* means if intresting findings pop up I plan to do a dedicated blog post for that item :)

after Lewis's seminar;

- test pl/sql commit inside loop anomaly; inside a db call such as an anonymous pl/sql block commits dont always become sync writes to lgwr. *
- test if autotrace traceonly on 10gR2 doesn't run for selects but just does an explain plan
- check the contents of 11g's v$sql_hints view
- test latching amounts and compare for below views on 10gR2
v$segment_statistics -> v$segstats
v$sqlarea -> v$sql or v$sqlstats 
- test select dbms_support.mysid from dual
- test if "alter session set sql_trace true" works without a "=" sign
- test if ASSM effectively goes for 16 freelists on 10gR2.
- check the column differences between 11gR1 and 10gR2 for v$sql
- check www.iozone.org and 11g's dbms_reosurce_manager.calibrate_io for dbms_stats.gather_system_stats. *
- test hint /*+ nocpu_costing */ to switch system stats of for a single query.
- test dbms_stats.convert_raw_value
- is CLUF needed or used by CBO for other purposes than INDEX RANGE SCAN access path. *
- test 10053 event on 11g for count(1) -> count(*)
- can CBO benefit from constraints if they are NOVALIDATED but ENABLED. *
- test sort merge join with two indexed row source to see 10gR2 will still need a sort step even data comes from index as sorted. *
- test /*+ IGNORE_WHERE_CLAUSE */ 
- test if "_optimizer_ignore_hints=TRUE or /*+ IGNORE_OPTM_EMBEDED_HINTS */ hint also disables useful hints like append, parallel query or qb_name
- investigate for a perl or shell script parser of event 10053 trace file since it is not so "human readable" when doing joins. a summary report output will guide a lot. *


after Woods's presentation;

- what is new in 11gR1 for tkprof, statspack and 10046 sql trace? after ash, awr and addm is Oracle still investing in these free of charge utulities?
- if there are enhancements run 11g's tkprof and statspack for 10g and compare the output results. *
- can I develop a similar 11g's SQL Performance Analyzer like software for the assistance of a 9i to 10g migrations. bind values are first challenge. *
- "timesten cache is 10X faster after 11g" saw in one of Wood's slides, do research for some numbers. *
- investigate for SQL developer 1.5's new features like sql trace reports, pl/sql profiler report and awr report. *
- after 10.2.0.4 11gR1's first patch 11.1.0.7 is coming, apply and test on vmware. *
- can I use 11gR1 active standby option for a 10gR2 primary, not for HA needs but reducing the load of reports and backups of production with a near real-time unstopped physical data guard. *
- test data masking for columns option for sensitive data(credit card number kind of) on test systems which are duplicated from production. *