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

st: RE: Req: Stata version in ssc archives


From   "Nick Cox" <n.j.cox@durham.ac.uk>
To   <statalist@hsphsun2.harvard.edu>
Subject   st: RE: Req: Stata version in ssc archives
Date   Mon, 22 Aug 2005 19:43:26 +0100

I note from this that you have privately modified 
my program -fixsort- and now want advice on why 
it still works.

The Statalist FAQ implies that once you modify a 
program, it is then yours and you are responsible 
for it! 

In this case, specifying -version 9- arose 
because of the internal use of the -clonevar- 
command. This was introduced in the lifetime 
of Stata 8, so anyone who has updated Stata 
8 after 5 October 2004 should be in your situation: 
they can hack at the program and have the outcome 
you describe. 

Setting -version 9- is far from the only solution here 
but it was the laziest at the time I wrote -fixsort-. 

The situation with -blist- is different. The 
description at http://ideas.repec.org/c/boc/bocode/s419301.html
is simply out-of-date in stating Stata 7, as the code
now there specifies Stata 9, although my guess
is that Stata 8 would be sufficient. The one 
detail here I noticed is whether a type must be specified
in -generate-ing a new string variable. I can't comment 
further. 

When you say that 

the required version represents the minimum
requirements needed by the command to work

this is indeed the main idea. But it is also 
possible for a programmer to specify, whether 
by accident or design, either a version that
is higher than this or a version that is lower
than this. This could arise for all sorts of 
reasons. Here are three examples. They are far 
from exhaustive. 

1. A programmer is developing a program under 
Stata 9. They thus believe that it works under
Stata 9. They have no time or interest in 
going back to Stata 8, or another version, 
to check that it would also work under that
version. Perhaps it will, perhaps it won't, but 
they can't bothered to check that out. 

2. Perhaps the help file was developed for 
a later version. That is, the program might 
work in an earlier version, but the help file 
won't show properly because it uses a later 
version of SMCL. 

3. A programmer could set say -version 8- 
and then accidentally use some 
feature of version 9 -- and not notice
that -version 9- is required. This 
is the opposite of #1. 

Otherwise put, programmers are busy people 
too and often make short-cuts or consider
their own convenience. 

Nick 
n.j.cox@durham.ac.uk 

Roberto De Miguel
 
> Making - ssc whatsnew - I found the command "blist"
> ............
> BLIST
>  module to list values of variables in as small a space as possible
>  Authors: Adrian Mander       Req: Stata version 7.0
> 
>  Revised: 2005-07-31
> ............
> 
> For my surprise, although it said: Req: Stata version 7.0, 
> when attempting
> to use it requested the version 9 of Stata.
> 
> Looking at the ado, I found that if I changed the line "version 9.0"
> meetly, it also runs in a previous version of Stata.
> 
> Something similar happened to me with the "fixsort" command 
> that although
> it require the version 9.0 can be used in a previous version changing
> the line that declares the version in the ado
> ............
> FIXSORT
>  module to sort variables and align in sorted order, with others fixed
>  Authors: Nicholas J. Cox       Req: Stata version 9.0
> 
>  Created: 2005-08-02
> ............
> 
> I always understood that the required version represent the minuimum
> requirements needed by the command to work. Isn't this correct?
> 
> Can somebody clear up this point?

*
*   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/



© Copyright 1996–2014 StataCorp LP   |   Terms of use   |   Privacy   |   Contact us   |   What's new   |   Site index