My point, though, worded poorly in the middle of the night, was that even software with a fundamentally sound design and architecture can be awful to use because of bugs and missing (or hastily implemented) features that are actually essential, but weren’t ready in time.
And yeah, lots of poorly designed software is out there too, stable or not.
*cough* so many Linux GUI apps *cough*.
I am in full agreement!
To me, checking if an idea is any good –– say, a GUI layout ––, involves at least a quick-and-dirty check. Back of the envelope stuff; maybe a quick form-y test in Glade or QDesigner, or even a quick HTML form page; then an unit-test-like back-end test program to see if the UI provides sufficient data in usable form to do the job, while still being intuitive and easy for users.
When I implemented web pages a couple of decades ago, a common problem was that people had ideas on what they'd like to see on the web, but were very reluctant to provide the content itself, especially in universities (at least in Finland). They had an idea they felt was good, and pushed really hard for their implementation, before even making sure the content needed is available. But, it turns out the (preliminary/un-proofread) content itself is the test if the idea is good, and the visual layout (that most people would instead dabble with) is just a final touch. When the design idea is "sold" before the content even exists, you get web sites with poor content and unintuitive navigation and structure... but when the idea of the site is based on existing content, and about that content, you can get damn good design
and implementation surprisingly quick. In my experience, the text content is the hardest to obtain on any website project.
And remember how in the mid 90s, VR was supposed to become the next big thing, so everyone started working on “interactive” virtual rooms and crap, giving us “Where’s Waldo?”-like interfaces on CD-ROMs that used an insufferably long animation to “walk” you from the lobby to the Products room when you clicked on that door. Your CD-ROM drive would frantically seek for a few seconds before letting you inside, where you could click illegibly tiny thumbnails of their product boxes, which would fly off the shelf in another insufferably long animation to show you the front box art in dithered 256-color, 640x480 pixel glory. Then you could flip it over and read the back…
Oh wait, no, after you just force-quit the Macromind Director player in frustration, you just grabbed the catalog the damned CD-ROM came with to begin with…
Ha! I managed to avoid the VR craze myself! (I did a few Macromedia Director multimedia projects 1996-2005 or so (last version I used was 7.0); mostly for Macs, but also online for use with the Shockwave plug-in.)
We did use "room" and "wall" analogs in a couple of projects, but limited to 2D, because the environment just couldn't do 3D properly. (Well, I did do a couple of small (very limited pixel size) 3D examples for web use and such, even a couple of online games, so you could do 3D, just not VR or anything like that near full-screen in Director. Even "sprite" size had to be carefully restricted, to make animations 'smooth' (not horribly steppy).)
When somebody had a design idea, I did tiny "unit tests"/"demos" on a not-top-of-the-line PowerMac with placeholder visuals, to see if the idea works in practice or not, before allowing it to be accepted into the overall design: checking the idea is feasible in practice before letting anyone get excited about it. Worked really well, too: it's easy to avoid intra-team friction when everybody (including the artists!) can see and experience the issues in real life. It did mean I had to spend quite a lot of extra effort, but I was young and excited... Besides, those quick and dirty tests sometimes caused even better ideas to come up (when everyone got a better intuitive grip on the real-world capabilities of the environment), so I still consider it was definitely worthwhile.