The inner workings of our own Galaxy are more mysterious than those of galaxies located millions of parsecs away, mainly because of our unfavorable location in the mid-plane outskirts of a dusty disk. The Milky Way is a typical luminous spiral of the types seen to large distances. Understanding its structure and the fundamental processes and interactions between gas and stars within it would benefit all galactic and extragalactic research programs. A Space Infrared Telescope Facility (SIRTF) Galactic Plane survey will provide a comprehensive view of the entire stellar and dust content inner Galaxy. The broad scope of these data will provide a global understanding that narrowly confined selected regions cannot, and they will provide vital information regarding two basic scientific questions:

• What is the structure of the inner Galaxy, including the disk, molecular ring, and bar, as traced by the spatial distribution of stars and IR-bright star forming regions (SFRs)?
• What are the statistics and physics of star formation? What are the properties of various stages of star formation, and how do they vary with location?

The first question involves the appearance of the inner Galaxy to an outside observer, vital for accurately comparing it to other galaxies. The second question involves the details of the star formation process by considering young, but heavily obscured, clusters which can be resolved spatially for the first time with SIRTF. Only the broad view of a survey can answer these global questions. In addition, there are many other interesting scientific problems that can be illuminated by the survey data.

Source catalogs are a fundamental tool to answer the scientific questions posed above. Consequently, the generation of such catalogs, for both point and extended sources, is

one of our primary post-pipeline products and will be the initial focus of our analysis efforts.

We will produce a series of catalogs with the final products containing astrometric, photometric and morphological parameters for all sources detected above the 5s level. The first product will be a Bright Source Catalog (BSC), with the limit set at approximately the 20s level. The release of the first BSC is timed to

be available for Guest Observer proposals, such as spectroscopic follow-ups. The GLIMPSE Source Catalog (GSC will be released in increments). Each release should reach the 5s level, but will be restricted to a limited range in galactic longitude. Updated versions of released catalogs will occur periodically (about every 2 months) in order to incorporate any improvements in the subtraction of various background sources: zodiacal light, diffuse galactic background, extended red emission, galaxies, and solar system bodies. In addition, updates will incorporate the re-processing of the Basic Calibrated Data products (by SIRTF Science Center — SSC) due to changes in SIRTF calibration pipeline.

The ability to examine regions larger than an individual image can be crucial in addressing many science issues. The manner in which our survey will be produces images of adjacent fields which are likely to be rotated with respect to each other. The issue of geometric distortion must also be addressed. To aid in the analysis process, we will produce a series of mosaics at each observed wavelength that can be combined to generate images of a region of an arbitrarysize and location.

The database will be implemented in Oracle 8i probably residing on a Sun Microsystems Ultra 80 (2x450 MHz) computer. The core database schema will be implemented in the first 6 months of the program. The basic schema will likely consist of separate tables for each wavelength and post-pipeline processing level (i.e., Level A, Level B, etc.). Such

a structure provides an efficient and natural way to import/export data, as well as to update and propagate changes in the state of the SIRTF and ancillary data. Refinements of the database structure will be modified and implemented in a modular way as the post-pipeline products and tools are developed. Revisions and modifications during Phase 1 will be executed on a separate ``development'' database. Changes to the database can be easily accomplished by using products like Rational Rose Enterprise Software coupled with Oracle's Enterprise Manager. All revisions will be implemented in a scheduled manner to avoid conflicts with other ongoing development tasks. Backups of the database will be executed on a regular time schedule in both cold (shutdown) and hot (in-service) modes, providing both long and short-term recovery of all transactions.

In addition to database transactions from incremental processing of the data (i.e., reprocessing of individual frames, source extraction), allowances in data entry must be made for the delivery of bulk data sets from. Such bulk data loads will be accomplished through a series of modularized scripts (implemented in Perl) that will include a wide selection of data integrity, validity, and constraint checking (e.g. Level A processing). Only the data that pass the initial integrity checks will be loaded into the database using the automated load processes.

FRAME: A single 256x256 image (plus metadata), associated with data obtained a specific time and a specific wavelength. A unique observation identification (FRAME_ID) will be assigned in Level A processing.

Calibration images are associated with FRAME_ID, but many frames will share the same files. Calibration data should be a separate GLIMPSEdb table

V1.1 (30-Apr-2001): restructure table "LEVELA"

minor typo corrections