Nevertheless, Shriram, I think you're picking on one of parts of the system that follows these ideals the best. Better to pick on slideshow. :)
Robby On Thu, Aug 19, 2010 at 9:22 AM, Shriram Krishnamurthi <s...@cs.brown.edu> wrote: > You put it through Spidey. If you didn't have any code, what would > you have put through Spidey? Nothing. > > The code came first. It had long since been deployed before you did > this. > > Performance testing came when we set out to write a paper and wanted > to do measurements. They were initially designed only to measure > performance, not to actually *stress*. They morphed into stress tests > later. By that time, too, the system was long since deployed. > > Shriram > > On Thu, Aug 19, 2010 at 9:59 AM, Matthias Felleisen > <matth...@ccs.neu.edu> wrote: >> >> I, with Paul's help, worked the entire Web server through MrSpidey and >> eliminated all but those checks that Herman-Meunier later showed how to >> eliminate with their ICFP paper. That's far more than testing in some sense >> even if it doesn't show that it serves. >> >> Paul set up automatic stress tests back then already. We reported them in >> our paper. >> >> I call this judgment inaccurate. >> >> >> >> >> >> On Aug 19, 2010, at 9:56 AM, Shriram Krishnamurthi wrote: >> >>> Neither would the PLT Web server. I'm pretty certain Jay's own edits >>> to it were not preceded by tests. >>> >>> Jay's changes to the PLAI language most certainly were not, because >>> the current PLAI that is bundled with DrRacket is broken. >>> >>> (Sorry, Jay, but it's your prose.) >>> >>> Shriram >>> >>> On Thu, Aug 19, 2010 at 9:50 AM, Casey Klein >>> <clkl...@eecs.northwestern.edu> wrote: >>>> On Thu, Aug 19, 2010 at 8:44 AM, Sam Tobin-Hochstadt <sa...@ccs.neu.edu> >>>> wrote: >>>>> On Tue, Aug 17, 2010 at 3:57 PM, Jay McCarthy <jay.mccar...@gmail.com> >>>>> wrote: >>>>>> Please comment. >>>>> >>>>> I think that this: >>>>> >>>>> "Your first task when changing old code is to build an adequate test >>>>> suite to ensure you do not introduce new mistakes as you attempt to >>>>> improve it. Thank you for improving the world for future generations!" >>>>> >>>>> is too demanding. There are enormous areas of our code that don't >>>>> have a test suite. How comprehensive a test suite do I need before >>>>> changing slideshow? Or scribble (which has a test suite for the >>>>> syntax, but not the language)? >>>> >>>> Robby and Matthew, would Slideshow exist today if you'd be expected to >>>> build it with this process? >>>> _________________________________________________ >>>> For list-related administrative tasks: >>>> http://lists.racket-lang.org/listinfo/dev >>>> >>> _________________________________________________ >>> For list-related administrative tasks: >>> http://lists.racket-lang.org/listinfo/dev >> >> > _________________________________________________ > For list-related administrative tasks: > http://lists.racket-lang.org/listinfo/dev > _________________________________________________ For list-related administrative tasks: http://lists.racket-lang.org/listinfo/dev