Click to edit Master text styles
Second level
Third level
Fourth level
Fifth level
Thank NLA;
Before we begin…
Software architect, Digital Archive.
Hired to develop preservation plans – developed a roadmap
Contributed to Global Digital Format Registry.
WAFER – risk analysis used here.
Worked as a consultant, large software firm.
almost 40 years
membership cooperative
world's largest consortium of libraries
computer library services, research organization
50K libraries in 84 countries
Maintains WorldCat; Dewey numbering system
… the result of 16 months of development, painstaking reviews and prototyping.
Domesday Book; 1976 Viking landings on Mars
1000 formats
Inevitable march to obsolescence
Transition – Why are we talking about migration, emulation, and all that? Why do we need to do anything upfront?
Is it possible that any digital format is capable to be used as a preservation format?
Tailor – measure twice, cut once
Measurement is not sufficient
Other systems – FCLA, UVC, Virtual Remote Control (the INFORM method can be used to document decisions).
Transition – Imagine trying to communicate before meter and yard were invented
… before meter and yard were invented. Still, ‘short’ vs. ‘long’.
| |
Set out to develop a method that can build a foundation for analysis and
reduces amount of guessing out of preservation decision making.
Be able to use this method to
compare formats,
understand what is gained and what is lost if migrating from one to another,
define preservation goals in terms of which significant properties can be preserved, and end up with is the ability to judge the most cost effective way to achieve known preservation goals.
Transition – INFORM measurement system measures risks to preservation
- recognizing and assess risks.
preservation planning == a response to known and unknown risks posed by the promise to provide access over time to digital objects.
Risk-based assessment methods - widely used in software project management, security threat analysis and business planning.
I used - my expertise in security threat analysis models,
a number of project management risk-based methods,
ideas developed by Cornell University’s Prism project,
various other papers
– identified about 70 risk factors.
DRM, encryption, lost commercial viability, unique support across industry
Transition – it is a combination of opinion polling (as in politics) and risk aversion (as in investing)
- apply measurement units to a measurement path
Specific Specs, dependencies, software & hardware requirements, orgs, etc.
Ask experts to estimate the risk exposure. For example:
Specification or source code can be independently inspected (open source & shared source models)
Complex, or poorly documented
Widely accepted and used
Unique in its class (competing or similar formats exist)
DRM, encryption, signatures, watermarks allowed, mandated, used
Supported by a very small organization or community, or by irreplaceable expertise
What experts can be included
Attempting to create objective group assessment from subjective individual opinions
Transition - results
As good as the experts utilized.
Method is scalable: use whatever “experts” you can find. Over time, hopefully more, better.
Do what you can with the information at hand.
Transition – method is descriptive, not prescriptive
Same as investing: broker measures and communicates risk, investor’s aversion determines best course of action
Transition – forecast immediate and short term needs
Reuse existing platforms and methods, immediate & future needs --> reduce costs
It can be used for looking at …
a)Content creation & publication – suggest best formats to use
b)Acquisition – restrict what is collected, know in advance how difficult to handle a certain collection may be
c)Preservation planning – migration strategies, short-term emulation solutions
Give an example:
- PDFs converted to TIFF images, related together by a METS XML doc
- Or, if a collection has 120,000 JPEG images, should they be converted to TIFF?
Future use – guidelines for expected usage of formats for collection owners
Compare TIFF to JPEG: pick TIFF if new objects, don’t convert existing JPEG objects. Format quality is important.
Functionality loss (cut & paste), full-text search.
Transition: how it works – next steps
Store results in a global registry (GDFR comes to mind); all contents shared; the more contributors, the better.