Upgrade to Pro
— share decks privately, control downloads, hide ads and more …
Speaker Deck
Features
Speaker Deck
PRO
Sign in
Sign up for free
Search
Search
The advantage of using ’Eff’ in Scala Project
Search
machu
June 28, 2019
Programming
2
13k
The advantage of using ’Eff’ in Scala Project
machu
June 28, 2019
Tweet
Share
More Decks by machu
See All by machu
NBAチームから学ぶ強いチームの作り方
machuz
0
48
Authorization to implement with Extensible Effect
machuz
0
440
アルプの 認証/認可分離戦略と手法
machuz
3
770
AuthzCtx - Alp社内共有会
machuz
0
81
アルプのEff独自エフェクト集 / Alp-original ’Eff’ pearls
machuz
1
2.2k
Scalebaseバックエンド構成について/the backend design of Scalebase
machuz
0
6.5k
SQL Meisterへの道 ~更新編~ / sql-meister-CUD
machuz
0
2.3k
SQL Meisterへの道 ~基礎〜参照編~ / sql-meister-R
machuz
0
2.7k
Authz
machuz
0
310
Other Decks in Programming
See All in Programming
Rubyで鍛える仕組み化プロヂュース力
muryoimpl
0
220
認証・認可の基本を学ぼう後編
kouyuume
0
250
Java 25, Nuevas características
czelabueno
0
120
生成AIを利用するだけでなく、投資できる組織へ
pospome
2
420
Tinkerbellから学ぶ、Podで DHCPをリッスンする手法
tomokon
0
150
公共交通オープンデータ × モバイルUX 複雑な運行情報を 『直感』に変換する技術
tinykitten
PRO
0
170
DevFest Android in Korea 2025 - 개발자 커뮤니티를 통해 얻는 가치
wisemuji
0
180
GoLab2025 Recap
kuro_kurorrr
0
780
AIコーディングエージェント(Gemini)
kondai24
0
300
Go コードベースの構成と AI コンテキスト定義
andpad
0
150
AI 駆動開発ライフサイクル(AI-DLC):ソフトウェアエンジニアリングの再構築 / AI-DLC Introduction
kanamasa
11
4.5k
AIの誤りが許されない業務システムにおいて“信頼されるAI” を目指す / building-trusted-ai-systems
yuya4
6
4k
Featured
See All Featured
Breaking role norms: Why Content Design is so much more than writing copy - Taylor Woolridge
uxyall
0
120
Getting science done with accelerated Python computing platforms
jacobtomlinson
0
82
コードの90%をAIが書く世界で何が待っているのか / What awaits us in a world where 90% of the code is written by AI
rkaga
58
41k
How Fast Is Fast Enough? [PerfNow 2025]
tammyeverts
3
410
The Impact of AI in SEO - AI Overviews June 2024 Edition
aleyda
5
680
Navigating Algorithm Shifts & AI Overviews - #SMXNext
aleyda
0
1k
End of SEO as We Know It (SMX Advanced Version)
ipullrank
2
3.8k
Reality Check: Gamification 10 Years Later
codingconduct
0
2k
GraphQLの誤解/rethinking-graphql
sonatard
74
11k
<Decoding/> the Language of Devs - We Love SEO 2024
nikkihalliwell
0
100
Let's Do A Bunch of Simple Stuff to Make Websites Faster
chriscoyier
508
140k
Building Adaptive Systems
keathley
44
2.9k
Transcript
The advantage of using ’Eff’ in Scala Project Scala Matsuri
2019 @ma2k8 4DBMBϓϩδΣΫτͰ&⒎Λ༻͢Δར
About Me ▶Tsubasa Matsukawa - SWE&SRE at Recruit Marketing Partners
Co.,Ltd. & Quipper Co.,Ltd. @wing_007 ϦΫϧʔτϚʔέςΟϯάύʔτφʔζ݉2VJQQFSॴଐͰ 48&ͱ43&ྖҬΛ୲͍ͯ͠·͢ɻ
Our Products ▶Study Sapuri English - English learning app [
Web/iOS/Android ] ओʹελσΟαϓϦ&OHMJTIͷ։ൃΛ͍ͯ͠·͢ɻ ӳޠֶशʹڵຯ͕͋Δํੋඇ
Scala has an elegant syntactic sugar ‘for expression’. To make
the most of this force…. 4DBMBʹૉఢͳҥߏจGPS͕ࣜ͋Γ·͢ɻ ͦͷྗΛ࠷େݶར༻͢ΔͨΊʹɾɾɾ
We needed to decided which ‘types’ to primarily use to
when starting Scala projects. ࢲୡ4DBMBͷϓϩδΣΫτΛ։࢝͢ΔࡍɺओʹͲͷʮܕʯΛ ༻͢Δͷ͔ΛܾΊΔඞཁ͕͋Γ·͢ɻ
These ‘types’ are often chosen. ▶Future[A] ▶Future[Try[A]] ▶Future[E Either A]
▶EitherT[Future, E, A] ※There is also case that uses "Task" instead of "Future" ͜ͷΑ͏ͳܕ͕Α͘બΕΔͰ͠ΐ͏ɻ 'VUVSFͷΘΓʹ5BTLΛ͏͜ͱ͋Γ·͢
Our team is trying…. Eff[R, A] ࢲୡͷνʔϜͰ&⒎Λࢼ͍ͯ͠·͢
Each pros and cons were…? ͦΕͧΕͷϝϦοτσϝϦοτʁ
Future[A]
▶ Easy ॴ؆୯͞ʹਚ͖·͢ɻ Pros of Future[A]
Cons of Future[A] ▶ Rough error handling. ▶ Semaphore hard.
※Task solve Semaphore problem. ॴηϚϑΥͷ੍ޚ͕͍͠ͱɺ ΤϥʔϋϯυϦϯά͕େࡶʹ͔͠ߦ͑ͳ͍Ͱ͢ɻ
Future[Try[A]] ɹ
Pros of Future[Try[A]] ▶ Error handling is possible. ॴΤϥʔϋϯυϦϯά͕ՄೳʹͳΔͰ͢ɻ
Cons of Future[Try[A]] ▶ Only throwable can be used ▶
Syntax is not simple(Coding 'try ~ catch' every time is very troublesome. ॴΤϥʔϋϯυϦϯάʹ༻Ͱ͖Δܕ͕ 5ISPXBCMFʹݶΔͱγϯλοΫε͕ෳࡶʹͳΔͰ͢ɻ
Personally, I often use ‘Try’ when wrapping Java libraries, but
since I convert to ‘Either’, they do not appear in signatures. ݸਓతʹ5SZ+BWBϥΠϒϥϦͷ8SBQͰΑ͍͘·͕͢ɺ &JUIFSܕʹม͠ɺγάχνϟʹग़͠·ͤΜɻ
Future[E Either A]
Pros of Future[E Either A] ▶ Flexible Error handling is
possible. ▶ E is ADT can be flexible expression. ▶ Every time force error handling. ॴ5SZΑΓॊೈͳΤϥʔϋϯυϦϯά͕ ՄೳʹͳΔͰ͢ɻ
Cons of Future[E Either A] ▶ Rough error handling. ▶
Every time need to match the type even if you don't need error handling. ॴΤϥʔϋϯυϦϯά͕ ෆཁͳՕॴͰܕ߹Θ͕ͤඞཁʹͳΔͰ͢ɻ
EitherT[Future, E, A]
Pros of EitherT[Future, E, A] ▶ Can solve part of
the complexity of nesting. ॴ͋Δఔωετͯ͠ GPSࣜΛγϯϓϧʹอͯΔͰ͢ɻ
Cons of EitherT[Future, E, A] ▶ It gets slower as
types are stacked. ▶ May change results in evaluation order. ▶ Too many lift. ॴɺධՁॱͰ݁Ռ͕มΘΔՄೳੑ͕͋ΓɺϞφυ ελοΫʹ٧Ήܕ͕૿͑Δͱɺܕ߹Θ͕ͤࡶʹͳΔͰ͢ɻ
Eff[R, A]
Pros of Eff[R, A] ▶ ‘Eff’ can solve other ‘types’
disadvantage. ▶ Simple syntax. ▶ Can simply use various effects. ▶ Can solved performance problem when types are stacked. ▶ A result doesn't change for the order of evaluation. &⒎͜Ε·Ͱհͨ͠ܕͷσϝϦοτͷ େ෦ΛղফͰ͖·͢ɻ
Cons of Eff[R, A] ▶’Eff’ is very simple syntax. But
not easy. ▶ difficult effects wire. ▶ Anything is lazy evaluation. ▶ Complexity of type specification. &⒎ͷσϝϦοτɺ؆୯Ͱͳ͍Ͱ͢ɻ ͋Δఔͷ׳ΕඞཁͰ͠ΐ͏ɻ
Our project used ‘Eff’ on clean architecture. զʑͷϓϩδΣΫτͰɺ$MFBO"SDIJUFDUVSFͰઃܭ͠ɺ Ҿ͖ճ͢ܕʹ&⒎Λ༻͢ΔߏΛಋೖ͍ͯ͠·͢ɻ
͜ͷΑ͏ʹෳͷޮՌΛϑϥοτʹهड़Ͱ͖ΔͷͰɺ γϯϓϧʹهड़Ͱ͖͍ͯ·͢ɻ
How to use ‘Eff’ &⒎ͷ͍ํ Library: https://github.com/atnos-org/eff
Define ADT "%5Λఆٛ͠·͢ɻ
Define Smart-constructer "%5Λ&⒎ʹੵΉεϚʔτίϯετϥΫλΛఆٛ͠·͢ɻ
ੵΜͩ"%5Λղऍ͢ΔΠϯλϓϦλΛఆٛ͠·͢ɻ Define Interpreter
Ready to use ‘Eff’. ͜ΕͰ&⒎Λར༻͢Δ४උ͕Ͱ͖·ͨ͠ɻ
&⒎Λར༻ͨ͠ϓϩάϥϜΛॻ͖·͢ɻ Write program!!
ΠϯλϓϦλʹ͔͚ͯ&⒎Λ࣮ߦ͠·͢ɻ SVO99ͷؔͦΕͧΕJNQMJDJUDPOWFSTJPOͰੜͯ͠·͢ɻ Run program!!
‘Eff’ is Good! &⒎͍͍ͧʂʢհهࣄͱίʔυαϯϓϧష͓͖ͬͯ·͢ Introduction article: https://tech.recruit-mp.co.jp/server-side/post-18728/ɹ Code sample: https://github.com/ma2k8/eff-arch-example
Thanks! ͋Γ͕ͱ͏͍͟͝·ͨ͠ʂ