Graphics and Archaeology: Interpreting the Past

Snowbird, Utah
20-23 May 2000

 

Panel:

Digital Archiving of 3D Artifacts

Presenters:

  • Marc Levoy
  • Lon Addison
  • Chuck Hixon
  • Kevin Cain

Summary:

  • Scribe's note - this session record is an approximation of the lively discussion.. Elements below are in chronological order and position presentations are interleaved without headers.

Lon started off the conversation by demonstrating his copy of the Andrea Palladio Le ville CD-ROM. It is a World Heritage site, yet only 50 copies of the CD-ROM were ever made. It is a multi-media kiosk presentation made several years ago. All of the villas are presented by name and image. It is an extremely comprehensive archive of photos, text, 3D rendered models, QuickTimeVRs, full bibliographies, and temporal navigation. It is a fantastic database, but many questions arise:

  • not many people have access to this
  • also, people cannot access this data format well today
  • websites - come and go - links die for many types of reasons
  • why can't we keep our digital archives alive?

Marc next got up and offered a list of issues for consideration and discussion:

  • Archiving
    • expense of storing 3D data [250GB - who can store this practically?]
    • central versus distributed archiving
    • authenticity, versioning, variants
    • how long will the archive last?
      • how do we cycle the data forward in time and space?
  • Distributing
    • permissions, distribution, payments
    • robust 3D digital watermarking
      • mapmakers ploy - change something small in image to track variants
    • detecting violations, enforcement
    • choice of media, bandwidth, compression
  • Viewing
    • real-time viewing on low-cost PCs
    • cataloging, searching
    • making measures from the data
    • extracting data from the archive
  • Protecting
    • who owns the data?
    • images versus 3D models
    • raw data versus elaborated models
    • scholarly versus commercial use
    • virtual versus physical copies

Someone raised the question - copyright law states there is some [75 years] limitation of ownership. Museums in general claim control of the artifact to control use of the image. A counter point was offered - do we need to take a stance as digital archivers - it's old, it's free? There is ongoing debate (Getty website, NEA) on digital preservation. All the current attention is being done for the 3D world - the 3D perspective has not really been considered consistently. Governments need to be lead through this - digital copyright extends in many directions.

It was noted that a recent panel tried to address issues of digital rights. There was a very hotly contested final report. The National Academy of Sciences panel suggested that it was not time to make laws right now - the market should sort itself out first, especially in the wake of recent music ownership controversies.

It was also noted that there is a need to tread very carefully because some elements are strong centers of national pride. In Greece, for example, if you are found with a shard of pottery, you can be jailed for 15 years.

Dave offered if everything is desired to be open, rushing in may cause others to close down / isolate / deny access to material and information. Someone offered that appropriate use of data and control of data was important. There was a question, however - when something permeates society to a great degree - e.g., Xerox, Kleenex - what is the boundary of public interest/domain? What if I took every possible photograph (50+ years old or at least those with expired copyrights) and created a CG model for the public domain - is this not legal?

Marc noted that the state of the art will always be painful. Technology will go forward, but the desires will always exceed that edge.

Ruth offered that in Turkey, they believe the control is around the artifact. Our data is one step removed - at what point does data become interpretation and should be unchanged? There has to be some way of preserving original sets of data.

Chuck noted that we archive our CAD files for projects. Our policy is that if you modify our CAD files (after you have received them as the client), you are liable for any changes that you make. Contractually, anything we create, we own. If ownership is important, that must be established up front.

Duncan noted that two types of data were being talked about:

  • original source - Michelangelo remains
  • excavation - destroys while discovering - data becomes the only persistent record - it has to be as complete as possible

David offered an example of Packard at University of Chicago / Getty Trust. There was a multi-year survey and it was conducted traditionally. There was no problem to release data to the public. There is more concern over 3D data - what can be reproduced and used from this data?

Duncan noted that physical artifacts have gatekeepers called curators. Items in a store room are accessed so infrequently, it is easy to manage. Internet access, however, could possibly mean over 300 accesses a day!

Someone commented that we need to consider the purpose of the data; some types of data needs to be distributed. A question came up - is the issue about not getting royalties for use? The answer was "yes, but also there is a possibility of tasteless applications."

Impermanence of formats and storage came up as a barrier to archiving, especially for Anne who's slide collection continues to age but might not make it forward in digital formats, too. Lon offered that at Berkeley, we've had to restart our archiving process - but we've learned valuable things each time. Anne should know about the issues of format and lossy-or-not.

Chuck added that we need to keep up with technology - his costs also account for keeping archives alive. Duncan noted in archaeology, our biggest cost is space. Digital archives have the advantage of being smaller. Stuart added that in Scotland, the Archaeological Data Service works to fund projects to address this issue.

Lon pointed out that this is a social issue. We need to be responsible in anticipating this problem. Marc added that we almost lost the 1960 census because we couldn't read the format. Paul then added the concept of physical versus virtual storage. Ubiquity will help preserve our chances of being connected. Distribution can help preserve it. We need well-defined image formats (at least for 2D). Even digital-to-digital copies. There's generational loss - roundoff error (8bit -> 12bit). There needs to be known conversion consistency. Checksums can help guarantee integrity.

Lon ended with a comment "yes, look at the Library of Alexandria - localized information lost in a disaster versus a distributed model". Perhaps the risk is self-evident. Anne, jump into your digital conversion with eyes wide open (but jump all the same!)

Conclusions / Issues:

  • The needs and issues of this topic are greater than SIGGRAPH and Eurographics
  • Archiving technology - there is no standard and it is always changing
    • there was great concern that when we present it, the "impermanence" issues might scare people away from even trying it
    • Anne is concerned about not trying at all - will there be a discontinuity, however, that thwarts the investment?

 

[ Previous ] [ Index ] [ Next ]