Who is SMILing on the Web?

Dick Bulterman Oratrix Development Valeriusplein, 30 in Amsterdam, The Netherlands E-mail:[email protected] INTRODUCTION The amount of streaming media available on the web is orders of magnitude less than the amount of text and the resources for locating relevant information contained in streaming media [4], such as search engines, web directories and portals are not as well developed as those for text. The indexing challenges offered by the unstructured, spatio-temporal nature of media together with the lack of universal media standards largely accounts for this. The W3 Synchronized Multimedia Integration Language (SMIL) [6] standard partially addresses these issues with the support for creation and presentation of complex streaming media across multiple formats. Our focus here is the characterization of web pages that are SMILing or streaming complex media presentations. Since exhaustive enumeration of web pages is impractical, we take a sampling approach. Experimental validation shows that less than 1% of web pages contain links to streaming media, of which 65% is RealMedia, 11% is Windows Media, 2% is SMIL and 22% is other formats. We also investigate the distribution of links from web pages to streaming media and find it skewed to small numbers of both in (to media files) and out (from web pages) links.