On Thu, Aug 28, 2003 at 11:53:14PM -0600, Jacob Anawalt wrote: > Just my 2c on the rewrites from scratch. I also agree, and have one > theory on why it happens anyway based on my experiance with other > programmers who I've seen assigned to improve on existing code. > > I think often 'rewriters' give up before getting their head around the > existing code and figuring out why it works the way it does. They feel > it is taking them longer to understand the code than it would to write > it their way.
I've done a rewrite from scratch where the first step involved completely understanding everything the old code did, nonexistent comments, bugs and all... because the old source was about the only "documentation" on exactly what the system was supposed to do. Without the rewrite, it would have been possible to fix most of the bugs and work around the rest, but it would still have been horrendously inefficient - it was processing large batches of data, and the rewrite got the run time down from 6 hours to 20 minutes. I do agree about swapping old bugs for new ones... -- Pigeon Be kind to pigeons Get my GPG key here: http://pgp.mit.edu:11371/pks/lookup?op=get&search=0x21C61F7F
pgp00000.pgp
Description: PGP signature
