OERInterop-Search

<- Back to OER Interoperability Principles 生活随笔 会计服务 货运公司 功率放大器 笔记本维修 软件著作权申请 新胜达 断桥铝合金 花卉租摆 温度控制器 阳光房 素食 北京保洁公司 压路机 北京物流 工商注册 电脑回收 温控器 航空货运 代理记帐 电脑租赁 展览展示 吹膜机 鱼鳞病 升降平台 kvm 仓储货架 汽车租赁 铜门 铜窗 铜装饰 铜扶手

Enable search and discovery of the OER of interest.
 * Provide as much authoritative metadata as you can
 * Dublin Core, educational standards relevant to your target audience,
 * Facilitate the provision of as much non-authoritative metadata as you can
 * People generated: Ratings, annotations, bundle neighbors
 * Machine generated: Number of downloads, most recently accessed
 * Encourage and enable formal vetting (lensing) of content
 * link your tools into normal workflow of author, teacher and learner
 * provide "different" kind of metadata
 * attention metadata, location metadata, etc.
 * not search, discoverability and findability

Group ideas:
 * Are we talking about publishers?
 * What metadata are needed?
 * Are publishers just one role in a larger community responsible for providing OER content?
 * What are the goals for a user? Engagement with the whole OER community? What is the environment that would need to exist to make this possible?
 * Even if there are different strategies for finding and engaging with materials, there are probably common strategies that underlie the facilitation of those processes. The scale at which we want users to be able to engage will affect the degree to which specific, standardized formats are necessary.
 * People will want what they want when they need it. How does this affect our recommendations?
 * We need to unpack the word "quality". What branding mechanisms are available? What role for trust?
 * Is trust really that big a deal? Can't people learn how to do the evaluation themselves?
 * Is this trust issue relevant to search and discovery?


 * OK, the way to indicate trust and sponsorship is to ensure that the metadata contain attribution and or marks of origin.
 * What about state standards, etc.
 * As there are collections of services evolving, how do we reclaim community-generated metadata?
 * What types of metadata are not being widely applied but should be?
 * User feedback. Ratings. # of downloads. Etc.
 * Emerging standards, Dublin Core, etc.


 * If a user has never heard of OER, how will they discover it?
 * Through normal search engines.
 * For learners, they probably engage with "social graph" (e.g., Facebook, etc) more than they query Google.
 * If they stumble on OER, how will they become aware of the breadth of opportunities, rather than just the site on which they landed?
 * Issue is to get the content into the learners' environments, not to get the learners into the content sites.

Barriers to the dream:
 * Metadata must be present.
 * If present, make it clear how to access/contribute to the metadata.
 * If rich media, transripts must be time-coded.
 * If no metadata present:
 * What can be generated automatically?
 * If fields are missing, how to fill in?
 * Autogenerated metedata are at least as good as person-generated metadata.
 * Is there a mechanism for getting metadata from users back into the system so that others can benefit? FreeBase?
 * Perhaps it would help to distinguish among providing metadata versus maintaining metadata?
 * Do we need more educational materials to explain the logic of metadata and/or licensing?

What about search optimization?
 * What conditions are needed?
 * How to enable search and discovery within the OER community regardless of the entry point?
 * OER recommender does this, but we need to get it out of javascript. Talk to COSL folks because they can give the XL that will enable deep linking to related materials.
 * Can use a registry, with added info, especially related to the technical formats, etc.

Some basic principles - Hal:
 * Accessibility of the URLs. Not all URLs are accessible (i.e., findable automatically).
 * Identifying things below the level of the page.
 * For any site, you should be able to export the metadata in an extensible format. Is it important that the tags have unambiguous meanings? Delicious example: no formal meaning or structure, but works pretty well.
 * Is there any advantage to namespaces versus just using existing RSS standards? If you use RSS, what should be fed?

Specific recommendations:
 * URLs must be accessible.
 * Not behind a password.
 * Are there ways to divide (or mirror) sites so that they are a mixed open/closed system?
 * We must have access to the metadata and knowledge of the services that are built around them.
 * Ideally, it should be made available in XML.
 * There may be specific formats for the XML that are better than others?
 * Is there a contrast? There seems to be a debate in the tech community about what is actually the best practice.
 * Perhaps there is a need to distinguish among the infrastructure that will enable standards to evolve, versus the actual social behavior that interfaces with that architecture (either successfully or not)?
 * If registries are being created, the registries themselves should be interoperable.

A possible tool of interest:
 * Should we build a visible brand - OER as a mark. The mark would provide legal and technical info?
 * Define OER.
 * Have a checklist or some such thing to help people identify whether they are actually offering OER.
 * Serves as an educational tool in its own right. Can also track derivatives and the growth of the space.
 * ccLearn has already considered such a tool, but as David Wiley says, moving on this would require some consensus on defining OER.
 * Is there room for more then one definition?