The HEA/JISC funded OER pilot project positively encourages the exploitation of technically sophisticated web-based services such as YouTube, Flickr etc. to present resources released by the project. The challenge is to then create a search facility that is able to pull together appropriate resources spread over a number of services on the web yet offer more information and presentational flexibility than the de-facto benchmark search tool, Google.
This paper presents an attempt to embrace the sophisticated interconnectivity available from these services and the human, technical and metadata considerations. The aim is to avoid creating another database that duplicates (possibly incorrectly) information held elsewhere eliminating the confusion that may cause and the maintenance load that checking and updating would cause. These issues will reduce the sustainiblity of such a project output.
Activity 2-unit 2-update 2024. English translation
Pull yourself together! Remote searching of multiple sources to best present OER materials
1. Pull yourself together! Remote searching of multiple sources to best present OER materials Lisa Rogers, l.j.rogers@hw.ac.uk Rob Pearce rob@engsc.ac.uk
2. What this presentation will cover Project overview / Search demo The premise: the conventional approach and why it’s creaky the distributed approach and why it’s creaky Technical development Demo of the distributed Search “Supersearch” Further work 1059 Pull yourself together! Remote searching of multiple sources to best present OER materials. Rob Pearce rob@engsc.ac.uk
3. Project overview OER – take teaching resources, clear legal ownership issues, give them away for others to use as they wish. A new take on an old idea OER project encouraged release OERs using innovative web services, e.g; YouTube. Also put all your stuff in the JorumOpen 1059 Pull yourself together! Remote searching of multiple sources to best present OER materials. Rob Pearce rob@engsc.ac.uk
4. …so what is a Supersearch? 1059 Pull yourself together! Remote searching of multiple sources to best present OER materials. Rob Pearce rob@engsc.ac.uk
5. Project overview The OER pilot project for Engineering disseminates its resources, where possible, through YouTube, Flickr, SlideShare, Vimeo, Scribd and others. Instead of building (yet another!) local database of these resources to create a cross search service, I decided to investigate using the “APIs” from each service, as well as third party tools such as Yahoo! Pipes and Google custom Search. This presentation summarises the progress so far. 1059 Pull yourself together! Remote searching of multiple sources to best present OER materials. Rob Pearce rob@engsc.ac.uk
6. …why this approach? Plan “A” – a conventional databases: Lo-Risk Pros: normal approach, well understood, easy to control, Cons – data repeated in many different places time consuming data always inconsistent or out of date 1059 Pull yourself together! Remote searching of multiple sources to best present OER materials. Rob Pearce rob@engsc.ac.uk
7. …why this approach? Perfection: one source of data, so easy to control / better version control some guarantee of a consistent service no duplication quick to update, lo-propagation delay data always consistent 1059 Pull yourself together! Remote searching of multiple sources to best present OER materials. Rob Pearce rob@engsc.ac.uk
8. …why this approach? Plan “B” – compromise, the Supersearch Pros: reduces repetitionof info to a minimum whilst still meeting projectrequirements saves potential users from having tovisit multiple sites can provide better answers one-stop-shop for resourcesand project dissemination pulls together multiple media thatshould be linked together provides partner specific materials’ views Cons: depends on consistentservice levels from free services time consuming “APIs” not consistent 1059 Pull yourself together! Remote searching of multiple sources to best present OER materials. Rob Pearce rob@engsc.ac.uk
9. The technical approach 1059 Pull yourself together! Remote searching of multiple sources to best present OER materials. Rob Pearce rob@engsc.ac.uk
10. 1st test: Google Custom Search Google provides a custom search engine facility. Results can be limited to a list of predetermined websites or can be filtered by keywords. This search engine limits the results to the following domains.Flickr http://www.flickr.com, Scribd http://www.scribd.com, SlideShare http://www.slideshare.net, YouTube http://www.youtube.com as a demonstration the project produced: http://www.google.co.uk/cse/home?cx=007182910873444472376:bdz1enadj3a search limited with keyword “engscoer” which is the identifying tag for the project. irregular results, for example, results may appear in the refined results though not be present in the unrefined search results. This does however seem to be gradually improving over time. easily extensible; more sites could be included without too much difficulty. using labels, could refine a search to photo sharing sites such as Flickr and Photo Bucket. Cons: Options for embedding limited to JavaScript embed code. Results a little inconsistent, lack of fine control over look. 1059 Pull yourself together! Remote searching of multiple sources to best present OER materials. Rob Pearce rob@engsc.ac.uk
11. 2nd test: Yahoo! Pipes Yahoo pipes is described as “a powerful composition tool to aggregate, manipulate and mash up content from around the web”. The demonstration pipe is available here http://pipes.yahoo.com/pipes/pipe.info?_id=3046acdeb4f3af164c7abc1ed83a388a for Scribd and SlideShare, the pipe pulls in RSS feeds from these sites of resources tagged engscoer, then filters these results based on a keywords in the title and/or description. for YouTube an RSS feed URL is constructed based upon the project tag and keywords entered in the query. This seems to be more effective than filtering an existing RSS feed. for Flickr, the Yahoo! Pipes Flickr module was used, allowing users to perform a query to find a particular number of images, matching their search criteria as well as the project tag. This again seems to be an effective mechanism. the output from Yahoo Pipes can be delivered as an RSS feed, JSON or PHP. JSON and PHP allow much more flexibility in which metadata elements from the results are displayed, though this requires more advanced programming skills than the Google CSE embed code. example of yahoo pipe output embedded within a webpage http://www.icbl.hw.ac.uk/engscoer/yahoopipe.php Cons: Can be complicated, slow, depends on a free “none core” Yahoo service, RSS feeds may only contain the latest 20 items therefore older resources would be overlooked. Embedding options limited, results a little inconsistent, lack of fine control over look. 1059 Pull yourself together! Remote searching of multiple sources to best present OER materials. Rob Pearce rob@engsc.ac.uk
12. 3rd test: Search “APIs” APIs provided by the five main file sharing platforms. each API requires the search query to be formed in a unique way, differ in the way the queries are performed and may not return all the metadata fields that are required. a certain amount of effort is required to configure each API. Luckily, as these services are popular, there are community developers who have created PHP clients for some of the APIs. this became the approach of choice. tabs are in fact separate web pages, when a new tab is clicked upon, it performs the search for that service, rather than in he background. 1059 Pull yourself together! Remote searching of multiple sources to best present OER materials. Rob Pearce rob@engsc.ac.uk