Brewster Kahle’s (Internet Archive) ISWC talk on worldwide distributed knowledge

Universal access to all knowledge can be one of our greatest achievements.

The keynote speech at ISWC 2007 was given this morning by Brewster Kahle, co-founder of the Internet Archive and also of Alexa Internet. Brewster’s talk discussed the challenges in putting various types of media online, from books to video:

  • He started to talk about digitising books (1 book = 1 MB; the Library of Congress = 26 million books = 26 TB; with images, somewhat larger). At present, it costs about $30 to scan a book in the US. For 10 cents a page, books or microfilm can now be scanned at various centres around the States and put online. 250,000 books have been scanned in so far and are held in eight online collections. He also talked about making books available to people through the OPLC project. Still, most people like having printed books, so book mobiles for print-on-demand books are now coming. A book mobile charges just $1 to print and bind a short book.
  • Next up was audio, and Brewster discussed issues related to putting recorded sound works online. At best, there are two to three million discs that have been commercially distributed. The biggest issue with this is in relation to rights. Rock ‘n’ roll concerts are the most popular category of the Internet Archive audio files (with 40,000 concerts so far); for “unlimited storage, unlimited bandwidth, forever, for free”, the Internet Archive offers bands their hosting service if they waive any issues with rights. There are various cultural materials that do not work well in terms of record sales, but there are many people who are very interested in having these published online. Audio costs about $10 per disk (per hour) to digitise. The Internet Archive has 100,000 items in 100 collections.
  • Moving images or video was next. Most people think of Hollywood films in relation to video, but at most there are 150,000 to 200,000 video items that are designed for movie theatres, and half of these are Indian! Many are locked up in copyright, and are problematic. The Internet Archive has 1,000 of these (out of copyright or otherwise permitted). There are other types of materials that people want to see: thousands of archival films, advertisements, training films and government films, being downloaded in the millions. Brewster also put out a call to academics at the conference to put their lectures online in bulk at the Internet Archive. It costs $15 per video hour for digitisation services. Brewster estimates that there are 400 channels of “original” television channels (ignoring duplicate rebroadcasts). If you record a television channel for one year, it requires 10 TB, with a cost of $20,000 for that year. The Television Archive people at the Internet Archive have been recording 20 channels from around the world since 2000 (it’s currently about 1 PB in size) – that’s 1 million hours of TV – but not much has been made available just yet (apart from video from the week of 9/11). The Internet Archive currently has 55,000 videos in 100 collections,
  • Software was next. For example, a good archival source is old software that can be reused / replayed via virtual machines or emulators. Brewster came out against the Digital Millennium Copyright Act, which is “horrible for libraries” and for the publishing industry.
  • The Internet Archive is best known for archiving web pages. It started in 1996, by taking a snapshot of every accessible page on a website. It is now about 2 PB in size, with over 100 billion pages. Most people use this service to find their old materials again, since most people “don’t keep their own materials very well”. (Incidentally, Yahoo! came to the Internet Archive to get a 10-year-old version of their own homepage.)

Brewster then talked about preservation issues, i.e., how to keep the materials available. He referenced the famous library at Alexandria, Egypt which unfortunately is best known for burning. Libraries also tend to be burned by governments due to changes in policies and interests, so the computer world solution to this is backups. The Internet Archive in San Francisco has four employees and 1 PB of storage (including the power bill, bandwidth and people costs, their total costs are about $3,000,000 per year; 6 GB bandwidth is used per second; their storage hardware costs $700,000 for 1 PB). They have a backup of their book and web materials in Alexandria, and also store audio material at the European Archive in Amsterdam. Also, their Open Content Alliance initiative allows various people and organisations to come together to create joint collections for all to use.

Access was the next topic of his presentation. Search is making in-roads in terms of time-based search. One can see how words and their usage change over time (e.g., “marine life”). Semantic Web applications for access can help people to deal with the onslaught of information. There is a huge need to take large related subsets of the Internet Archive collections and to help them make sense for people. Great work has been done recently on wikis and search, but there is a need to “add something more to the mix” to bring structure to this project. To do this, Brewster reckons we need the ease of access and authoring from the wiki world, but also ways to incorporate the structure that we all know is in there, so that it can be flexible enough for people to add structure one item at a time or to have computers help with this task.

20071113b.jpg In the recent initiative ““, the idea is to build one webpage for every book ever published (not just ones still for sale) to include content, metadata, reviews, etc. The relevant concepts in this project include: creating Semantic Web concepts for authors, works and entities; having wiki-editable data and templates; using a tuple-based database with history; making it all open source (both the data and the code, in Python). has 10 million book records, with 250k in full text.

I really enjoyed this talk, and having been a fan of the Wayback Machine for many years, I think there could be an interesting link to the SIOC Project if we think in terms of archiving people’s conversations from the Web, mailing lists and discussion groups for reuse by us and the generations to come.


Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s