I was talking to one of our engineers yesterday (let’s just call him “A.”) about a certain senior engineer (let’s just call him “C.”) when A. revealed a very funny set of circumstances surrounding C. revealed by a third engineer (let’s call him “D.”).
C. is a programmer from Ye Olde Schoole – until recently, he coded in `ed`, telnetted to a Solaris 7 box to compile and code up until we disabled the last one and continues to read his email using `mail`. He does not play well with others and has a near-allergic reaction to source control programs such as CVS, Subversion, SCCS, RCS, etc. His preferred method of revision control is to:
- GZip all source files.
- Tar all gzipped files into a single .tar*
- Name the resultant .tar according to his versioning scheme
- Dump the .tar into a directory named after the particular project or code module
*Yes, I know this is an inversion. When asked why he didn’t use the preferred method of first tar’ing the files and then gzipping the resulting .tar, C. declared that older versions of Windows incorrectly handle/rename files with dual extensions (turning a .tar.gz into a .tar, for instance). When asked why he didn’t simply label the files with the acceptable alternative “.tgz”, developed specifically for just such circumstances, C. maintained his ignorance of such an alternative.
A. and D. were discussing a current project and the possible reuse of some older code when D. mentioned that he had remembered C. composing something that might be useful to the project at hand and thus felt compelled to “get down into the ‘tar pits'” to look for that which he sought. “Tar pits” (named for the endless directories filled with endless .tar files, any one of which could contain the desired, nay, prized! information in question) – haw! What an excellent coinage of phrase!