3. About Speaker
Jonathan Hikita
•GM, CData Software Japan LLC
Career summary
•15+ yrs of utilizing enterprise data
•10 yrs in Financial (JPN, Indonesia etc)
•Global Operation of Dev Component vendor
•Establishing CData Software’s Japan operation
•Love KARATE and Heavy Metal
@cdatajapan
@keisuke.hikita.5
https://qiita.com/jonathanh
6. About CData Software
Bi-directional Access to Live App, Database, & Web API Data Through Standard Drivers
・CData Software, Inc. / Started: 1994 (/nsoftware)
・Location: Chapel Hill, NC a spin-off of /n software
・CData Japan: 2016/6 (JV with Infoteria)
・20+ years of providing data-related development components
・100+ supported data sources
・”See the World as a Database”
7. 1 Chome-6-27 Chuo, Aoba Ward,
Sendai, Miyagi Prefecture
980-0021, Japan
Tel: 050-5578-7390
CData Japan
101 Europa Dr. #110
Chapel Hill, NC 27517 USA
Tel: (919) 928-5214
Fax: (919) 928-5455
US Headquarters
- Central & Eastern Europe
- Central China
Additional Development Offices
Global and Japan Local Operations
Worldwide Offices for Global Sales and Support
8. Largest enterprise data source coverage
Drivers for NoSQL, Big Data, & SaaS Connectivity
CRM & Marketing
Accounting
Collaboration & ERP
DB
Documents and file formats
SNSNetworking and protocols
EC and fintech
Others
9. CData standardize data in RDB ⇔ Web APIs
Enable real-time data integration with hundreds of applications, databases, and Web APIs
Data Drivers
Makes Web API accssible from
JDBC/ODBC/ADO standard
interface
API Server
Easiest way to create
professional REST APIs from DBs
13. What’s inside CData Drivers?
Give schema like RDB, enable standard SQL, give enterprise level security features
• JSON/XML data is
mapped to tables
• Auto-detects schema
from unstructured
data
• Enable SQL-92
• Full CRUD
• JOIN, filtering,
aggregation etc.
• Makes it easier from
application to access
• More sophysticated
framework use.
• Firewall, proxy, other
security
requirements
• Authentication, log
and other
management.
Table modeling
(schema detection)
SQL engine Standard Interface Security etc.
ODBC
15. OEM Partners-by Global #1 players,
by Japan’s #1 players
Embedded In The Leading BI, ETL, Data Integration, Data Virtualization, and Data Warehousing Tools
17. Target audience is non-API first services
• To make the base service users happy ---- YES
• To acquire new channels ---- YES
• API is the original way of using the service ---- NO
22. API use requires paper agreements
• NDA, agreement
• Worst was 10 pages checklist on
companies size, financials,
security, etc.
• Why require months to try APIs
23. API reference in PDF・Excel
• Excels, PDFs, Words
• Sometimes even CD-ROM
• How can you expect users to
use the latest API docs?
24. Crazy API Limit
• Sometimes, API rate limits hit with
just 1 complex request
• I often panic right before my
session when this happens
25. POST, POST, POST for everything
• There is only POST methods
• Retrieve, Delete, Insert, Update,
all are done with POST putting
parameters in Body
• Often has different URI for each
types of calls
26. No endpoint list
• This means that humans need to
refer to API docs
• No dynamic schema detection
• SCHEMA is what applications
need.
27. Why respond with “¥” mark?
・Currency fields sometimes respond with
strings with “yen” mark
・Stop making everything “strings”
・So many confusions on data types
28. Give summary, no records
・Can retrieve total, but no details
・Sometimes, APIs refused to give
details but only aggregated data. At the
same time you can POST individual data.
29. Mix of Resource oriented & Function oriented
・REST APIs supposed to be resource oriented.
However, many verbs in URIs.
Ex. https://XXX.com/getXXX
・Find, Search are often seen, too.
30. Allow POST only
• Can only do POST and CSV uploads.
• User cannot retrieve their data
through APIs
• This limits API use very much
38. Remember, your users are developers now
Underlying service
users
Non-Engineer
API users
Engineers
• Browser use
• Routine work
• Look and input
• manual
• Top-down decision making入
• Trust, stability
• Code
• Create new ways of usage
• Read schema and let machine decide
• AUTOMATION
• Fast evaluation cycle
• Change, innovation
39. Persona is different, redefine UX to DX
UX
User Experience
→
DX
(Developer Experience)
41. This gravity is very, very strong in Japan
• F2F sales oriented culture
• Large reliance on consultation and all-in services
• Tokyo-centric economy
42. How to get out from old UX
• API PM (if cannot, evangelist/advocate need to play some roles)
• API UI
• API portal (Dev portal) on web
• API documents (ideally auto-generated and match with meta-data)
• SEO
• SDK
• Drivers
• Others
• Give trials for API users
• No paper, no f2f unless requested
• Support that can handle technical questions
43. Message:
Just knowing this makes things different:
Gravity from underlying services exists
API UX can be and should be different from underlying service UX