ianCode

In the interest of consolidating my github accounts, I had them transfer the Dynamo repository to another account. Then I deleted that account. Whoops. For those of you who have visited the Dynamo site in the last couple of days, you’re welcome. It was my pleasure to entertain you with github’s 404 page. For those of you interested in the gory details of how I restored the repository, read on.

Luckily, git makes it really easy to restore a repository. I had to create a new repository on github. Like a newborn child, this repository came into the world with nothing, not even a README. Following github’s instructions for initiating a new repository on my machine, I stubbed out a local repository that was also empty, committed once, and pushed to github.  If you had visited Dynamo on github during this brief interlude you would have found a repository with only a README file, and you would have thought, “That Ian sure is brilliant. He figured out how to make Dynamo fit inside this one little text file.” Perhaps if I were that smart, I wouldn’t have deleted this entire project in the first place. I copied over the contents of the old repo, including the .git directory in the new local directory and did “git reset –hard HEAD”. Now, as far as git is concerned, my new repository has all of the tragic history of my old repository. All that was left to do was push. Something about pushing this reset repo to the remote repository made git angry however. Stumbling blindly forward, as is my wont, I then did “git push –force origin master” which is just as scary as it sounds. Presto! All of my files, branches, and my commit history are back on github.

Some of you may be wondering where the Dynamo repository went on github. So am I. It disappered some time yesterday and the folks at github don’t seem to know where it went. How do you lose a software repository? I will post again as soon as the repo is reestablished.

Thanks for your patience.

New Vasari installers are available, complete with Dynamo integration. Thanks to my friends at Autodesk for pushing this forward.

You can get the new Vasari installers here.

I’m working on a node that will access the sun and shadow settings for the active view. Currently, it will pipe out the azimuth, elevation, hour, month (oops I forgot days) and year. In order to make this work, I needed to make a dynamic model updater which responds to changes in the sun and shadow settings. That’s all working. What doesn’t work is getting the node to update automatically after the dmu runs. If I automatically tell the downstream nodes to update after the dmu, I violate my transaction lifecycle. I’ll have to give that some thought. I’m beginning to worry about using adaptive components in any dynamo workflow. They are so extremely slow to update that anything but the most simple model will take ages to rebuild.

WPF is, and I hate to admit this, one of the coolest frameworks I’ve used for doing UI work. Once you get the hang of it. I’ve implemented some capability around data binding that will clean up the UI code considerably. For now, just a couple of changes. I made the bezier paths scale correctly against the relative Y offsets of nodes so connectors aren’t out of whack when nodes are close together. And I made the shines on the nodes scale correctly. Next up, I’ll get the nodes to automatically re-size around their content and I’ll get the node coloration to work once and for all.

I’ve added the ability to query a divided surface. You can select a divided surface and generate a data tree from its associated panel instances. I wanted to do some attractor-type stuff with the panels so I also edited the distance between points functionality to recognize not only data trees full of reference points, but also data trees full of FamilyInstances. This way you can generally calculate the distance between a family instance’s location and your attractor. What I learned from this is that adaptive components are dog slow to update. A small 10×10 grid of simple components on a surface takes 10+ seconds to update.

I haven’t had a chance to look at dynamo in the last couple of weeks. In anticipation of a show and tell tomorrow, I’ve made some small updates. I changed node dragging to use the left mouse button because I was having to jump through hoops to make the right-dragging solution work with context menus. The installers have been upgraded too so they should actually work, which is nice. And dynamo now has its own button in the ribbon courtesy of Matt Jezyk. I also turned off drop shadows on the nodes because, although it looked sexy, it was terribly slow. I was afraid this would happen. Raster effects are still software rendered in WPF, so once you have about 10 nodes in the scene with shadows it starts getting clunky. Not acceptable. This is another place where Apple got it right, standardizing on UI effects being OpenGL-based so you don’t have this kind of lame problem.

Was working into the wee hours of last night getting Matt and Zach’s updates to dynamo from AU into the main line. It’s not quite done yet, and I have to add a couple other changes. I hope to finish that tonight so we have one clean line to develop on. I couldn’t get GitHub’s pull requests to work, but that would be the best way for people to contribute: by forking and issuing pull requests. If any other dynamo developers out there would like to help troubleshoot this by forking and issuing pull requests for testing, I’d be grateful.

When Matt Jezyk and Zach Kron from Autodesk contacted me a couple of weeks ago to inquire about using Dynamo in their lab “Energetic Supermodels” at Autodesk University, my response was…”It’s your funeral.” It’s not that I didn’t relish the opportunity for the Vasari wonder twins to demo my software in front of about 60 users slavishly devoted to all things parametric, it’s just that I feared that people would see it for what it is: a rough, finicky, somewhat ugly piece of software that doesn’t yet do much. Judging from the feedback they got after the class, and the number of people that accosted me to discuss the future of Dynamo, I’d say that Matt and Zach must be Jedi. You know the scene…”These aren’t the droids you’re looking for…” Execept here it was, “This is the software you’ve been waiting for…”

I’m really excited that people think Dynamo might turn into something interesting. I’m collating all of the cool ideas that I discussed with users at AU to add them to the road map on gitHub. I invite you to go to the Dynamo site on gitHub, download the source or the executeables and open an issue if there’s a feature you’d like to see implemented. I’m in the process of putting together a tutorial for how to create your own nodes. In the meantime, if you’re a developer and you want to take a crack at coding your own components for Dynamo drop me an email, and I’d be happy to help you get up and running.

Thanks again to Matt and Zach for some great exposure. Looks like I’ve got some work to do.

Why did I have a grid in Dynamo? No clue. Now I don’t. I’m also playing with WPF drop shadows. I believe these are still software rendered so they might get really clunky when workbenches get large.

I also made a significant change to the way that nodes update their downstream connections. Previously, a node which was connected multiple times to the same node downstream would call update on that node as many times as it was connected. This was stupid. It now triggers all of the connectors to send their messages downstream, but defers updating until all the messages are sent. Then it updates a node only once, even if it has multiple connector coming from the same upstream node.