Re: st: Convention for replacing an existing routine on SSC with a revision that has major syntax changes.

[David Elliott <dcelliott@gmail.com>]

Thank you to the usual suspects...

I actually like Nick's suggestion #2 best.  I'd like to keep the name
chunky because of the existing references to it in Statalist and many
other web locations.  I would then rename the old version and place a
reference to it in the help file.  (I'm considering renaming the old
version "clunky".)  It is not like I would be removing a line-indexed
file-splitting program from the Stata repertory, Roy Wada's -chewfile-
performs a similar function minus the ability to stick headers on each
of the chunks.

If the utility were more widely used I might consider incorporating
the old routine and it's syntax with testing for the version based on
the syntax employed.  However, part of the virtue of the new routine
is how uncomplicated it is in relation to the old one.


As an aside - using Mata's file I/O routines was a challenge to manage
at first but it is well worth it because one completely sidesteps
problems with " ` " or other problematic characters in the input file.
 It is really the only way to roll...

DC Elliott

On 25 August 2010 11:16, Nick Cox <n.j.cox@durham.ac.uk> wrote:
>
> Yes, that's it. The two versions can coexist with a branch on c(version). Or you can test for one syntax and given a syntax error allow a test for the other syntax.
>
> Nick
> n.j.cox@durham.ac.uk
>
> Austin Nichols
>
> David Elliott <dcelliott@gmail.com> :
>
> If the original version does not use Mata, it may be version 8 compatible; see
> ssc d ivreg28
> for how the authors of ivreg2 differentiate the old and new versions.
>
> On Wed, Aug 25, 2010 at 9:55 AM, David Elliott <dcelliott@gmail.com> wrote:
>
> > The subject line above almost says it all.
> >
> > I have completely rewritten my file chunking routine -chunky- to use a
> > different and far more efficient chunking strategy using Mata for the
> > file I/O.  It is basically a completely different program.  I am
> > currently calling it -chunky_mata- to during testing to distinguish it
> > from -chunky- but it is my intent to completely supplant the original
> > routine which, quite frankly, was an ugly kludge for which I am almost
> > ashamed to claim ownership.
> >
> > The syntax has changed from:
> > chunky using filename , index(#) chunk(#) saving(filename[, replace])
> > to
> > chunky using filename ,[ [peek(#) analyze] | [chunksize(#.#)
> > header(string) stub(string) replace]
> >
> > What is the convention for replacing a routine that will "break" any
> > previous programs written using it?
> >
> > I appreciate your guidance.
>
> *
> *   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/



--
David Elliott

Everything is theoretically impossible, until it is done.
Progress is made by lazy men looking for easier ways to do things.
 -- Robert A. Heinlein (American science-fiction Writer, 1907-1988)

*
*   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/