I suspect Richard is a little more zealous than the average Stata
user-programmer in this respect.
My own private default is usually Stata 8.2, as little things like where
you can put braces and how you can specify comments that I have
internalised changed with Stata 8. Also, anything graphical requires 8
at least. And if someone has 8, they should be able to -update- to 8.2.
But anything with Mata is automatically 9 at least.
But I've energetic and highly productive Stata friends who are publicly
generous with their Stata code, but draw the line at leaning over
backwards to maximise backwards compatibility.
Nick
[email protected]
-----Original Message-----
From: [email protected]
[mailto:[email protected]] On Behalf Of Richard
Williams
Sent: 26 January 2009 21:47
To: [email protected]; [email protected]
Subject: Re: st: RE: Stata version in user ados
At 03:57 PM 1/26/2009, Nick Cox wrote:
>* Even if an edited program appears to work satisfactorily, that is not
>a guarantee that your version produces identical results to the
>original.
My own little tidbit on this: I have Stata 7, 8, 9 and 10 on my
computer. I have routines on SSC that work on Stata 8. If I tweak
those programs, I do it in 10, but then I go back and run some tests
using Stata 8. I figure I don't want to zap the program
unnecessarily for Stata 8 users. If I find I have zapped it, I find
the fix is usually pretty straightforward, e.g. I have to find the
Stata 8 name for some function that got renamed by Stata 10; or use
slightly different code that accomplishes the same thing.
If I was doing really major changes though, I would probably just
create a Stata 8 version whose code was frozen and require that users
of the new version use Stata 10 or higher.
*
* For searches and help try:
* http://www.stata.com/help.cgi?search
* http://www.stata.com/support/statalist/faq
* http://www.ats.ucla.edu/stat/stata/
