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

RE: st: New version of -descsave- on SSC

From   "Nick Cox" <>
To   <>
Subject   RE: st: New version of -descsave- on SSC
Date   Wed, 27 Apr 2005 14:57:13 +0100

Sorry, but it is more complicated than that. User- 
programmers have at least four different habits: 

1. User revises -prog- and keeps the same name. 
Other users might find, at worst, that (a) 
the latest version requires a Stata later than 
their current one and (b) a previous version that 
will work with their current one is no longer 

2. User revises a program and uses a number 
to indicate _its_ version, usually a low
integer 2, 3, etc. This might be a revision 
of an official command, somebody else's user-written 
program or your own previous program.  

3. User revises a program and uses a number 
to indicate the version of Stata that the previous 
version will work with. So -prog- is the 
current version, and -prog7- will work with 
Stata 7 (usually, but not invariably, also Stata 
8 up). Users here usually make sure that older versions
remain available, but typically do not 
update them separately. 

2 and 3 are different numbering conventions but 
in practice any ambiguity is rare and should be resolvable
by looking at the help, or even the code. 

4. User maintains own website with separate 
versions for older Statas. (Roger Newson, 
perhaps others). 

In terms of official or semi-official sites, 

* STB/SJ versions have always been maintained 
separately so that older versions remain 
accessible. As an Editor of the SJ, I would 
like to emphasise that we encourage publication
of Software Updates. Some members will affirm 
that whenever an Update is made advisable by 
a bug fix or major change, we usually push 
and pull users into overhauling their programs, 
cleaning up the code and modernising help files. 

* SSC will not hold multiple copies of -prog.ado-. 
Kit Baum will happily indulge users who 
wish other versions to be stored under different 

I am not clear what Richard is asking here. 
There is no special process for updating 
programs that mirrors -update-. Rather with 
e.g. -ssc- you are supposed to think about 
what you want overwritten. The worst possible 
situation resembles 1, in which in principle
you might lose a valuable program by overwriting 
it by a later version that you can't use. Even 
this need not be insoluble, as public appeal 
would normally identify a dusty copy lurking 
somewhere on someone's machine. 

In essence, STB/SJ and SSC are in the position
of being able to lay down their own rules, 
and in neither case can I recollect 
protest about such rules. 

Users putting stuff up elsewhere are fully entitled  
to establish their own rules. For my part 
I respect whatever conventions users decide
they want to adopt for themselves, especially 
as we are talking about free programs!


Richard Williams

> When Stata does an update, the old 
> programs just sort of disappear, e.g. version 9 users won't 
> be using the 
> Stata 8 version of -mlogit-, they'll just automatically have the new 
> one.  User-written routines don't have such a luxury, and so 
> as a result, 
> you see things like old programs being re-released with new 
> names, e.g. 
> prog becomes prog7 or whatever.  So, you might have a bunch 
> of different 
> programs called -prog- floating around, one for V. 6, another 
> for V. 7, 
> etc, along with otherwise identical programs called prog6, 
> prog7, etc.  I 
> don't know of any convenient way around this but it would be nice if 
> updating user programs behaved more like updating Stata is, 
> i.e. you get 
> the version of the program that is appropriate for your 
> version of Stata 
> without having to re-release old programs under new names.
> I'm also working on a user-written Stata 8 program, which, in 
> a few months, 
> I imagine I will rewrite for Stata 9.  I'm trying to write 
> this program so 
> it can be easily supported by Long and Freese's -spost- 
> routines, but if 
> the program has different names depending on the version of 
> Stata you are 
> using then support will be harder (e.g. there might be a 
> -prog-, which gets 
> renamed -prog8- and is replaced by another -prog- written for 
> Stata 9, 
> which eventually gets renamed -prog9- and is replaced by a 
> -prog- written 
> for Stata 10, etc.)  Maybe I need to tell them to reserve a 
> string of names 
> for me for future compatibility! (e.g. prog, prog7 prog8, 
> prog9, prog10, 
> whatever.)
> -------------------------------------------
> Richard Williams, Notre Dame Dept of Sociology
> OFFICE: (574)631-6668, (574)631-6463
> FAX:    (574)288-4373
> HOME:   (574)289-5227
> EMAIL:  Richard.A.Williams.5@ND.Edu
> WWW (personal):
> WWW (department):
> *
> *   For searches and help try:
> *
> *
> *

*   For searches and help try:

© Copyright 1996–2017 StataCorp LLC   |   Terms of use   |   Privacy   |   Contact us   |   What's new   |   Site index