, Shaw and Garlan, Accelerate).
- 3 were carried from earlier reviews of this repository (Garousi 2019, Tofan 2014, Gloaguen 2026).
- Excluded after reading a title: hits on robotics, digital twins, cyber-physical systems, security and sustainability architectures, not counted; 5 more relevant works located and not read (three microservice migration studies; two tactics reviews).
- Grey:
- 244 Q/A records were screened by title; web results were about 10 per search over 7 searches and 19 pages read directly, not counted per string.
- 30 were included as new: 11 Q/A threads, 13 web sources scored, and 6 books listed but not opened.
- Excluded: Prime Video's page (moved; read only in secondary write-ups), 233 Q/A records, 5 pages not read.
- Added by snowballing: none.
- Existing reviews found: 14, listed in review.md strand 0.
The figure draws the stages of the table for architecture design.
Formal sources
"narrative" means carried from an earlier review of this repository. Quality words follow the README grades.
- "Abstract read": the abstract was read in full from the publisher, a repository or a preprint.
- "In a search summary" or "in a listing": it was read in another source's account (secondary).
- "Summary line": a one-sentence summary only.
- "Located only": no result is claimed beyond the title.
| Key | Kind | Found via | Quality | Used in |
|---|---|---|---|---|
| Garousi 2019 | formal | narrative (earlier pass) | method paper | protocol |
| Tofan 2014 | formal | earlier pass; R1 | systematic mapping; linked, full text read in design-decisions | review 0 |
| Gloaguen 2026 | formal | earlier pass | preprint; linked only | review 11 |
| Hofmeister 2007 | formal | lead | comparison of five industrial methods; full text read through a library | review 1 |
| Falessi 2011 | formal | lead (found by title) | survey of decision-making techniques; abstract read in a search summary (secondary) | review 0, 1 |
| Shaw 1996 | formal | lead (restored) | book, not opened; library record and table of contents read | review 3 |
| Parnas 1972 | formal | lead | argument from one worked example; abstract read in a search summary (secondary); full text not reached | review 2 |
| Stevens 1974 | formal | lead | argument; located only, abstract not reached (publisher gives none) | review 2 |
| Baldwin 2000 | formal | lead | book, not opened | review 2, 6 |
| MacCormack 2006 | formal | lead | empirical, two products; abstract read | review 2 |
| MacCormack 2012 | formal | lead | empirical test of mirroring; abstract partly read | review 6 |
| Colfer 2016 | formal | lead | review of 142 studies; full text read (open access) | review 6 |
| Cataldo 2008 | formal | lead | empirical framework; summary line read (secondary); abstract not reached | review 6 |
| Cataldo 2013 | formal | lead | empirical, two large projects; abstract read in a search summary (secondary) | review 6 |
| Mauerer 2022 | formal | found by search summary | measured, 25 open source projects; abstract read | review 6 |
| Herbsleb 1999 | formal | lead (found by title) | case study; abstract read in a search summary (secondary) | review 6 |
| Nagappan 2008 | formal | lead (found by title) | measured, one company; abstract read | review 6 |
| Basili 1996b | formal | found by title | measured, 8 projects; abstract read | review 2 |
| Briand 2000 | formal | found by title | measured; abstract read in a summary, numbers (secondary) | review 2 |
| Xiao 2014 | formal | lead (Mo, Cai, Kazman) | measured, three open source projects; abstract read | review 2, 7 |
| Mo 2015 | formal | lead (Mo, Cai, Kazman) | measured, nine open source projects and one commercial; abstract read in a search summary (secondary) | review 2, 7 |
| Mo 2016 | formal | lead (Mo, Cai, Kazman) | measured, 108 open source and 21 industrial projects; abstract read in a search summary (secondary) | review 2, 7 |
| Kazman 2015 | formal | lead (Kazman) | industrial case study; abstract read in a search summary, no numbers (secondary) | review 7 |
| Nayebi 2019 | formal | found by search summary | longitudinal industrial case; abstract read | review 7 |
| Prechelt 2001 | formal | found by search summary | controlled experiment, professionals, nine tasks; abstract read on the group's page | review 3; against |
| Prechelt 2002 | formal | lead | two controlled experiments; draft abstract, secondary | review 3 |
| Zhang 2012 | formal | lead | systematic mapping; abstract read | review 0, 3 |
| Khomh 2008 | formal | F12 neighbourhood | empirical; abstract of the technical report version read in a search summary (secondary) | review 3; against |
| Vokac 2004 | formal | F12 neighbourhood | industrial case, 500,000 lines over three years; abstract read in a record (secondary) | review 3; against |
| Ameller 2013 | formal | found by title | interviews; located only, abstract not reached | review 4 |
| Bellomo 2015 | formal | found by title | descriptive, ATAM data; abstract read | review 4 |
| Kazman 2000 | formal | lead | technical report describing the method and one case; full text read | review 4 |
| Babar 2004 | formal | R6 | comparison of four methods; full text read (university repository copy) | review 0, 4 |
| Dobrica 2002 | formal | found by title | survey of eight analysis methods; abstract read | review 0, 4 |
| Abrahamsson 2010 | formal | lead | position; introduction read in a search summary, abstract not reached in full | review 5 |
| Waterman 2015 | formal | lead | grounded theory, 44 participants; abstract read in a listing | review 5 |
| Boehm 2008 | formal | found by title | measured, 161 projects; abstract read | review 5 |
| Li 2022 | formal | R4 | systematic mapping, 73 studies; abstract read | review 0, 7 |
| de Silva 2012 | formal | lead | survey; abstract read in a search summary (secondary) | review 0, 7 |
| Le 2018 | formal | R5 | empirical; located only, opening of the abstract read | review 7 |
| Besker 2018 | formal | R3; lead | systematic review; abstract read in a search summary, no count (secondary) | review 0, 7 |
| Verdecchia 2021 | formal | lead | grounded theory and focus groups; abstract read | review 7 |
| Ernst 2015 | formal | found by title | survey, 1,831 participants; secondary | review 7 |
| Mumtaz 2021 | formal | R5 | systematic mapping; abstract read | review 0, 7 |
| Baabad 2020 | formal | R4 | systematic review, 74 primary studies; full text read | review 0, 7 |
| Taibi 2017 | formal | found by title | interviews; secondary | review 8 |
| Di Francesco 2018 | formal | found by title | industrial survey, 18 practitioners; full text read (open copy) | review 8 |
| Blinowski 2022 | formal | found by title | measured, controlled experiments; abstract read | review 8 |
| Mendonca 2021 | formal | G1 | company report in a peer-reviewed magazine; abstract read | review 8 |
| Ghemawat 2023 | formal | found by title | prototype and position; abstract read | review 8 |
| Su 2024 | formal | R9 | multivocal review; abstract read | review 0, 8 |
| Vera-Rivera 2021 | formal | R9 | systematic review, 29 of 326 papers; abstract read | review 0, 8 |
| Tang 2010 | formal | lead (Tang) | experiment on design effectiveness; summary line read (secondary); abstract not reached | review 12 |
| Tang 2012 | formal | lead (Tang) | observation of two design teams; abstract read in a search summary (secondary) | review 12 |
| Tang 2015 | formal | lead (Tang) | experiment on design decisions; summary line read (secondary); abstract not reached | review 12 |
| Razavian 2016 | formal | lead | multiple case studies, 12 student groups; abstract read through a library | review 12 |
| Mohanani 2020 | formal | lead | systematic mapping, 65 articles; preprint abstract read | review 0, 12 |
| Zannier 2007 | formal | lead | interviews with designers; summary line read (secondary); abstract not reached | review 12 |
| Forsgren 2018 | formal | lead (restored) | book, not opened; publisher page read | review 6 |
| Esposito 2025 | formal | R8; R12 | multivocal review; abstract read | review 0, 11 |
| Schmid 2025 | formal | R12 | systematic review, 18 articles; abstract read | review 0, 11 |
| Dhar 2024 | formal | F14 | exploratory study; abstract read | review 11 |
| Li 2024 | formal | F14 | benchmark case study; abstract read | review 11 |
| Hong 2023 | formal | F14 | framework; abstract read | review 11 |
| Su 2026 | formal | R12; F14 | mining of 980 decision records; abstract partly read | review 11 |
Grey sources, scored
The eight subtotals appear as authority, methodology, objectivity, date, position, novelty, impact and outlet, as in the method note. Under 10 of 20, a source is an example of practice only.
| Key | Kind | Found via | Quality (total, then subtotals) | Used in |
|---|---|---|---|---|
| Conway 1968 | grey tier 1 | P | 15.5 of 20 (4 / 2 / 3.5 / 1 / 1 / 2 / 1 / 1) | review 6 |
| Lewis 2014 | grey tier 3 | P | 14 of 20 (4 / 2.5 / 3 / 1 / 0.5 / 2 / 1 / 0) | review 8, 9 |
| Fowler 2015 | grey tier 3 | P | 10.5 of 20 (3.5 / 1.5 / 2.5 / 1 / 0.5 / 1 / 0.5 / 0) | review 8, 9 |
| Noonan 2018 | grey tier 3 | G1 | 12.5 of 20 (3 / 3.5 / 2 / 1 / 0.5 / 1.5 / 1 / 0) | review 8, 9 |
| Müller 2020 | grey tier 3 | P | 12.5 of 20 (3 / 3 / 2.5 / 1 / 0.5 / 1.5 / 1 / 0) | review 2, 9 |
| Loukides 2020 | grey tier 1 | G4 | 14 of 20 (3 / 4 / 2.5 / 1 / 0.5 / 1 / 1 / 1) | review 8, 9 |
| DORA 2025 | grey tier 2 | P | 14.5 of 20 (4 / 3 / 3 / 1 / 1 / 1 / 1 / 0.5) | review 6, 9 |
| ThoughtWorks 2014 | grey tier 1 | P | 10.5 of 20 (3 / 1.5 / 2 / 1 / 0.5 / 1 / 0.5 / 1) | review 6, 9 |
| ThoughtWorks 2017 | grey tier 1 | G2 | 10.5 of 20 (3 / 1.5 / 2 / 1 / 0.5 / 1 / 0.5 / 1) | review 9 |
| Microsoft 2025 | grey tier 1 | P | 12.5 of 20 (3.5 / 2.5 / 2 / 1 / 0.5 / 1 / 1 / 1) | review 3, 9 |
| AWS 2026d | grey tier 1 | P | 10.5 of 20 (3.5 / 2 / 2 / 0 / 0.5 / 0.5 / 1 / 1); a blind second rater gave 9.5, so an example of practice only | review 4, 9 |
| ISO 2023c | grey tier 1 | G6 | 14 of 20 (3.5 / 2 / 3.5 / 1 / 1 / 1 / 1 / 1) | review 4, 9 |
| Team Topologies 2026 | grey tier 2 | P | 8.5 of 20 (3 / 1 / 1.5 / 0 / 0.5 / 1 / 1 / 0.5) | review 6, 9 |
Below the threshold: Team Topologies 2026 (example of practice only). Books, not opened, not scored (named, no claim rests on them): Cervantes 2016, Bass 2021, Evans 2003, Ford 2017, Skelton 2025, Ousterhout 2021.
Grey sources from Q/A sites
These threads are grey tier 2, with the date read 2026-10-29. The subtotals are as above, and the rules are in the shared protocol. The votes, views and answers per thread, as read on 2026-10-29, are in bibliography.
| Key | Kind | Found via | Quality (total, then subtotals) | Used in |
|---|---|---|---|---|
| SESE 2014d | grey tier 2 | AR1 | 12 of 20 (1 / 3.5 / 3.5 / 1 / 0.5 / 1.5 / 0.5 / 0.5) | review 10 |
| SESE 2017h | grey tier 2 | AR7 | 17 of 20 (3 / 4.5 / 4 / 1 / 1 / 2 / 1 / 0.5) | review 8, 10 |
| SESE 2018d | grey tier 2 | AR7 | 15 of 20 (2.5 / 4 / 4 / 1 / 0.5 / 1.5 / 1 / 0.5) | review 8, 10 |
| SESE 2020b | grey tier 2 | AR7 | 14.5 of 20 (2.5 / 4 / 4 / 1 / 0.5 / 1.5 / 0.5 / 0.5) | review 8, 10 |
| SESE 2020c | grey tier 2 | AR8 | 12.5 of 20 (2.5 / 3 / 3.5 / 1 / 0.5 / 1.5 / 0 / 0.5) | review 8, 10 |
| SESE 2025a | grey tier 2 | AR2 | 11.5 of 20 (1.5 / 3.5 / 3.5 / 1 / 0 / 1.5 / 0 / 0.5) | review 8, 10 |
| SESE 2012k | grey tier 2 | AR9 | 12.5 of 20 (1.5 / 3 / 4 / 1 / 0.5 / 1 / 1 / 0.5) | review 2, 10 |
| SESE 2020d | grey tier 2 | AR9 | 14 of 20 (2 / 4 / 4 / 1 / 0.5 / 1.5 / 0.5 / 0.5) | review 2, 10 |
| SESE 2022b | grey tier 2 | AR15 | 14.5 of 20 (2.5 / 4 / 4 / 1 / 0.5 / 1 / 1 / 0.5) | review 10 |
| SESE 2011k | grey tier 2 | AR5 | 11 of 20 (1.5 / 3 / 3.5 / 1 / 0 / 1 / 0.5 / 0.5) | review 2, 10 |
| SO 2010c | grey tier 2 | AR4 | 9 of 20 (2 / 1.5 / 3 / 1 / 0 / 1 / 0 / 0.5) | review 10 |
Below the threshold: SO 2010c (example of practice only).
Adoption counts, descriptive, read 2026-10-29
Public repository pages, read without identity; stars measure interest, not use.
| Count | Value | Caveat |
|---|---|---|
| stars: ArchUnit (a Java library to assert architecture rules) | 3,882 | created 2017-04 |
| stars: Spring Modulith (modular applications with Spring Boot) | 1,206 | created 2022-07 |
| stars: Packwerk (Shopify's boundary checker, Müller 2020) | 1,928 | created 2020-09 |
| stars: modular-monolith-with-ddd (a sample) | 14,101 | a worked example, last push 2024-06 |
| stars: awesome-software-architecture (a list) | 11,748 | a reading list |
| repositories with topic software-architecture | 4,221 | self-tagged |
| respondents using microservices (Loukides 2020) | 77% of 1,502 | mailing-list sample |