Pidev Eesti avaliku sektori veebisaitide hindamine WCAG 2.0 reeglite suhtes
Date
2016
Authors
Journal Title
Journal ISSN
Volume Title
Publisher
Abstract
Ligipääsetavus avaliku sektori veebilehtedele on hiljuti tunnistatud üheks eesmärgiks nii Euroopa liidu kui ka muude valitsuste poolt. Selleks, et ligipääsetavuse astet mõõta, on võetud mõõdupuuks WCAG 2.0 juhised, mille järgimise hindamine on pooleldi automatiseeritud. Kõikide kriteeriumite hindamine vajab siiski veel palju inimtööjõudu ja subjektiivsust. Inimressursi kokkuhoidmiseks piiratakse hindamine enamasti mõne leheküljeni iga domeeni kohta. Antud lõputöös püütakse parendada automatiseerimist: esiteks uuritakse inimhindajate strateegiat, teiseks analüüsitakse, kas suurema arvu lehtede hindamine ühes domeenis\n\rmõjutab lõplikku hinnangut. Eksperimentaalsed tulemused näitavad, et inimeste hindamine on lähemal pooleldi lubavale ja keelavale strateegiale.\n\rKõrgema arvu lehtede hindamine ei mõju positiivselt hindamise täpsusele võrreldes inimeste hinnangutega.
Accessibility of public sector Web sites has been recently recognized as one of the objectives of governments of EU member states and countries\n\relsewhere. In order to measure in which extent the accessibility has been achieved, WCAG 2.0 guidelines have been adopted as a benchmark measure.\n\rAlthough conformance to the guidelines has been partially automated, there is still a lot of human effort and subjectivity involved in the evaluation process. Furthermore, due to human involvement, evaluation is mostly narrowed down to a limited set of Web pages of a domain under evaluation. This study aims to make another step toward evaluation automation by 1) reverse-engineering strategies of humans evaluators and 2) analyzing whether higher number of evaluated Web pages will have positive effect to the final ranking. The experimental results show that human ranking is closer to semi-permissive and restrictive evaluation strategies. Furthermore, we show that higher number of evaluated pages will not have positive impact on the evalution precision wrt human rankings.
Accessibility of public sector Web sites has been recently recognized as one of the objectives of governments of EU member states and countries\n\relsewhere. In order to measure in which extent the accessibility has been achieved, WCAG 2.0 guidelines have been adopted as a benchmark measure.\n\rAlthough conformance to the guidelines has been partially automated, there is still a lot of human effort and subjectivity involved in the evaluation process. Furthermore, due to human involvement, evaluation is mostly narrowed down to a limited set of Web pages of a domain under evaluation. This study aims to make another step toward evaluation automation by 1) reverse-engineering strategies of humans evaluators and 2) analyzing whether higher number of evaluated Web pages will have positive effect to the final ranking. The experimental results show that human ranking is closer to semi-permissive and restrictive evaluation strategies. Furthermore, we show that higher number of evaluated pages will not have positive impact on the evalution precision wrt human rankings.