The new tools need a new UI. The old one is hopelessly mired in ancient ways of doing things (isn’t that astounding to say about .NET code). To focus on the core features, the UI took a back seat with a temporary UI with lots of shortcuts. Several of these shortcuts were just wrong…
One is to organize by template type (core object, child list, readonly object, readonly list, hieararchy) and within this organization include files for editable and handcrafted files. I haven’t converted this project (and may not) to use partial methods, so I’ve got a three class strategy – base, derived designer, derived handcrafted. I have overwritten the files in the editable about 65 times because of this organization.
Which brings me to the second shortcut we should never have taken: This version is the first harness I’ve used in over six years that overwrites handcrafted files. I have no clue how you folks that use most existing tools survive. I’m just dying under the rewrites and become totally dependent on source code recovery – something I almost never do with normal code.
Unless you've got a good reason to do it otherwise, organize as Project/Generation Group/Template Group NOT Project/Template Group/Generation Group. IOW, use Project/Autogenerated/RootObject/files and Project/Editable/RootObject/files NOT Project/Root Object/Autogenerated/files and Project/RootObject/Editable/files. This allows easier file management such as deleting directories and managing source control.
I'm not planning on fixing this in the temporary GenDotNet CTP user interface because I hope to replace that UI with a configurable one that gives you full control over your project structure.
OK, I’m blogging again. Why the silence? The last month was tremendously painful as I tried to keep up with clients, train a puppy, get slides ready for DevConnections next month, and tackle getting the first GenDotNet tools CTP out. The last proved to be about an order of magnitude more difficult than I expected. As such, I need to set expectations for this CTP.
I’m not sure how interesting the last is, except to say it’s amazing the things new sets of eyes see. Particularly with a large project where it’s tough to get a toe hold. And it is big. There are over 1000 files, over two thirds of which are legacy stuff temporarily in play.
This is a CTP – for the things that it focuses on, it’s beyond the proof of concept that I’ve put in my blog. But for many, many things, it’s duct tape and baling wire to get something standing up to test. I’ve written on this in the Codeplex/GenDotNet opening page, which I think is a more appropriate place.
To be honest, I wouldn’t have gone public if CodePlex didn’t demand it. But then it was probably a good thing for me and the project to have a hard deadline. The project is now up and ready for people to shoot at it and contribute. Just don’t shoot at it because its grossly incomplete. Shoot at it because of the parts I focused on – which I’ll talk about more in the next few days.
Look for far more usable drops in the next two months as I and other people have a chance to focus on bringing our vision for other portions of the project to fruition. I'm really excited about where we're going, but can't help but contemplate the image of my son wrinkled and dirty when he was first born to know heading off for graduate study at one of the finest math departments in the world (he hasn't made a decision yet on which of them). This is breathing. It's started, and it will grow. And I promise, it will not take 21 years to come into its own.
And, the puppy is beautiful and very, very smart. Photos bound to appear soon.