Re: Wish list: clear stata result screen

[Phil Schumm <pschumm@uchicago.edu>]

> A small addition to the Stata Wish list: a command for "Clear stata result
> screen".  I think there is no such a command (I looked hard for it).

Unless I missed it, there has not yet been any indication of why 
someone would want to do this.  Let me suggest two possible reasons: 
(1) after poking around awhile with a dataset and/or changing one's 
mind about an analysis, one wants to look at a fresh, clean set of 
results without seeing what came before (I won't comment on the 
psychology of this, but only admit that I occasionally feel this way 
myself), or (2) one wants to print a clean analysis without previous 
mistakes and/or false starts.  There may be other reasons, but off 
the top of my head I can't think of any.

Note that both of these objectives can easily be accomplished by 
using the viewer window together with log files.  For example, I have 
a little personal command called -newlog- which does nothing more 
than close the current log (if one is open) and open a new log in 
/tmp (these files are automatically named with a datetime stamp so 
that all I have to type is the command name itself and so that each 
successive filename is unique).  I also have a command called -vlog- 
which opens the current log in the viewer (or, alternatively, 
refreshes the contents of the viewer with the current log).  So, if I 
get to a point in a session where I want to examine (or to print) a 
set of results by themselves, I just use -newlog- to start a new log, 
issue the commands whose results I wish to view (or print), and then 
use -vlog- to open the viewer (I can then use the "Print Viewer..." 
command from the File menu if I want to print the results).  Very 
simple, very fast.  Moreover, if I ever do need to go back through 
the contents of my results window, I don't have to scroll through a 
bunch of blank lines.

One final note.  While it is often useful to perform part or all of a 
Stata session entirely interactively, I would argue that any real 
data analysis should always be done via one or more do-files.  This 
is not to say that there won't be interactive moments, but once you 
separate the wheat from the chaff, it's almost always a good idea to 
move the resulting code into a do-file so that it can be replicated 
later on.  In fact, for those with a tightly integrated text editor, 
the distinction between interactive and do-file begins to vanish. 
The point to all this, though, is that once your final analysis is in 
a do-file, it is then simple to open a new log, execute the file, and 
view (or print) the "clean" results.


-- Phil
*
*   For searches and help try:
*   http://www.stata.com/support/faqs/res/findit.html
*   http://www.stata.com/support/statalist/faq
*   http://www.ats.ucla.edu/stat/stata/