Stata The Stata listserver
[Date Prev][Date Next][Thread Prev][Thread Next][Date index][Thread index]

RE: st: Royston's mvis ado and version questions

From   "Nick Cox" <>
To   <>
Subject   RE: st: Royston's mvis ado and version questions
Date   Thu, 30 Sep 2004 15:49:50 +0100

Nick's first statement is very good 
advice, but his last statement is

No; this is not really what -version- does. 
There are lots of counter-examples. 

For example, the way that -set trace- works 
has been changed in Stata 8. As far as I can 
tell, no amount of version control will 
bring back the old behaviour. 

As another example, setting -version- does not 
restore bugs in a previous version. Good news
normally, but bad news if what Stata deemed a bug 
was a feature for the programmer whose program 
you are using. 

The best statement of what -version- does is, 
surprise, the details in the help for -version-. 
The key is more that old behaviour can be arranged
if it was thought that someone might seriously 
miss it, but there's a bundle of judgement calls


Nick Winter
> Note also, if you are trying to rewrite a program for an 
> earlier version of 
> Stata, it is crucial to run the program in the earlier version of 
> Stata.  That is, if you change the version statement on a program to 
> "version 7", and then run it successfully in Stata 8, that is 
> no guarantee that it will not crash when run by Stata 7.
> Stata 8 still allows some new features even when version is 
> set to 7 (or 
> less).  All the version statement ensures is that any 
> behavior that changed 
> from 7 to 8 will operate as it used to.

*   For searches and help try:

© Copyright 1996–2022 StataCorp LLC   |   Terms of use   |   Privacy   |   Contact us   |   What's new   |   Site index