Screen index file

Chemistry Retrieval Subsystem. The chemistry retrieval subsystem file design and organization is being described in detail for publication elsewhere [ ]. Briefly the system consists of two numeric index cross reference files, a screen index file, and a master structure file. It contains about 270,000 unique structures and occupies 8 disk packs of 37 million characters each. 

The screen index file is the first file accessed for any structure or sula-structure search. It contains all the information necessary to determine structure matches. When the structure matches have been located, additional information for each structure, such as the structure picture, may be retrieved via the accession number. Because its orgcuiization is index-sequential, the screen index file may iye accessed either sequentially, or selectively by use of the its Indexes. 

The partitioning factor is conceptually similar to Hbde s bucket index [4]. It is a 12 bit code derived from the screen through a series of AND and OR operations in such a way that the inclusion properties of the screen are preserved. Its function in the screen index file is to allot records to one of 4096 partitions with a theoretically uniform distributation. 

The usefulness to the WRAIR Staff of searching for substructures on-line has yet to be determined. Since the chemistry retrieval system uses the screen and connection teible generated by the query to search to screen index file for either a whole or a sub-structure no more information (disk-drives) is needed on-line for sub-structure searching than for identity matches, so one is no more costly than the other, in terms of disk resources. But the elapsed time needed for such sub-structure searches varies greatly and may be large. The time needed may be kept small if the query is specific and many screen bits are set. 

Up to six data files can be plotted on a single axis or one file on each of three axes. To plot the user simply selects the file or files of interest from the index screen and clicks on the plot option. If more than one type of curve is available, the user is provided with a selector box to assign curves to axes. Figure 3 shows a sample of the plot window. 

One problem with documents kept in any system is that other people given the document will need the same application to view the document. Sending a file in Microsoft Word format requires that the recipient also own a copy of Microsoft Word. Without it, the document cannot be viewed. This problem has been solved by a new standard known as Acrobat from Adobe systems. By using Acrobat, anyone wishing to publish a document uses the Acrobat software to create a version readable by the Acrobat Reader. The reader software is freely distributed, so the recipient of a document need not buy a copy of the original software. Besides being able to create documents viewable by anyone, this new standard also allows the document to be indexed. Any part of the document can also be cross-referenced to other sections. These features are especially useful in documentation, particularly in NDA submissions. Volumes of paper documents can now be stored on a single disk and instantly cross-referenced. A reviewer could have an entire NDA available on the computer screen. 

File Segments are a type of screen used to limit the answer set retrieved to a particular class of compounds. Using them, it will be possible to select only records for documents in certain Central Patents Index (CPI) sections, or to limit the search to coordination complexes or other types of structures. These segments will be searchable in the commercial release of the system, and Boolean logic capabilities will be subsequently available. 

The CSD is released, twice each year, as an indexed sequential binary file that contains additional data items derived from the basic input. Principally, these are bit-encoded screens that are required to enhance search speeds within the CSD software. Most screens are derived from the 2D connectivity... 

Screen capture of radiograph by Lou Cornicelli, Minnesota Department of Naturai Resources big game program coordinator, showing the reiative iack of fragmentation of a shotgun siug, even when it hits bone, (http //files.dnr.state.mn.us/fish wildlife/lead/index.htm)... 

Then, because the files are indexed by an abbreviation of the screen, i.e., the partitioning factor, they need not be read completely, thus eliminating most of the input/output time necessary to read the entire file. (Of course additional time will also be saved because such a situation will limit the number of iterative matches needed as well.) The worst situation is a query so general that the entire screen file must be read, and most of it must be searched iteratively. The time for usch a case depends largely on the size of the file and the competition for CPU time from other multi-programmed jobs. The estimated worst-case search would require about 35 minutes elapsed time in an otherwise empty machine, for a file of 250,000 structures. 

---