Y2K retrospective / was Re: Algol vs ...

Ian King IanK at vulcan.com
Fri Feb 12 14:39:13 CST 2010


> -----Original Message-----
> From: cctalk-bounces at classiccmp.org [mailto:cctalk-
> bounces at classiccmp.org] On Behalf Of Russ Bartlett
> Sent: Friday, February 12, 2010 9:19 AM
> To: General Discussion: On-Topic and Off-Topic Posts
> Subject: Re: Y2K retrospective / was Re: Algol vs ...
> 
[snip]
> 
> In the early days when we wrote programs they were written onto coding
> forms.  Once punched onto 80 column cards we would check each card to
> ensure that had been key punched correctly.  Checking an 8000
> statement  program took time.  Once checked we would have it listed
> (used little machine time).  We would then dry run through the program
> looking for logic errors.  Having done that we would have it compiled.
> Typically there would be a couple of development slots or so a week for
> testing so we had ensure that we had done due diligence.  If it
> compiled we would schedule a test slot and run against test data.
> Debugging consisted of analyzing dumps and correcting the code.
> Contrast this against interactive source debuggers.  Today machine time
> is inexpensive and many compiles and test shots may be performed in a
> day.  "Workbench" tools allow the programmer to run
>  their program without even hitting the mainframe.  A totally different
> world.
> 
And yet, despite those inexpensive tools and resources, one of my most persistent challenges as a test manager at a certain large software company was to convince developers that they should do local builds (which could be done incrementally) before checking in their changes.  Some devs were good about 'buddy builds' and the like, but usually only after some test manager they had encountered in their careers managed to convince them that the small amount of time it took to do a build was minimal cost compared to the productivity hit of a broken main build.  

One of the tactics I used to educate one team was a public recognition of 'he/she who broke the build' - a rubber chicken hung from that dev's office relight until either the end of a week or until someone else broke the build.  One dev came to my office steaming mad about it - he thought this was childish and a waste of his time.  I pointed out that by breaking the build, he was wasting *everyone's* time, including his and mine.  He eventually became one my staunchest supporters regarding QA practices.  

And yes, I had my experience with punched cards - FORTRAN IV.  One of the best habits I ever developed was sequence numbers.  -- Ian 



More information about the cctalk mailing list