kchax4377 / lucene-net-benchmark Goto Github PK

Change the ReportTask so it factors numbers better.  For example, have it turn 
seconds into minutes, or memory into gigs, when appropriate.

Check on the files that declare threadStatic variables.  I ended up removing
the static state, but maybe that wasn't right.  It looks to me like the intent 
was to have a static value created when the thread starts, and then to copy the
values to local variables for local manipulation.

Even when you preform new rounds, the added docs number is always a grand 
total; look into this behavior.

The Extract Wikipedia code has gone through a line by line modification of the 
code to C#; however, it has not been tested nor has the XML reader been 
programmed.  Further work needs to be done to fix this.

Add the ability to create a simple chart pro-grammatically and use data in a 
dynamically generated spreadsheet.

Add a data point to the ReportTask that allows you to report statistics about 
the Lucene index at each run (eg. Total File Size and Total Files).

Add a feature that allows you to define header cells and note them in format 
(e.g. make them bold).  Turn off word wrap for header cells and have them auto 
size.

Want to make this change to be more inline with the Lucene.NET implementation.

The quality scorer has gone through the initial read through for changes from 
Java to C#; however, the code has not been run or debugged.  These changes 
should include placing the code in a separate project and turning it into a 
command line tool.

create some way for tests to be paused (possibly even shut down) and then 
resumed.  This is usually an issue for stats more than anything else.  It would 
be nice to be able to create tests that run in parts, but still be able to have 
total summary data.

Sometimes Excel doesn't like number (particularly 0) in string columns.  Look 
into how to fix this behavior so that columns that have "-" and number data in 
them are formatted correctly automatically.

The column widths are odd on the reports.  Add a feature that allows you to 
define the column widths so the log and the reports have different formats.

Another step just to stick to standards.

Just need to go through with the object designer and do some documentation