Recent NSR Modifications
Based on NSR user feedback the following improvements have been made:
NSR Modifications - May 2007
During 2006-2007 a new version of the NSR database and web application have been developed. This page outlines the major changes.
The database shema has been modified somewhat and most of the supporting Java code has been rewritten. The most important aspect of the code modification is the creation of an entirely new "selector generation" code. This is the program that reads the keyword abstract for each entry and generates a "selector" field, which is then used for indexing the references. As a result, selector fields for all NSR entries have been re-generated for this release.
Authors now indexed by last name and two initials, rather than last name and one initial. Now searches on "A.B.Smith" and "A.R.Smith" will result in different reference lists. This required re-generating the author dictionary, to account for "aliases" correctly. In the process, introduced aliases for collaboration names (i.e. searching on "Fermilab E581" and "FNAL E581" will retrieve the same references). It is likely that some author aliases have been missed - this is fairly easy to correct if brought to the attention of the NSR manager.
Parent and Daugter nuclides:
At the request of the NSDD, daughter nuclides are now calculated and indexed, where possible, as selectors in references with RADIOACTIVITY keywords. Daughter and parent nuclides now have different identifiers in selectors. Previously, there was T for target and N for nuclide. Now there is T and N, plus P for parent and G for daughter (D was already in use in selectors for "deduced"). This makes it possible to search specifically on daughter and parent nuclides.
Multiple selectors per keyword phrase:
Unlike the previous version of nsrprep, a given word or phrase can now generate multiple selectors. For instance, the word "octupole" will generate the indexed selector "OCTUPOLE" and the more general "MULTIPOLE". (In the past, both "octupole" and "dipole" generated just "QUADRUPOLE". Not sure why it was done this way). Similarly, all cluster decays generate "CLUSTER-DECAY" selector in addition to selectors for specific decays (i.e. "12C-DECAY").
Subject generation from reaction text:
The selectors "FISSION" and "POLARIZATION" can now be generated by appropriate reaction strings. Previously, they required specific words elsewhere in the keywords (i.e. "analyzing power" for POLARIZATION).
Minor topics not indexed:
"Minor topics" have been deprecated, and are not indexed. These appeared after the major topic, and before the bulk of the keywords. Examples are "Mesic-atom", "ICPND", etc. The "mesic-atom" selector will now be generated based on specific phrases found in the keyword abstract, such as "pionic atom", etc.
"Mass" selector now generated:
The word "mass" will generate the selector "MASS" if the major topic is "ATOMIC MASSES". In the past, the "MASS" selector existed in principle but was never generated.
A few selectors were added, including "HYPERFINE" and "HALO".
The major topic "ISOTOPE SHIFTS" has been dropped.
Changes to web application are fairly minor at this point. Some code was modified to work with new schema - i.e. two author initials, searches on parent or daughter nuclides.
General changes to "Indexed Search":
Drop-down lists on "indexed search" page include indents to help clarify the relationships among various options. (i.e., to show that "measured", "deduced", etc are subsets of "subject").
Changes to nuclide search:
Searches on "nuclide" will list references that have that nuclide in the selectors as "N" (nuclide), "P" (parent), or "G" (daughter). Searches on target, parent, or daughter will only pick up those types in selector list.
Changes to author search:
Author specification can now include zero, one, or two initials along with last name. If only one initial is entered, it will return all references that have an author with that first initial and last name. In other words, searching on "A.Smith" will pick up references for A.Smith, A.B.Smith, A.G.Smith, etc.
One completely new feature, experimental at this point, is called "list analyze". To some extent, it is an attempt to do something similar to what "INSPEC" and some publisher sites can do with reference lists. It is accessible via the "combine/view" page, after one or more queries have been run. It lists information on the following for a given list of references by clicking the "analyze" button: nuclides studied, publication years, authors, journals. Because this option is somewhat memory-intensive, it is only offered on lists that contain between 1 and 1000 references.