《The Management of Crowdsourcing in Business Processes》由会员分享，可在线阅读，更多相关《The Management of Crowdsourcing in Business Processes（2页珍藏版）》请在金锄头文库上搜索。
1、The Management of Crowdsourcing in Business ProcessesStephen Curran1, Kevin Feeney1, Reinhard Schaler2, David Lewis1Centre for Next Generation Localization:Trinity College Dublin1 and Localization Research Centre, University of Limerick2Stephen.Currancs.tcd.ie, Kevin.Feeneycs.tcd.ie, Reinhard.Schale
2、rul.ie, Dave.Lewiscs.tcd.ie 1. IntroductionOrganizations increasingly place value on the devolution of decision-making authority both within and beyond traditional organizational hierarchies and boundaries. The sacrifice of centralized management control is exchanged for increased engagement from st
3、aff, customer and partners and the more agile decision making and improved commitment to implementing decisions this implies. However, most existing IT service management and workflow management systems assume a centralised management model and offer very limited management support for devolved deci
4、sion making. In this paper, we examine an approach to devolved decision-making in industrial workflow through the trend to crowdsource elements of the localization workflow for digital content, by leveraging the linguistic talents of online user communities which use the content or associated produc
5、ts. .In general, both IT service management and workflow management solutions target the enterprise market and are therefore marketed toward individual companies. Management support for inter-organization workflows adopt a value-chain world view that aims to optimize the position of the organization
6、 operating the system and treats people and companies outside of the organization as either downstream suppliers or upstream consumers. Underlying the use of crowdsourcing in workflows is a Value Network world view allee, where the value to the end user is prioritized by all participants in the netw
7、ork, and benefits are sought beyond direct financial exchange such that customer good-will and loyalty, community cohesion and free knowledge exchange are explicitly valued. However, the management of crowdsourcing in an enterprises workflow poses challenges for the centralizing assumption embedded
8、in workflow standards such as BPEL bpel and IT management procedures. 2. Challenges in Managing Crowdsourcing in the Localization ProcessLocalization workflows address the translation of digital content, such as software UI text or product manual, into different target market languages. Usually para
9、llel translation is required to multiple languages simultaneously, within a fixed deadline. The most common model is for much of this task to be mediated by language service providers, who manage a cohort of translators who often work remotely, e.g. from home. Increasingly previous translation of si
10、milar material is retained and reused in the form of a Translation Memory (TM) available to all workers, so that no phrase need be translated more than once. In addition, techniques such as statistical machine translation are starting to be used to assist the translator. Managing the quality of tran
11、slation against the cost and time (measured per word) across this value-chain remain a major challenge for industry. It relies on human quality control by specialist post-editors, and is sensitive to factors such as the source and destination languages, the content domain, the repetitiveness of the
12、content, complexity of sentences, availability of TM, and is therefore regarded as relatively subjective.One potential advantage of crowdsourcing is to address the shortcoming of subjective quality assessment by supporting “wisdom of the crowd” style rating of translation. This may be especially adv
13、antageous if those involved in rating are representative consumers of the translations, since their satisfaction is the ultimate gauge of quality. Successful examples of crowdsourcing have required significant investment of time and energy in platforms that enable the organization to effectively lev
14、erage the input of the crowd. In order for users to take the trouble to contribute, they need to be provided with some motivation to do so. A user must feel a sense of community and that their contributions are valuable. A system that aims to build a community must support communication between the
15、participants. The platform should provide tools such as forums, blogs, wikis etc. to support community building and enable collaboration. These tools should be managed to ensure that they are used for the intended purposes.The crowd normally consists of many overlapping groups with various different
16、 shared attributes each with a different relationship to the organization. Sometimes their contributions should be treated differently. For example, some translatable content may contain confidential information and should have restricted access while some translatable content may involve the translator to have highly specialized knowledge. In order to allow organizations to leverage user-generated content in such a way that the management cost of the system does not grow more
