A ten hour flight from the Bricsys Int'l Conference in Amsterdam. Every bit of reading material has been consumed. Plenty of time to develop crazy plans, like how contenders like BricsCAD could upset the reigning heavyweight champion, AutoCAD.
Yes, it must be capable (See Pt 1, C is for Capable), but that will mean nothing if the software is not dependable...bulletproof, even. It must be able to take everything thrown at it and not wince. Customers know they can go with the market leader and they will be safe. It will take only one bullet to be shot down.
Years ago, when I was leading a CAD division of a contender, we went on tour. So proud to show our new baby to the world. We had stacked the latest, greatest release with all sorts of features. You'd have to be a fool to pay thousands when we had it all for a couple of hundred bucks. We got into the office of a major engineering publication. The product manager was on a roll, going through a litany of cool features. Oh, we were on a roll. Until the program crashed. Seeing the blue screen of death, I tried to distract the editor with some spiel about company philosophy. The editor was attentive, polite. The demo was recovered and completed. But we all knew it was over. Neither of us had the courage to open the magazine for the rest of the year.
Making a software bulletproof takes a commitment. A company has to have excellent quality assurance. Before the software ever gets out the door. After that, it needs to be responsive to whatever the users find, and fix it. It's not glamorous, like having all the cool features. And it takes lots of money. Big companies run endless tests to determine fallibility with large files, ferret out incompatible geometry, untranslated styles, more...and for AutoCAD contenders, the biggest and most important questions may be "How does it work with my 3rd party applications?" or "will my LISP routines work?" This is life and death for companies who have had years with AutoCAD. Getting a program to work like AutoCAD is one thing. Getting program to work with everything that has grown around AutoCAD could be the real challenge.
Bricsys has told us that how much effort it took to match AutoCAD sheet set capability, almost delaying their ship date of BricsCAD 13. Bricsys is trying really hard. Could it just be that little engine that could?