5D Library › Computational Aesthetics, Visual Saliency Maps and Cognitive Ergonomics in User Interface Design · 中文
Chapter 7 Reading the Classics: Examining Curves and Aspects, and Timber as the Root
This chapter reads the earlier discussion of proportion, scale and “looking good” beside two ancient Chinese texts on craft and building: the opening of the Kaogongji (“Record of Trades”) in the Rites of Zhou, on the “hundred craftsmen”, and the rules in juan 4 of Li Jie’s Yingzao fashi (“Treatise on Architectural Methods”, Northern Song) that make the timber section cai the basic module. The aim is not to show that the ancients understood visual saliency or ergonomics, but to see how they expressed “make things according to the material and its disposition” and “derive all dimensions from one basic unit”, and to mark clearly the limits between ancient craft rules, ritual hierarchy and modern mathematical models. Quotations from the Kaogongji are intended to follow the edition the author specified, Sun Yirang’s Zhouli zhengyi (Zhonghua Shuju, punctuated edition), pending checking (孙诒让 撰,王文锦、陈玉霞 点校 1987); for the Yingzao fashi, Liang Sicheng’s annotated edition is the intended reference, also pending checking (李诫 撰;梁思成 注释 1983). Both are cited only by work, section and juan. This round of checking had access only to online transcriptions (the Kaogongji on the Chinese Text Project and electronic versions of two annotated readers of the Yingzao fashi), not to the printed editions; variant characters and punctuation are noted below. English renderings are this book’s own. Book 31 has already discussed the orientation by shadows and site selection in the “Master Builder” section of the Kaogongji, which is not repeated here.
The Kaogongji: examining curves, aspects and dispositions to work the five materials
Source text
国有六职,百工与居一焉。或坐而论道;或作而行之;或审曲面埶,以饬五材,以辨民器;或通四方之珍异以资之;或饬力以长地财;或治丝麻以成之。坐而论道,谓之王公;作而行之,谓之士大夫;审曲面埶,以饬五材,以辨民器,谓之百工;通四方之珍异以资之,谓之商旅。……知者创物,巧者述之守之,世谓之工。百工之事,皆圣人之作也。……天有时,地有气,材有美,工有巧。合此四者,然后可以为良。(孙诒让 撰,王文锦、陈玉霞 点校 1987)
(Rites of Zhou, “Offices of Winter”, Kaogongji)
In plain terms: the state has six occupations, and the hundred craftsmen are one of them. Some sit and discuss the Way; some rise and carry it out; some examine the curves, aspects and dispositions of materials so as to work the five materials into the wares people use; some circulate the rare goods of the four quarters to supply what is needed; some apply their strength to increase the wealth of the land; some work silk and hemp into cloth. Those who sit and discuss the Way are called kings and dukes; those who rise and carry it out, officers and ministers; those who examine curves, aspects and dispositions to work the five materials into the people’s wares, the hundred craftsmen; those who circulate the rare goods of the four quarters, merchants and travelling traders. … The wise create things; the skilful transmit and preserve them, and this, carried on through generations, is called craft. All the works of the hundred craftsmen were created by the sages. … Heaven has its seasons, earth its qi, materials their excellence and craftsmen their skill; only when these four are combined can something fine be made.
On 埶 and 势. The author’s framework and the traditional commentary editions read 审曲面埶, with 埶 read as 势 (“disposition”, “force of circumstance”); the online transcription available in this round (Chinese Text Project) reads 审曲面势. Here 埶 is read as 势 (elsewhere 埶 is read yì, as the original form of 藝, “art”). Following the practice of the edition the author specified, the book provisionally prints 埶 and, in the editorial notes, asks the author to settle the character against the printed Zhouli zhengyi. Commentators differ: according to secondary sources, the commentary of the elder Zheng (Zheng Sinong) reads “审察五材曲直方面形势之宜”, and Zhang Binglin proposed another reading; the phrase is generally understood as examining the curvature, facing and disposition of the material. The original commentaries were not checked for this book, and no further philological judgement is made here.
Historical context. The Kaogongji is a pre-Qin technical text on the crafts, generally dated to the late Spring and Autumn or Warring States period, later inserted into the Rites of Zhou to replace the lost “Offices of Winter”. It records rules for making carriages, weapons, musical instruments, ritual vessels, buildings and more; it is not a discussion of vision or cognition. Two points bear on this book. First, “examining curves, aspects and dispositions to work the five materials” places the starting point of making in the properties of the material itself: the craftsman first sees the curvature and disposition of the material and then decides how to work it, rather than fixing an abstract proportion and cutting the material to fit. Second, “heaven has its seasons, earth its qi, materials their excellence and craftsmen their skill; only when these four are combined can something fine be made” is a multi-factor view of quality: a fine object depends on season, place, material and skill together, and no single factor (such as a particular proportion) can guarantee it.
Reread in this book’s terms there is a structural resemblance, which must be labelled an analogy: the conclusion of Chapters 1–6 is that whether an interface element “works” depends on the users’ tasks, experience and visual system (Chapters 2–5), not on a universal proportion (Chapters 1 and 6); “examining curves, aspects and dispositions” likewise starts from the properties of the particular object. This is only a resemblance between two ways of thinking; the Kaogongji has no notion of reaction time, contrast or saliency.
The Yingzao fashi: timber as the root
The Yingzao fashi, compiled by Li Jie of the Directorate of Buildings and Construction, was printed in 1103 (the second year of Chongning) as the officially issued standard for building design and construction. Juan 4, “Rules for Major Carpentry, Part 1”, opens by defining the basic module of the whole book, the cai.
Source text
凡构屋之制,皆以材为祖;材有八等,度屋之大小,因而用之。
第一等,广九寸,厚六寸。(右殿身九间至十一间则用之。) 第二等,广八寸二分五厘,厚五寸五分。(右殿身五间至七间则用之。) 第三等,广七寸五分,厚五寸。(右殿身三间至殿五间或堂七间则用之。) 第四等,广七寸二分,厚四寸八分。(右殿三间、厅堂五间则用之。) 第五等,广六寸六分,厚四寸四分。(右殿小三间、厅堂大三间则用之。) 第六等,广六寸,厚四寸。(右亭榭或小厅堂皆用之。) 第七等,广五寸二分五厘,厚三寸五分。(右小殿及亭榭等用之。) 第八等,广四寸五分,厚三寸。(右殿内藻井或小亭榭施铺作多则用之。)
各以其材之广,分为十五分,以十分为其厚。凡屋宇之高深,名物之短长,曲直举折之势,规矩绳墨之宜,皆以所用材之分,以为制度焉。(李诫 撰;梁思成 注释 1983)
(Yingzao fashi, juan 4, “Rules for Major Carpentry, Part 1: Cai”)
In plain terms: all rules for building take the cai as their root; cai come in eight grades, chosen according to the size of the building. Grade 1 is nine cun deep and six cun broad, used for main halls of nine to eleven bays; … grade 8 is four cun five fen deep and three cun broad, used for coffered ceilings inside halls or small pavilions with many bracket sets. For every grade, the depth (guang) of the section is divided into fifteen fen, and ten of these fen are its breadth (hou). The height and depth of buildings, the lengths of all members, the disposition of curves and of roof pitch, and the fitting of compass, square and ink line are all fixed in fen of the cai used. The same juan also sets the filler block qi at six fen deep and four broad: “a cai with a qi added is called a full cai (zucai)”.
Textual check. The passages quoted in the author’s framework, “凡构屋之制,皆以材为祖;材有八等,度屋之大小,因而用之” and “各以其材之广,分为十五分,以十分为其厚”, agree with the transcriptions seen; only the punctuation after the first clause differs, one transcription using a semicolon and another a full stop. The usage notes in brackets are small-character notes in the original.
The program below checks these dimensions grade by grade and makes an arithmetic comparison.
Calculation
# The cai-fen module of the Yingzao fashi (juan 4): eight grades of cai, each "divided
# into 15 fen along its guang (depth), with 10 fen as its hou (breadth)". Sizes in cun
# as given in the text. Checks: guang/hou of every grade; 1 fen in cun; ratios between
# grades. Then an arithmetic comparison, labelled as an analogy only: the 3:2 section,
# the golden ratio and common screen ratios against the extent of the binocular visual
# field (taken here as about 200 deg wide and 130 deg high; textbook values vary).
import math
GRADES = [(9.0, 6.0), (8.25, 5.5), (7.5, 5.0), (7.2, 4.8), (6.6, 4.4), (6.0, 4.0),
(5.25, 3.5), (4.5, 3.0)] # (guang, hou) in cun, grades 1-8
PHI = (1 + 5 ** 0.5) / 2
if __name__ == "__main__":
print("grade guang hou guang/hou 1 fen (cun) guang / next grade")
for k, (g, h) in enumerate(GRADES, 1):
nxt = f"{g / GRADES[k][0]:.3f}" if k < 8 else " -"
print(f" {k} {g:5.2f} {h:5.2f} {g / h:9.3f} {g / 15:11.2f} {nxt:>16s}")
print(f" grade 1 / grade 8 = {GRADES[0][0] / GRADES[-1][0]:.2f};"
f" every section 15:10 = {15 / 10:.3f}; golden ratio {PHI:.3f}"
f" ({1.5 / PHI - 1:+.1%})")
zu = (15 + 6) / 10
print(f" zucai (cai + qi of 6 fen) = 21 fen : 10 fen = {zu:.2f}"
f" ({zu / PHI - 1:+.1%} from phi)")
print("analogy only: ratios vs an approximate binocular field of 200 x 130 deg")
field = 200 / 130
for name, r in (("visual field 200:130", field), ("cai section 15:10", 1.5),
("golden ratio", PHI), ("16:10 screen", 1.6),
("16:9 screen", 16 / 9), ("4:3 screen", 4 / 3)):
print(f" {name:22s} {r:6.3f} {r / field - 1:+7.1%} vs field")
alt = [w / h for w in (180, 190, 200, 210) for h in (120, 130, 135, 150)]
print(f" field 180-210 x 120-150 deg: ratio {min(alt):.2f} to {max(alt):.2f},"
f" covers 4:3, 3:2, 16:10, phi")
for w, h, d, lab in ((53.0, 30.0, 60.0, "24-inch 16:9 monitor at 60 cm"),
(7.0, 15.0, 30.0, "phone held at 30 cm")):
aw = 2 * math.degrees(math.atan(w / 2 / d))
ah = 2 * math.degrees(math.atan(h / 2 / d))
print(f" {lab:30s}: {aw:4.1f} x {ah:4.1f} deg,"
f" {aw * ah / (200 * 130):.1%} of the field's bounding box")
# grade guang hou guang/hou 1 fen (cun) guang / next grade
# 1 9.00 6.00 1.500 0.60 1.091
# 2 8.25 5.50 1.500 0.55 1.100
# 3 7.50 5.00 1.500 0.50 1.042
# 4 7.20 4.80 1.500 0.48 1.091
# 5 6.60 4.40 1.500 0.44 1.100
# 6 6.00 4.00 1.500 0.40 1.143
# 7 5.25 3.50 1.500 0.35 1.167
# 8 4.50 3.00 1.500 0.30 -
# grade 1 / grade 8 = 2.00; every section 15:10 = 1.500; golden ratio 1.618 (-7.3%)
# zucai (cai + qi of 6 fen) = 21 fen : 10 fen = 2.10 (+29.8% from phi)
# analogy only: ratios vs an approximate binocular field of 200 x 130 deg
# visual field 200:130 1.538 +0.0% vs field
# cai section 15:10 1.500 -2.5% vs field
# golden ratio 1.618 +5.2% vs field
# 16:10 screen 1.600 +4.0% vs field
# 16:9 screen 1.778 +15.6% vs field
# 4:3 screen 1.333 -13.3% vs field
# field 180-210 x 120-150 deg: ratio 1.20 to 1.75, covers 4:3, 3:2, 16:10, phi
# 24-inch 16:9 monitor at 60 cm : 47.7 x 28.1 deg, 5.1% of the field's bounding box
# phone held at 30 cm : 13.3 x 28.1 deg, 1.4% of the field's bounding boxThe depth-to-breadth ratio of every one of the eight grades is 1.500, that is 15:10 = 3:2; one fen runs from 0.60 cun in grade 1 to 0.30 cun in grade 8, and the depth of grade 1 is exactly twice that of grade 8. Ratios between neighbouring grades range from 1.042 to 1.167 and are not a fixed ratio (grades 3 and 4 differ by only 1.042); Guo Qinghua’s study argues that grade 3, the most commonly used, was the basic unit from which the other grades were derived (Guo 1995), and the competing explanations of the grading are not judged here. 3:2 differs from the golden ratio 1.618 by 7.3 %; the full cai with its qi is 21:10 = 2.10, 29.8 % away. The basic section of the Yingzao fashi is therefore the simple whole-number ratio 3:2, not the golden ratio; using it to support the claim that “ancient Chinese architecture also follows the golden section” is wrong, and it serves as a counterexample to “the golden ratio can be found everywhere” in Chapter 1.
The second half of the program is an analogy, not an argument: the aspect ratio of the human binocular visual field compared with other ratios. A common textbook statement is that the binocular field is about 200 degrees wide and 130 degrees high, a ratio of 1.538; but different sources give rather different values (for example, clinical field data reported at second hand put a single eye’s field at more than about 100 degrees temporally, about 60 degrees nasally, about 60–70 degrees above and about 75–80 degrees below; Strasburger and colleagues’ review says the horizontal extent of peripheral vision is “nearly 180 degrees” (Strasburger et al. 2011)). Letting the width vary between 180 and 210 degrees and the height between 120 and 150, the ratio ranges from 1.20 to 1.75, covering 4:3, 3:2, 16:10 and the golden ratio alike. “Screen proportions should match the proportions of the visual field” therefore has no testable content: the edges of the field are vague, peripheral acuity is extremely low, and a 24-inch 16:9 monitor at 60 cm subtends only about 47.7 × 28.1 degrees and a phone at 30 cm about 13.3 × 28.1 degrees, 5.1 % and 1.4 % of the field’s bounding box. What governs legibility is the visual angle and spatial frequency of Chapter 5, not the proportions of the outer frame.
Historical context. The immediate purpose of the cai–fen system was standardisation: one basic unit, scaled by grade, fixed the dimensions of every member, so that official buildings could choose a grade by scale and timber supply and labour could be standardised and estimated. It was also a system of rank: which grade was used depended on the class of building (main hall, hall, chamber, pavilion), tied to ritual. Feng Jiren’s study discusses Song culture and metaphor in the Yingzao fashi (Feng 2012). Comparing cai–fen with the “design tokens” or “8-point grid” of modern design systems is therefore an apt structural analogy as far as modularity goes: both derive all dimensions from one basic unit, and changing the grade (or the base font size) rescales the whole; but the ancient grades also encoded status and ritual, while modern grids encode screen pixels and legibility, and their purposes differ.
Structural comparison and its limits
| Ancient text | Modern concept | Limit of the comparison |
|---|---|---|
| “Examining curves, aspects and dispositions to work the five materials” | Design from the properties of the object (user, task, visual system), not a universal proportion | The Kaogongji is about materials and craft; it has no psychological or information-theoretic concepts |
| “Heaven has its seasons, earth its qi, materials their excellence and craftsmen their skill; only when these four are combined can something fine be made” | A multi-factor view of quality: saliency, legibility, choice and pointing jointly determine usability | Only a resemblance of thought; the ancient “fine” includes ritual propriety and durability, not task time |
| “Timber as the root”; “all fixed in fen of the cai used” | Modular design systems (base unit, proportional scaling) | An apt structural analogy, but the grades of cai were also ranks of ritual |
| All eight grades 15:10 | A simple whole-number ratio, not the golden ratio | A factual check, not an analogy |
| Visual field about 200° × 130° and screen proportions | “Screens should match the visual field” | The field’s edges are vague and its values vary by source; the range of ratios covers several proportions at once; an arithmetic analogy only |
These comparisons are only structural resemblances. First, the Kaogongji is a pre-Qin craft text and the Yingzao fashi a Northern Song official building standard; neither is a study of vision or cognition. Second, the ancient texts have no notion of contrast sensitivity, information entropy or saliency, and it cannot be said that the ancients had “already discovered” the principles of modern interface engineering. Third, the proportions of the Yingzao fashi are whole-number ratios, and using them to support the golden section is an error. The book says only that the two can illuminate each other structurally, not that the ancients “long ago discovered” computational aesthetics.
Chapter summary
The Kaogongji’s “examining curves, aspects and dispositions to work the five materials” places the starting point of making in the properties of the material itself, and “heaven has its seasons, earth its qi, materials their excellence and craftsmen their skill” is a multi-factor view of quality; the variant 埶/势 should be settled by the author against Sun Yirang’s punctuated edition. Juan 4 of the Yingzao fashi takes the cai as its module; all eight grades have a 15:10 = 3:2 section, not the golden ratio, and the full cai is 21:10. The comparison of the human visual field with screen proportions is only an arithmetic analogy: the field’s values depend on the source, and the range of ratios covers several proportions at once. Cai–fen and modern design systems admit a structural analogy in their modularity, but the ancient grades were also ranks of ritual.
References
Fan, Y. (2026). Computational Aesthetics, Visual Saliency Maps and Cognitive Ergonomics in User Interface Design: From the Hick–Hyman Law and Spatial Frequency Channels to Saccadic Trajectory Modelling, Read Beside the Kaogongji and Classical Treatise on Architectural Methods [计算美学、视觉显著性图谱与人机界面认知工程] (Version 1.0.0). 5D Library. https://doi.org/10.5281/zenodo.23836258 · Licensed under Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International (CC BY-NC-ND 4.0).