|
 After two years of work, an innovative project using Web-based
technologies to speed researcher access to a large body of new
scientific data has demonstrated that not only access to but also the
quality of the data has improved markedly. A new paper* on the
Web-enabled ThermoML thermodynamics global data exchange standard notes
that the data-entry process catches and corrects data errors in roughly
10 percent of journal articles entered in the system.
A
landmark partnership between the National Institute of Standards and
Technology (NIST), several major scientific journals and the
International Union of Pure and Applied Chemistry (IUPAC), ThermoML was
developed to deal with the explosive growth in published data on
thermodynamics. Thermodynamics is essential to understanding and
designing chemical reactions in everything from huge industrial
chemical plants to the biochemistry of individual cells in the body.
With improvements in measurement technology, the quantity of published
thermophysical and thermochemical data has been almost doubling every
10 years.
This vast flood of information not only presents a
basic problem for researchers and engineers--how to find the data they
need when they need it.-- but also has strained the traditional
scientific peer-review and validation process. "Despite the peer-review
process, problems in data validation have led, in many instances, to
publication of data that are grossly erroneous and, at times,
inconsistent with the fundamental laws of nature," the authors note.
The
ThermoML project began as an attempt to simplify and speed the delivery
of new thermodynamic data from producers to users. The system has three
major components -- ThermoML itself, an IUPAC data format standard
based on XML (a generic data formatting standard) customized for
storing thermodynamic data; Software tools developed at the NIST
Thermodynamic Research Center (TRC) to simplify entering data into the
system in formats close to those used by the original journal
documents, displaying it in various formats and performing basic data
integrity checks; and The ThermoData Engine, a sophisticated expert
system developed at NIST, that can generate on demand recommended,
evaluated data based on the existing experimental and predicted data
and their uncertainties.
Authors
writing for five major journals that are partners in the program, the
Journal of Chemical and Engineering Data, the Journal of Chemical
Thermodynamics, Fluid Phase Equilibria, Thermochimica Acta, and the
International Journal of Thermophysics, participate in the process by
submitting the data for their articles using GDC software (available
from NIST). The data are evaluated, and any potential inconsistencies
reported back to the authors for verification. Based on two years of
experience and some 1,000 articles, the authors write, an estimated 10
percent of articles reporting experimental thermodynamic data for
organic compounds contain some erroneous information that would be
"extremely difficult" to detect through the normal peer-review process.
Related Links:
ThermoML :: An XML-Based IUPAC Standard for Storage and Exchange of Experimental Thermophysical and Thermochemical Property Data.
|
