From excellent (txt) README (!) :
1 Rights and Conditions of Use
The Archimedes Palimpsest data is released with license for use under
Creative Commons Attribution 3.0 Unported Access Rights. It is
requested that copies of any published articles based on the
information in this data set be sent to The Curator of Manuscripts,
The Walters Art Museum, 600 North Charles Street, Baltimore MD 21201.
2 Intended Audience and Consumers
The Archimedes Palimpsest Digital Product is intended to serve any
interested user or party.  However, its content is focused on serving
the following groups.
- Scholars of Greek and mathematics
 
- Application providers
 
- Libraries and archives
 
- Image scientists, and scientists in other disciplines interested
    in the production of the images
 
3 Digital Project Data Set Purpose
The Archimedes Palimpsest Digital Product provides all the digital
information available on the Archimedes Palimpsest in a single digital
data set, with a standard structure.  Its purposes are threefold:
- 
Serve as the authoritative digital data set of images in a
    standardized format that meets the needs of users, information
    providers, archives and libraries.
 
- 
Provide derived information (i.e. transcriptions, processing
    information) in the context of digital images of the original
    manuscript in a single integrated package.
 
- 
Offer a standard product sustainable by users to which current or
    future contributors can add additional standardized information
    (e.g. alternate texts, image analyses or conservation
    information).
 
4 Data Set Contents
This data set consists of:
- 
a core content set digital images and transcriptions of the
     Archimedes Palimpsest, each with accompanying metadata and
     checksums
 
- 
project-generated and third-party documentation of all included
     components
 
- 
supporting functional files, including XML schemas, and cascading
     style sheet files
 
- 
supplemental versions of the transcriptions by treatise and work
 
- 
a directory for researcher contributed content files, not a part
     of the core data set