September 15th Reading Notes

Data compression (Wikipedia)
- Process of encoding information using fewer bits than the original representation would use.
- Compression helps reduce the consumption of expensive resources, such as hard disk space or transmission bandwith.
- Lossless compression algorithms usually exploit statistical redundancy as a way to represent the sender's data more concisely without error.
- Close connection between machine learning and compression: a system that predicts the posterior probabilities of a sequence given its entire history can be used for optimal data compression, while an optimal compressor can be used for prediction.
- Data differencing: data compression is a special case of this. Consists of producing a difference given a source and a target.

Data compression basics
- Lossless compression algorithm, called "run-length encoding."
- RLE is very simple and fast to execute.
- A typical RLE image compressor will reduce the size of an uncompressed 24-bit photograph by less than 5%.
- LZ algorithms compress data by locating and eliminating redundancy. Performance is still highly influenced by the number of colors used.
- Entropy coding is a technique that assigns codes to blocks of data in a way that the length of the code is inversely proportional to the statistical probability of that symbol.

"Imaging Pittsburgh: Creating a shared gateway to digital image collections of the Pittsburgh region"
- Focus created a single Web gateway for the public to access thousands of visual images from photographic collections held by the Archives Service Center (Univ of Pitt), Carnegie Museum of Art, and the Historical Society of Western Pennsylvania.
- Increases public access to significant collections of historic material documenting the growth and development of Pittsburgh.
- Challenge working with the different organizations - different missions and institutional cultures.
- Metadata is "glue that holds these collections together" but creating of metadata has been a challenge.
- Difficult to develop consistent copyright and permission statements for the site and accompanying images.

"YouTube and libraries: It could be a beautiful relationship"
- Could upload instructional videos.
- Method of introducing resources that are available on campus.
- Teach students to use databases.