The Luxury of Ignorance: An Open-Source Horror Story (catb.org)

Posted Feb 26, 2004 22:22 UTC (Thu) by dlang (guest, #313)
Parent article: The Luxury of Ignorance: An Open-Source Horror Story (catb.org)

I attempted to setup CUPS on my network a few months ago and ran into similar problems. I ended up going back to lpd becouse I couldn't imagine any 'features' of this 'new, simpler' system that could possibly be worth the hassle of getting it setup.

ESR is right, every developer should take the software they write and get someone who knows nothing about it and just sit and watch waht they do to try and use the software. when they get really stuck (not just stalled for a little bit, let them try for help before stepping in). the purpose of stepping in isn't to get thigns going, but to try and gather info about problems further on in the software.

the hard part of this is that after doing a few trials like this that particular user will no longer be suitable for this sort of testing (they will have learned how you do things)

those who remember the heyday of Heathkit may remember that they did this type of thing with their kits and a person who got into their testing program would only go a handful of 'Beginner' level kits before they were considered 'to experianced' to be a valid test for them and had to move up to testing the next more complicated kits. This type of testing (and fixing the issues they find) would do wonders for software useability.

and by the way I don't exempt myself from having this problem. I have been guilty of producing software that seemed perfectly obvious to run (to me) and I couldn't understand why people were having so much trouble with it until I actually saw someone trying to use it and realized that they had missed the fundamantal structure of the software.