1 Cyberinfrastructure in - PowerPoint Presentation

Slide1

1

Cyberinfrastructure

in practice

Donald Sturgeon

Harvard Universitysturgeon@fas.harvard.eduSlide2

OverviewInfrastructureWhat it is and why

it mattersApplication Programming Interfaces (APIs)Simple examples in widespread use

Current examples in Chinese studiesChinese Text Project (ctext.org

) APIQuestions and challenges going forwardWhat infrastructure do we needHow are we going to maintain it2Slide3

3

Infrastructure

I

s

instrumentalNot an end but a means to an endMakes the possible (but difficult) easierMakes the impractical practicalRequires up-front investmentRequires ongoing maintenanceSlide4

Infrastructure in the humanities4

Electricity, hardware, etc.

“Real-world” infrastructure

General-purpose operating systems

Low-level software infrastructure

Code LibrariesGeneral-purpose infrastructures?Special-purposeresearch infrastructuresServices

Humanities research tools

Domain-specific tools,

databases, etc.

?

Software

Standards

ctext

API

CBDB API

CTSSlide5

Types of humanities infrastructureData formats

Text Encoding Initiative (TEI)Application Programming Interfaces (APIs)IIIF Image APIOpen source software and libraries

Stanford CoreNLPServicesUniversity Library provision of IIIF-compliant data

5Slide6

Application Programming InterfacesWhat do they do?

Predictable mechanisms for data exchangeIncreasingly: web APIsWhat are they useful

for?Making materials/services available in consistent wayAbstraction from implementation detailsAllowing the creation of “derived products”“Mashups”

consisting of independently maintained partsMining and analysis of data“Offloading” part of the development process

6Slide7

Google Maps

7

Google Maps

user interfaceAllows access to Google-defined servicesGoogle Maps API

Allows building upon Google’s map servicesCreate things Google alone would never createSlide8

Google Maps + Housing Rental Data

8

Distribution

of effort:Google:

has no control over

apartment datamaintains map data and map interfaceRental search company:has limited control over map data & interfacemaintains apartment rental informationResulting mashup is nevertheless a

cohesive product

Works as

if created by

a

single

group

In

fact

maintained

by

two

independent

groupsSlide9

Google Maps + Disease Data

9

Economies

of scale:Google:

concentrates on one thing

only: mapsOther groups:concentrate on their own contentbenefit from centralization of map-related code & dataLower barriers to entry for subsequent projectsNot a closed, one-off collaborationInstead: an open invitation to others to collaborateSlide10

APIs in

practice: ctext API

10Slide11

APIs in practice:

ctext

API

11

ctext

, MARKUS, Text Tools, etc. all communicate via public APIEveryone has access to the API and its documentationLowers barriers to entry for subsequent projectsNot a closed, one-off collaborationInstead: an open invitation to others to collaborateAnyone can create and distribute a new API client / pluginSlide12

APIs in the humanities

How

to create economies

of scaleWhat areas benefit from standardizationWhat areas benefit

from decentralizationHow to let

projects concentrate on core workDigital projects increasingly complexNo single team can expect to do everything wellWhat infrastructure is most urgently neededWhich components have greatest reuse potential12Slide13

Cyberinfrastructure: challengesComponents need to be maintained over time

How will this be guaranteed institutionallyStandardization is beneficial but not easyHumanities data is complex

TEI is one example demonstrating this complexityCyberinfrastructure needs coordinationNot just a function of a single groupMany stakeholders

Data creators, disseminators, consumers, end users, etc.13Slide14

14

Thank you!