Refactor your specs! Øredev 2013
4 likes1,862 views
The document discusses the importance of refactoring specifications in software development, emphasizing the need for TDD, BDD, and DDD practices to simplify and enhance clarity. It highlights patterns for recognizing opportunities for refactoring and invites collaboration during specification workshops to improve documentation and adapt to a business perspective. The goal is to foster effective communication and create living documentation that evolves with the business needs.
![Break the model
[Chris Matts]
examples -> model -> more examples](https://image.slidesharecdn.com/refactoryour-specscyriux-131212075545-phpapp01/85/Refactor-your-specs-Oredev-2013-70-320.jpg)
![Systematic
credit application
Credit
score
OK?
Compute
Credit
Limit
[]
References
Check
Record
Outcome
l s te p
io n a
Op t](https://image.slidesharecdn.com/refactoryour-specscyriux-131212075545-phpapp01/85/Refactor-your-specs-Oredev-2013-92-320.jpg)
![Systematic
]
[
Compute
Credit
Limit
l s te p
anua
M
References
Check
Record
Outcome](https://image.slidesharecdn.com/refactoryour-specscyriux-131212075545-phpapp01/85/Refactor-your-specs-Oredev-2013-95-320.jpg)
![Living documentation
[Gojko Adzic]
m p le s
Exa
n te n t
I
e n ts)
c omm
(Fre e](https://image.slidesharecdn.com/refactoryour-specscyriux-131212075545-phpapp01/85/Refactor-your-specs-Oredev-2013-164-320.jpg)
![Visible Workings
[Brian Marick]
o c le t
t om D
C us
p o rt
i me e x
Run t
o rt. . .
e x e xp
L aT](https://image.slidesharecdn.com/refactoryour-specscyriux-131212075545-phpapp01/85/Refactor-your-specs-Oredev-2013-169-320.jpg)
Refactor your specs! Øredev 2013
- 1. Refactor your Specs! Cyrille Martraire - @cyriux -
- 2. what if specs could be improved?
- 3. Specs
- 4. Specs
- 5. Specs
- 10. Soft ware Craftsmanship bar " g the aisi n "R g c o de or k in "W o ugh" ot en is n
- 11. Soft ware Craftsmanship abou t h at W y we h e wa t e c i f y? sp
- 13. A project...
- 16. Case 1
- 17. Code It!
- 18. (with BDD) 3M ond on EURIBOR ate b en a floating r 5M EUR Giv f 1 And a nominal o 15 ate of 2011/06/ And an issue d of 2012/06/14 And an end date UAL calculation period And an SEMI_ANN es: URIBOR 3M evolv When the E 5% | 2011/09/15 | 3. | 0% | 2012/03/15 | 4. | re: e cash-flows a0 | Then th 2300 | 2011/12/13 | 0 | 12/06/14 | 2550 | 20
- 19. DONE double sum = 0; for (CashFlow cf : cashflows){ double df = 1 + rate * t; sum += cf.getAmount() / df; } return sum;
- 21. Case 2
- 22. DONE double sum = 0; for (CashFlow cf : cashflows){ double df = pow(1 + rate), t); sum += cf.getAmount() / df; } return sum;
- 24. Spot the difference double sum = 0; for (CashFlow cf : cashflows){ double df = 1 + rate * t; sum += cf.getAmount() / df; } return sum; double sum = 0; for (CashFlow cf : cashflows){ double df = pow(1 + rate), t); sum += cf.getAmount() / df; } return sum;
- 25. Spot the difference double sum = 0; for (CashFlow cf : cashflows){ double df = 1 + rate * t; sum += cf.getAmount() / df; } return sum; double sum = 0; for (CashFlow cf : cashflows){ double df = pow(1 + rate), t); sum += cf.getAmount() / df; } return sum;
- 26. Duplication?
- 28. Refactor!
- 29. Cases 1 or 2 double sum = 0; for (CashFlow cf : cashflows){ double df = sum += cf.getAmount() / df; } return sum; 1 + rate * t pow(1 + rate), t)
- 30. Cases 1 or 2
- 31. What business sense? What name?
- 32. Ask the experts
- 33. Oh, I see what you mean...
- 34. This is the ‘convention for interests calculation’
- 35. Thx, we can grow our Ubiquituous Language now!
- 36. Cool. ...
- 37. Next ? ...
- 38. What’s its name?
- 39. We call that the ‘coupon’, or just ‘cashflow’
- 40. Thx!
- 41. Cashflows ...
- 45. Combinations
- 46. Combinations
- 47. Combinations
- 48. Discover Domain in-depth bon R ib C o s t ume
- 49. Systematic bon R ib + C o s t ume = te k i t t y Cu
- 50. Combinations mo de l A of the dom a i n
- 51. pattern 1 Systematic Combination large variety from a few primitives
- 60. 4+4 -> 4*4
- 61. 4+4 -> 4*4
- 62. 4+4 -> 4*4 Less code Lessu bugs o w i n! Y
- 65. Technically: - Strategy pattern applied several times - A (rigid) domainspecific language
- 67. Suggestion by combination i ve ic t ed Pr el od m Ex t rap o l a te
- 68. Suggestion by combination I t ’s v a l id le ib ss o p Ye Im s, g re at ide a !
- 69. Suggestion by combination e g n l le de l a o h m C e h t
- 70. Break the model [Chris Matts] examples -> model -> more examples
- 71. Beware! t o o Do n’t b e n’t g o Do t h an e r i c sm a r t e r ge n make Mus t e nse e ss s b us i n i ne s s b us p e o p l e ...) us t a l i t t le bi t (o r j f or f un o t jus t N
- 72. And now for some more patterns...
- 73. pattern 2 Different, but not really same purpose, same thing
- 74. 1. Concrete examples 2. visualize 3. Recognize refactoring opportunities
- 75. Plain Text "On each anniversary date, the bond pays a coupon" "At the end of each period, the swap pays a cashflow"
- 76. Recognize opportunities "On each anniversary date, the bond pays a coupon "At the end of each period, the swap pays a cashflow" S am io !? e n ar e sc
- 78. Equivalentanager r a f und m Fo bond $ $ swap time time
- 79. Same purpose bond $ e n ar io ! ame s c S $ swap time time
- 80. Tip: Focus on one context ≠ ≠ =
- 81. Systematic in this context ≠ ≠ = Le g a l o ne y M n age me n t Ma
- 82. Systematic in this context an sp ort Tr gu l at io n Re ≠ = go Car S h i pp i n g
- 83. Systematic in this context Dp oDt ans D r Tr u at o n golunided ReB ≠exts Cont = go Car S h i pp i n g
- 84. pattern 3 Degenerate Case Nothing is just a special case of something
- 85. Tax calculation VAT VAT on net price Not Applicable at io n a lc ul C ?
- 86. Introduce Special Case g is a o t hi n Do -N se of ial c a spe c l at io n c a lc u
- 87. Systematic VAT VAT on net price Not Applicable at io n a lc ul C
- 88. Technically: - Null Object pattern (Fowler)
- 89. pattern 3 bis Optional Element Systematic thanks to an optional element
- 92. Systematic credit application Credit score OK? Compute Credit Limit [] References Check Record Outcome l s te p io n a Op t
- 93. pattern 3 ter Manual as special case of automatic Systematic thanks to a alternate element
- 95. Systematic ] [ Compute Credit Limit l s te p anua M References Check Record Outcome
- 96. pattern 4 Not my business Systematic by moving non-intrinsic logic away
- 97. Toys inventory 50x EU R 1. 50 100x USD 3. 25 25x U R 5. 70 M t a l v a l ue? To
- 98. e nds "De p duc t " e pro on th
- 99. e nds "De p duc t " e pro on th EUR product qty * price non-EUR product qty * price * fx rate ccy/EUR MUR, IQD... qty * price * fx rate ccy/USD * fx rate except MUR, IQD... product USD/EUR
- 100. duc t " e p ro n cy o n t h u r re e nds "De p c EUR product qty * price non-EUR product qty * price * fx rate ccy/EUR MUR, IQD... qty * price * fx rate ccy/USD * fx rate except MUR, IQD... product product USD/EUR
- 101. pe nds "De e n cy" e cur r on th EUR product non-EUR product except HKD, ZAR... HKD, ZAR... product qty * price * currency conversion
- 102. Systematic qty * price * currency conversion r re n cy Se e C u ( s i ne s s io ns b u o n ve rs C h e re ) e s e lse w r ul
- 103. pattern 5 Principles over rules Systematic at the underlying principle level
- 104. Try it yourself ar ke t upe r m S r ic i ng p K at a C o de http://codekata.pragprog.com/
- 105. .65 $0
- 106. 3 for $1 (what if I buy 4, or 5?) $1.99/pound (cost for 4 ounces?) .65 $0
- 107. Pricing by quantity decision tree 1 2 3 4 $0.65 $1.00 $1.30 $1.30 $1.65 + suggest buy 3 + suggest buy 6 .65 $0
- 108. .65 $0 Pr i n c i p le : t ’s b e s t " Wh a f or t he u s t o me r " c
- 109. .65 $0 Pr i n c i p le : t ’s b e s t " Wh a C o u ld b e the f or e r " i m p le me n te d c us t om l i te ra l l y. . .
- 110. Technically: - Declarative code (brute force) - Rules engine...
- 115. Traditionally
- 116. Specifications Describe the system we want to build
- 120. User Stories As a ne w v is i t o r I w a nt t o re g is te r I n o rde r t o re c e i ve s p am 15 p o in t s: 10
- 121. Conversations I w a nt t h is b u tto n m o re to t h e left .. .
- 122. BDD scenarios ve n I’m l o g ge d i n Gi e Logof f bu t ton W h e n I c lic k t h n I’m l o g ge d o f f The
- 125. Equations
- 126. Rules all not be ly Taxes sh troactive hanged re c
- 127. State machines
- 128. Workflows
- 129. Common situation + Gi ve n I’m log ge d in bu tt on W he n I click th e Lo go ff Th en I’m log ge d of f more details
- 132. Too much upfront
- 133. One-way
- 134. Boring & Repetitive
- 136. So How to Really improve this specs thing?
- 137. Refactor your specs
- 138. the w abou t t hi n k ay yo u Refactor your specs
- 139. The 3 amigos
- 140. Just tell me what to do
- 141. No, we ne e d yo u r s k il ls t o s im p li f y a n d s ug ge s t a b s t rac t io n s TDD, refactoring, patterns...
- 142. Ju st gi m m e exa mp le s, I’ ll reve rse it al l in TDD
- 143. No, we ne e d a t r ue b us ine s s p e rs p e c t ive DDD Ubiquitous Language
- 144. I ca n on ly he lp on ce it’s do ne
- 145. No, yo u’re even mo re use fu l alo ng th e road, wi th yo ur sk ills on ch allenging ever yt hing 3 amigos & specification workshops
- 147. IRL LOL
- 148. I ne ed sp ec s to de ri ve te st ca se s
- 149. Why no t co lla bo rate du ring th e wo rk sh op s? We’ll do th e tes t ca ses at th e same time Specifications by Example
- 150. Why no t co lla bo rate du ring th e s ce n ar io s wo rk sh op s? We’ll do th e tes t ca ses at th e same time Specifications by Example
- 152. My jo b is to so lve p ro blem s & te ll w hat to do
- 153. "90% of time BA’s bring solutions" - former head of BA’s
- 155. Yo ur jo b is to ch al le nge th e proble m in-dep th Feature Injection, Impact Mapping
- 156. If I do n’t w ri te a do cu m ent I’m useles s!
- 157. Re la x ! Th e do cume n t is a lways late to th e p a rt y
- 158. Re a l wo r k is h a pp e n ing in fo rm a ll y, w h ile we ta lk .
- 159. An d yo u a ls o h ave to faci li tate a n d e x p la in th e b usi ne ss dom a in to e ve r yo ne
- 160. Train the team
- 161. new collabor
- 162. e ar q ue n t n d f re ly a Specs -> Conversations l l 3 am ig o s b e t we e n a
- 163. e ar q ue n t n d f re ly a Specs -> conversations bou t at a u t wh B h e n? io n t n t at c ume do
- 164. Living documentation [Gojko Adzic] m p le s Exa n te n t I e n ts) c omm (Fre e
- 165. Living documentation I n te n t
- 166. Living documentation re te e xam p le s C onc
- 167. Living documentation Fre e c omme n t s
- 168. Living documentation
- 169. Visible Workings [Brian Marick] o c le t t om D C us p o rt i me e x Run t o rt. . . e x e xp L aT
- 170. Source code as reference - Annotate domain-relevant classes - Custom Doclet to export Excelformatted glossary of every domain concept
- 171. Source code as reference Sent directly to end customers
- 172. Runtime exports
- 173. Runtime exports (1+r)^{frac{t_1 - t_0}{360}} 1+r.frac{t_1 - t_0}{365} 1+rt (1+r)^{t} sum_{i=0}^{n}frac{ }{ }
- 174. Think critically
- 175. Challenge things
- 176. Refactor the code
- 177. Refactor at the domain level too
- 178. Exploit opportunities As long as it’s in line with the business domain
- 179. What about you? - Introduce Just-In-Time specification workshops with the 3 amigos - Consider Living Documentation
- 182. Merci