TomWaltz wrote:
The AC library was limited enough that our company quickly came to the conclusion that instead if relying on GS and Objects Online, we would make our own.
We got the old (1998) SmartParts library after about ten minutes of using the AC (5.1) library. If you think it's bad now...
I started doing GDL a year after that. A year later, we were using only my own windows and doors. I'm on about my fourth version of them now.
The only parts we use from the AC library are things like furniture, appliances, faucets. We view these items as symbolic, and they don't have to be perfect, which, hoo-boy, they aren't. (Personal peeve: sinks can't operate in SEOs, so they can't cut the counter slab.) We won't invest time in making a perfect sofa object. (For other users, sofas might be critical.)
(Near-)Perfect windows, doors, structure elements, symbols, and annotations
are
required, and we have invested a lot of time in creating/maintaining them. Pretty much every object in our projects that isn't furniture/fixture is home-grown.
Yes, the libraries are bad, although they've been worse. Things like the interior/exterior sash material make me wonder if they've ever talked to a single american user, though I know they have. Truly bafflingly bad, and so easy to fix. My personal GDL hobby-horse is scale sensitivity, and the AC library makes very poor use of it. But they could fix all the big problems, and the libraries still couldn't meet the needs of thousands of users. User customization will always be needed.
It's a shame, since GDL is really an amazing technology.
I am permanently amazed that GS is actually sort of middle-of-the-pack in appreciating GDL's power. The power of objects isn't geometry (you can get geometry anywhere), it's parametrics. GDL is really the only tech for building custom parametric models from scratch.
The investment I describe above, or Kitchen & Associates' investment, is not feasible for smaller, lower-margin, or solo-practice type firms. GDL isn't exactly hard, but starting it is a little hard, and it uses different skills from architectural, or drafting, practice. Yet customized libraries are critical to using AC to its potential IMO. Even if the AC libraries were great, they can't possibly meet everyone's custom needs. Look at all the parts. Look at all the
kinds
of parts. On top of developing AC itself, the AC library is hugely complex.
What is needed is a graphical method for creating parametric objects. It needs to be easier for architects to make building elements that respond to the environment. If you want more people to make their own objects, and stop griping about the shipping libraries, the price of creating our own has to come down.
Objects are critical to the success of the AC platform. GS has a choice: offer perfect libraries, or lower the bar for everyone doing it for themselves. The first isn't realistic, so here we are.
With the development investment needed to make half the user base happy with the libraries, they could instead develop an environment to radically simplify the whole process, and let more people do more of it by themselves (and working together).
The other thing that never seems to take off is manufacturers' creating their own stuff. Why on earth am I scripting a Marvin window? Why should Graphisoft? Again, if it were cheaper (easier) there would be more of it. Manufacturers also demand intellectual property protection. Whether they demand it reasonably is another matter, but GS should provide this facility for people who want it. Inability to control copying is a deal-breaker in the current IP climate.
Update the tools, and work with the manufacturers, and the end-user library problems will quickly become tolerable.