A Cross-Domain Collaborative Website Accessibility Crowdsourcing Evaluation Method

Abstract

A cross - domain collaborative website accessibility crowdsourcing evaluation method includes: First, the requester initiates an evaluation task to the system. After the task is started, the server automatically crawls the site in batches, extracts web page information and locates detection points according to the accessibility standard items required by the requester, and batch - generates website accessibility crowdsourcing tasks corresponding to the items. Then, disabled or able - bodied users can access the web pages to be tested and the system result submission page across domains on a single interface by installing the browser plug - in supporting the system to perform crowdsourcing tasks. Finally, the system integrates all evaluation results and outputs a task report for the requester to review. The present invention realizes the automatic batch crawling of website information, locates and extracts accessibility information according to the requirements of relevant standard items, and thus combines the extraction results with the task evaluation requirements to pre - generate crowdsourcing tasks related to accessibility items.

Description

[0001] Technical Field: The present invention relates to a cross-domain collaborative website accessibility crowdsourcing evaluation method, belonging to the field of computer technology software. Background Art:

[0002] Among the Internet user group, in addition to able-bodied people, there are also some special groups with information acquisition barriers, including visually impaired people, hearing impaired people, and the elderly. According to the estimation of the World Health Organization, there are approximately 285 million people with various degrees of visual impairment globally, among which about 39 million are completely blind. Additionally, there are approximately 466 million people suffering from hearing loss, accounting for about 3.7% and 6.1% of the global population respectively. Due to physical defects and disabilities, these people have barriers in information acquisition and interaction. How to enable special groups to benefit from information technology and access and utilize information equally, conveniently, and without barriers is an urgent requirement in the current information era. Inaccessible assistance technologies mainly include the built-in accessibility assistance tools of computer systems (such as "Narrator" for Windows and "VoiceOver" for MacOS) and other third-party screen readers. These software can provide some functions for disabled people, such as accessing via keyboard shortcuts and reading text aloud, to assist in perceiving and understanding information.

[0003] However, for a web page to support good access by accessibility assistance tools, it needs to conform to certain accessibility development specifications. In actual product development, due to the lack of understanding of accessibility development specifications by website designers and developers, most websites do not provide good accessibility support, which causes difficulties for the vast number of disabled people in accessing websites without barriers. Although, according to accessibility standards such as "Technical Requirements and Test Methods for Internet Content Accessibility of Information Technology" (GB / T 37668—2019) or Web Content Accessibility Guidelines (WCAG) 2.1, some accessibility items can be automatically detected for compliance through programming, but most items, especially high-level accessibility standard items, have characteristics such as strong interaction and semanticization, and are difficult to be automatically detected by ordinary technical means. Therefore, it is still of great significance to build a manual crowdsourcing evaluation system for website accessibility compliance. The main difficulties in building such a system are as follows: (1) How to achieve scientific website automatic crawling and information parsing and extraction, so as to pre-generate crowdsourcing tasks related to the requirements of accessibility items; (2) How to support cross-domain compatible access between the web page to be tested and the evaluation result submission page on a single interface in the system; (3) How to support disabled people to participate in crowdsourcing evaluation. Summary of the Invention:

[0004] In view of the above problems and difficulties, the present invention proposes a cross-domain collaborative website accessibility crowdsourcing evaluation method.

[0005] The specific functional steps of a cross - domain collaborative website accessibility crowdsourcing evaluation method are as follows:

[0006] S101: The evaluation requester initiates an evaluation task on the system, specifies the evaluation configuration, and designates the scope of accessible standard items to be evaluated;

[0007] S102: After receiving the evaluation task, the system server starts the automatic batch crawling of the site page according to the configuration;

[0008] S103: During the site page crawling process, the system server dynamically locates and extracts the occurrence situations and information of relevant item detection points in the web page according to the accessible standard item requirements configured by the requester;

[0009] S104: The system server batch - generates website accessibility crowdsourcing tasks corresponding one - by - one to the items based on the extraction situations of the relevant item detection points;

[0010] S105: Disabled or able - bodied users install the crowdsourcing browser plugin downloaded in the system in the browser client, open this plugin in the developer interface and log in to the system account, apply for ordinary crowdsourcing tasks or review crowdsourcing tasks, and perform tasks in both the developer interface and the browser main tab interface, bypassing the different - domain problem between the page to be tested and the result submission page;

[0011] S106: After receiving the crowdsourcing task application request, the system server randomly distributes the pre - generated crowdsourcing tasks to the crowdsourcing evaluators;

[0012] S107: After receiving the task, the disabled or able - bodied user's client plugin automatically opens the corresponding task web page in the browser main interface tab according to the crowdsourcing task information, and displays the information of this crowdsourcing task, the corresponding standard item description, and the evaluation method in the developer interface;

[0013] S108: The disabled or able - bodied user switches the focus to the developer interface, obtains the evaluation item information and evaluation method by reading aloud through the accessibility assistive tool, then switches the focus to the browser main interface tab web page, and accesses and evaluates the web page through the accessibility assistive tool;

[0014] S109: The disabled or able - bodied user switches the focus to the developer interface and submits the evaluation result of the measured web page or gives up;

[0015] S110: If the disabled or able - bodied user encounters problems when accessing the page to be tested, reports the web page as an error page, and automatically applies for the next crowdsourcing task;

[0016] S111: After receiving the results of all crowdsourcing tasks of a certain evaluation task, the system server integrates them to form a report for the evaluation requester to review.

[0017] Specifically, the evaluation task information and configuration initiated by the evaluation demander on the system in step S101 include: S201: evaluation website name, evaluation application organization, evaluation website homepage URL, evaluation rule scope, evaluation website login account, login password and test description, etc. The test login account, test login password and test description are optional.

[0018] Specifically, step S102, wherein the method for automatically crawling site pages in batches is as follows: S301: Add the homepage URL of the site to be tested to the queue of web page URLs to be tested, (1) take out a URL from the queue and open this page. If it is the homepage, then (2) extract all URLs in the DOM of the opened page that are in the same domain as the website required by the evaluation demander and add them to the queue to be traversed, save this page DOM, and go to (1); otherwise, (3) calculate the PQ-Gram edit distance between this page DOM and all previously saved web page DOMs, if there is a distance less than the set threshold α, filter this page without doing anything, repeat (1) and subsequent steps; if the distances are all greater than the threshold α, repeat (2) and subsequent steps. When the number of crawled web pages reaches the number configured by the demander, or the depth of the crawled web pages reaches the depth configured by the demander, or the URL queue is empty, the crawling ends.

[0019] Specifically, in step S103, the demand side configures the accessibility standard items in the form of: S401: Accessibility standard items, that is, the scope of the test rules needs to be determined by independently checking several items of accessibility standards such as "Technical Requirements and Test Methods for Accessibility of Information Technology Internet Content" (GB / T 37668-2019) or Web Content Accessibility Guidelines (WCAG) 2.1 according to the needs;

[0020] Specifically, in step S104, the information of the website accessibility evaluation task generated corresponding to each item is as follows: S501: the evaluation page URL with the relevant detection points of the item, the evaluation item, and the evaluation method;

[0021] Specifically, in step S105, the definitions of ordinary tasks and review crowdsourcing tasks are as follows: S601: The system generates crowdsourcing tasks directly corresponding to the number of people required to conduct crowdsourcing for a single webpage and a single item configured by the demander, which is called an ordinary crowdsourcing task. When the number of crowdsourcing people is an even number, and the number of people who pass and fail in the final submission is equal, the task automatically generates a review crowdsourcing task, which is reviewed by an evaluator with higher review authority;

[0022] Specifically for step S105, the specific method to bypass the problem of different domains between the web page to be tested and the system's result submission page is as follows: S701: By viewing information and submitting results of the evaluation task in the developer interface, while accessing the page to be tested in the main tab of the browser. The access to both pages is independent, without cross-domain issues, but they are in the same browser interface, achieving the access to two pages in the same interface, which is convenient for evaluators to carry out crowdsourcing tasks;

[0023] Specifically for step S106, the management method of the system server for distributing and assigning evaluation tasks is as follows: S801: The requester specifies the number of crowdworkers for each crowdsourcing task when initiating the evaluation task. When the system server receives a crowdsourcing task application, it randomly selects a crowdsourcing task that has not reached the upper limit of the assigned number of crowdworkers and distributes it. At the same time, the assigned number of this crowdsourcing task is incremented by one. Within the timeout threshold of β seconds, this crowdsourcing task belongs to the evaluator who received this distribution and cannot be distributed repeatedly. This evaluator can submit the results normally; if the threshold is exceeded, this distribution times out and is invalidated, the assigned number of this crowdsourcing task is decremented by one, waiting for re-distribution. If this evaluator submits, a task timeout prompt will be given, and the evaluator needs to re-apply for a crowdsourcing task;

[0024] Specifically for step S107, the interactive functions in the developer interface include: S901: Switch the task to a normal task or a review task, personal center entry, logout button, view detection item information, submit evaluation results, abandon this task, report an error page;

[0025] Specifically for step S107, the crowdsourcing task information, its standard item descriptions, and evaluation methods displayed in the developer interface include: S1001: Title of the web page to be tested, title of the detection rule, rule description, detection method, passing conditions;

[0026] Specifically for step S109, the evaluation results of the web page to be tested submitted by disabled or able-bodied user evaluators include: S1101: Whether this web page to be tested passes the tested accessibility items. If it passes, submit directly; if it does not pass, fill in the reason for non-passing and then submit;

[0027] Specifically for step S109, after a disabled or able-bodied user evaluator abandons a crowdsourcing task, the processing method of the system server for this task is as follows: S1201: The assigned number of this crowdsourcing task is decremented by one, waiting for re-distribution;

[0028] Specifically for step S110, the conditions under which a disabled or able-bodied user can report an error page are as follows: S1301: This page is inconsistent with the page to be tested, or this page is a garbled page, or this page fails to open normally, with error codes such as 404 or 500 being prompted;

[0029] Specifically, for step S111, the criteria for the report content and items generated by the server are as follows: S1401: The report content includes the levels, content, and passing status of each standard item, sample error web pages, and a consolidated errata list of sub-standard items for all error web pages. If, among the reviewers, the number of people who consider the item to pass on the web page is more than those who consider it to fail, then the item passes on that web page; otherwise, it fails. If the number of people is equal, the single review result shall prevail;

[0030] Specifically, for step S111, the report formats that can be downloaded by the experience and evaluation task requester include: S1501: PDF format document or Word format document.

[0031] In summary, the present invention creates a cross-domain collaborative website accessibility crowdsourcing evaluation method, which has the following beneficial effects: (1) It realizes scientific automatic website crawling and information parsing and extraction, thus pre-generating crowdsourcing tasks related to accessibility item requirements; (2) It supports cross-domain compatible access to the web pages to be tested and the evaluation result submission page on a single interface in the system; (3) The system design supports the participation of disabled persons in crowdsourcing evaluation. BRIEF DESCRIPTION OF THE DRAWINGS:

[0032] In order to more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the following will briefly introduce the drawings required for use in the description of the embodiments or the prior art. Obviously, the following drawings are only some embodiments of the present invention. For those of ordinary skill in the art, without creative efforts, other drawings can be obtained based on these drawings.

[0033] Figure 1 It is a workflow schematic diagram of a cross-domain collaborative website accessibility crowdsourcing evaluation method provided by the present invention;

[0034] Figure 2 It is a schematic diagram of the crowdsourcing task interface of a cross-domain collaborative website accessibility crowdsourcing evaluation method provided by the present invention; Specific implementation method:

[0035] The following will describe the exemplary embodiments of the present disclosure in more detail with reference to the drawings. Although the exemplary embodiments of the present disclosure are shown in the drawings, it should be understood that the present disclosure can be implemented in various forms and should not be limited by the embodiments set forth herein. On the contrary, these embodiments are provided so that the present disclosure can be more thoroughly understood and the scope of the present disclosure can be fully conveyed to those skilled in the art.

[0036] Taking the accessibility compliance evaluation task of a certain website as an example, a cross-domain collaborative website accessibility crowdsourcing evaluation method of the present invention includes the following specific steps:

[0037] S101: The evaluation requester initiates an evaluation task on the system, specifies the evaluation configuration and the scope of accessible standard items to be evaluated.

[0038] S102: After receiving the evaluation task, the system server starts the automatic batch crawling of the site page according to the configuration.

[0039] S103: During the site page crawling process, the system server dynamically locates and extracts the occurrence situations and information of relevant item detection points in the web page according to the accessible standard item requirements configured by the requester.

[0040] S104: Based on the extraction situations of relevant item detection points, the system server batch generates website accessibility crowdsourcing tasks corresponding to the items one by one.

[0041] S105: Disabled or able-bodied users install the crowdsourcing browser plugin downloaded in the system in the browser client, open this plugin in the developer interface and log in to the system account, apply for ordinary crowdsourcing tasks or review crowdsourcing tasks, and perform tasks in the developer interface and the main tab interface of the browser to bypass the problem of different domains between the page to be tested and the result submission page.

[0042] S106: After receiving the crowdsourcing task application request, the system server randomly distributes the pre-generated crowdsourcing tasks to the crowdsourcing evaluators.

[0043] S107: After receiving the task, the disabled or able-bodied user client plugin automatically opens the corresponding task web page in the main interface tab of the browser according to the crowdsourcing task information, and displays the information of this crowdsourcing task and the corresponding standard item description and evaluation method in the developer interface.

[0044] S108: The disabled or able-bodied user switches the focus to the developer interface, obtains the evaluation item information and evaluation method by reading aloud through the accessibility assistance tool, then switches the focus to the web page in the main tab interface of the browser, and accesses and evaluates the web page through the accessibility assistance tool.

[0045] S109: The disabled or able-bodied user switches the focus to the developer interface and submits the evaluation result of the measured web page or gives up.

[0046] S110: If the disabled or able-bodied user encounters a problem when accessing the page to be tested, the user reports that the web page is an error page and automatically applies for the next crowdsourcing task.

[0047] S111: After receiving the results of all crowdsourcing tasks of a certain evaluation task, the system server integrates them to form a report for the evaluation requester to review.

[0048] The present invention discloses a cross - domain collaborative website accessibility crowdsourcing evaluation system and method. The method realizes the automatic batch crawling of website information, locates and extracts accessibility information according to the requirements of relevant standard items, and thus combines the extraction results with the task evaluation requirements to pre - generate crowdsourcing tasks related to accessibility items. The system and method also implement cross - domain access support for the web pages to be tested and the system result submission page on a single interface in the form of a browser plugin, and build a complete barrier - free access system to support disabled persons to participate in crowdsourcing evaluation. Generally speaking, the entire system and method include the following functional steps: First, the requester initiates an evaluation task to the system. After the task is started, the server automatically batch - crawls the site, extracts web page information and locates detection points according to the accessibility standard items required by the requester, and batch - generates website accessibility crowdsourcing tasks corresponding to the items. Then, disabled or able - bodied users can realize cross - domain access to the web pages to be tested and the system result submission page on a single interface by installing the browser plugin supporting the system to perform crowdsourcing tasks. Finally, the system integrates all evaluation results and outputs a task report for the requester to review.

Claims

1. A cross - domain collaborative website accessibility crowdsourcing evaluation method, characterized in that It includes the following functions: S101: The evaluation requester initiates an evaluation task on the system, specifies the evaluation configuration and the range of accessible standard items to be evaluated; S102: After receiving the evaluation task, the system server starts the automatic batch crawling of the site page according to the configuration; S103: During the site page crawling process, the system server dynamically locates and extracts the occurrence status and information of relevant item detection points in the web page according to the accessible standard item requirements configured by the requester; S104: The system server batch generates website accessibility crowdsourcing tasks corresponding one by one to the items based on the extraction status of the relevant item detection points; S105: Disabled or able-bodied users install the crowdsourcing browser plugin downloaded in the system in the browser client, open this plugin in the developer interface and log in to the system account, apply for ordinary crowdsourcing tasks or review crowdsourcing tasks, and perform tasks in the developer interface and the browser main tab interface at the same time, bypassing the problem of different domains between the page to be tested and the result submission page; S106: After receiving the crowdsourcing task application request, the system server randomly distributes the pre-generated crowdsourcing tasks to the crowdsourcing evaluators; S107: After receiving the task, the client plugin of the disabled or able-bodied user automatically opens the corresponding task web page in the browser main interface tab according to the crowdsourcing task information, and displays the information of this crowdsourcing task, the corresponding standard item description and the evaluation method in the developer interface; S108: The disabled or able-bodied user switches the focus to the developer interface, obtains the evaluation item information and evaluation method by reading aloud through the accessibility assistive tool, then switches the focus to the browser main interface tab web page, and accesses and evaluates the web page through the accessibility assistive tool; S109: The disabled or able-bodied user switches the focus to the developer interface and submits the evaluation result of the measured web page or gives up; S110: If the disabled or able-bodied user encounters a problem when accessing the page to be tested, reports that the web page is an error page and automatically applies for the next crowdsourcing task; S111: After receiving the results of all crowdsourcing tasks of a certain evaluation task, the system server integrates and forms a report for the evaluation requester to consult.

2. The cross-domain collaborative website accessibility crowdsourcing evaluation method according to claim 1, wherein: In step S101, the evaluation task information and configuration initiated by the evaluation requester on the system include: evaluation website name, evaluation application institution, evaluation website home page URL, evaluation rule range, evaluation website login account, login password and test description; the test login account, test login password and test description are optional items.

3. The cross-domain collaborative website accessibility crowdsourcing evaluation method according to claim 1, characterized in that: The step S102, where the method for automatically batch crawling the site page is as follows: Add the URL of the home page of the site to be tested to the URL queue of the web pages to be tested. (1) Take out a URL from the queue and open this page. If it is the home page, then (2) extract all the URLs in the DOM of the opened page that are in the same domain as the website required by the evaluation requester and add them to the queue to be traversed, save the DOM of this page, and go to (1). Otherwise, (3) calculate the PQ-Gram edit distance between the DOM of this page and all the previously saved web page DOMs. If there is a distance less than the set threshold α, then filter this page without any operation, and repeat (1) and the subsequent steps. If the distances are all greater than the threshold α, then repeat (2) and the subsequent steps. When the number of crawled web pages reaches the number configured by the requester, or the depth of the crawled web pages reaches the depth configured by the requester, or the URL queue is empty, the crawling ends.

4. A cross - domain collaborative website accessibility crowdsourcing evaluation method according to claim 1, characterized in that: In the step S103, the form in which the requester configures the accessibility standard items is as follows: The accessibility standard items, that is, the scope of the test rules needs to be determined by independently checking several items of the accessibility standards according to the requirements. In the step S104, the information of the website accessibility evaluation tasks generated corresponding to the items one by one is as follows: The URL of the evaluation page with the detection points related to this item, the evaluation item, and the evaluation method.

5. The cross-domain collaborative website accessibility crowdsourcing evaluation method according to claim 1, wherein: The definitions of the ordinary crowdsourcing task and the review crowdsourcing task in step S105 are as follows: The crowdsourcing task directly generated by the system according to the number of crowdsourcing people required for a single web page and a single item configured by the requester is called an ordinary crowdsourcing task. When the number of crowdsourcing people is an even number and among the final submitted numbers, the number of people who think it passes is equal to the number of people who think it fails, this task automatically generates a review crowdsourcing task, which is reviewed by the evaluators with higher review authority. The specific method for bypassing the problem that the pages to be tested and the result submission page are in different domains is as follows: By viewing the information of the evaluation task and submitting the results in the developer interface, and at the same time accessing the page to be tested in the main tab of the browser. The page accesses of the two are independent and there is no cross-domain problem, but they are in the same browser interface, realizing the access to two different domain pages in the same interface, which is convenient for the evaluators to carry out the crowdsourcing task.

6. The cross-domain collaborative website accessibility crowdsourcing evaluation method according to claim 1, wherein: The management method of the system server for distributing the evaluation tasks in step S106 is as follows: When the requester initiates an evaluation task, the requester specifies the number of crowdsourcing people for each crowdsourcing task. When the system server receives the application for a crowdsourcing task, it randomly selects a crowdsourcing task that has not reached the upper limit of the allocated number of crowdsourcing people and distributes it; at the same time, the allocated number of this crowdsourcing task is incremented by one. Within the timeout threshold of β seconds, this crowdsourcing task belongs to the evaluator of this distribution and cannot be distributed repeatedly. This evaluator can submit the results normally. If it exceeds the threshold, then this distribution times out and is invalidated, the allocated number of this crowdsourcing task is decremented by one, waiting for re-distribution. If this evaluator submits, it will be prompted that the task has timed out and needs to re-apply for a crowdsourcing task.

7. A cross-domain collaborative website accessibility crowdsourcing evaluation method according to claim 1, characterized in that: The interactive functions in the developer interface in step S107 include: switching the task to an ordinary task or a review task, the entrance to the personal center, the logout button, viewing the detection item information, submitting the evaluation results, abandoning this task, and reporting an incorrect web page. The crowdsourcing task information, its standard item descriptions, and evaluation methods shown in the developer interface include: the title of the web page to be tested, the title of the detection rule, the rule description, the detection method, and the passing conditions.

8. A cross-domain collaborative website accessibility crowdsourcing evaluation method according to claim 1, characterized in that: The evaluation results of the web page to be tested submitted by the disabled or able-bodied user evaluators in step S109 include: whether this web page to be tested passes the accessible items to be tested. If it passes, it is directly submitted. If it does not pass, the reason for non-passing is filled in and then submitted. After the disabled or able-bodied user evaluator in step S109 abandons a crowdsourcing task, the processing method of this task by the system server is: the number of people to whom this crowdsourcing task has been assigned is reduced by one, and it waits to be reassigned.

9. A cross-domain collaborative website accessibility crowdsourcing evaluation method according to claim 1, characterized in that: The conditions under which a disabled or able-bodied user can report an error web page in step S110 are: the web page is inconsistent with the web page to be tested, or the web page is a garbled web page, or the web page fails to open normally, and error codes 404 or 500 are prompted.

10. A cross - domain collaborative website accessibility crowdsourcing evaluation method according to claim 1, characterized in that: The report content generated by the server in step S111 and the passing criteria for items are: the report content includes the levels, contents, and passing situations of each standard item, the sample error web pages corresponding thereto, and the integrated errata list of all error web pages by sub-standard items; if among the evaluators, the number of people who think that this item on this web page passes is more than the number of people who think it does not pass, then this item passes on this web page, otherwise it does not pass; if the number of people is equal, the single review evaluation result shall prevail. The report forms that the evaluation requester can download in step S111 include: PDF format documents or Word format documents.

Images