In addition to all the data stored in this archive on this (UNIX-based) server the MCDC maintains an 
auxiliary (Windows-based)  site with additional sf32000 data sets.  These include a complete 
national collection of census tract and block group level summaries.  There is an xxtrbgs library 
with 50 datasets, one per state.  They have names such as txtrbgsph - which would be the data set 
containing P and H tables for the state of Texas at the census tract and block group summary 
levels (140 and 150 SumLev values). Note that we do NOT have CT tables at the census tract level. 
The Bureau publishes these but we did not deem them worth the extra resources needed to store 
them at this level.  There are no such tables published at the block group level.  These data 
sets can be accessed via the MCDC's sf3tabgen application (coming July, 2009) and when you click 
on the table drill-down links in a dp3_2k profile report.  But they cannot be accessed directly 
(for now) via Uexplore/Dexter.  There is, however, an xxtrbgs subdirectory of the sf32000x 
collection which IS accessible.  So you can access the sf32000x extracts derived from these 
complete table data sets.

There is also an xxsl090 library (i.e. directory with related data sets) on the auxiliary site.  The 50 
data sets in this collection have names such as wyph - Wyoming P and H tables and the split block 
group summary level, SumLev=090.  Summary level 090 is important because it is the smallest 
geographic level for which sample (e.g. SF3) data are available.  We can use these data to do 
custom aggregations to user-specified geographic areas.  We also used these as the source for the 
corresponding sf32000x data sets in the xxsl090 subdirectory of sf32000x.  Those are the data sets 
that we use in the CAPS report to build the circular area summary data; used because they are the 
smallest geographic units available.  

jgb, June, 2009.