R devel - Mar 2011 - rowsum

> with the entirely different rowSums, but it has been around
    > for a long time.)
    A lot longer than rowSums ...
    > Bill Dunlap
    > Spotfire, TIBCO Software
    ---
      This made me smile.  The rowsums function was originally an internal
    part of the survival package, used for fast computation of certain sums
    when there is a cluster() statement.  It was Statistical Sciences
    (S-Plus) who moved it to global status.  That is, they used it in enough
    other places that they decided to speed it up, took over the
    maintainance and ownership of the function (with my blessing), and
    ceased to label it as part of "survival" in the manual.  
      This "metabug" can't be laid at R's feet. 
    Terry Therneau

On Wed, Mar 30, 2011 at 1:56 AM, Terry Therneau <therneau at mayo.edu>
wrote:
> ? ?> with the entirely different rowSums, but it has been around
> ? ?> for a long time.)
> ? ?A lot longer than rowSums ...
> ? ?> Bill Dunlap
> ? ?> Spotfire, TIBCO Software
> ? ?---
> ? ? ?This made me smile. ?The rowsums function was originally an internal
> ? ?part of the survival package, used for fast computation of certain sums
> ? ?when there is a cluster() statement. ?It was Statistical Sciences
> ? ?(S-Plus) who moved it to global status. ?That is, they used it in enough
> ? ?other places that they decided to speed it up, took over the
> ? ?maintainance and ownership of the function (with my blessing), and
> ? ?ceased to label it as part of "survival" in the manual.
> ? ? ?This "metabug" can't be laid at R's feet.
The same process happened independently in R.  I ported the 'survival'
version of rowsum() to R, added it to base R in version 0.63.1, and
later it made it faster using hashing.

So perhaps it isn't entirely StatSci's fault, although it's likely
that R would eventually have added a rowsum() function for
compatibility.

  -thomas

-- 
Thomas Lumley
Professor of Biostatistics
University of Auckland