OK, let me try this post again... How demos are written is wicked important because: - Demos lay down the unwritten rules about how to structure apps and accomplish tasks. - They will be cut and pasted and form the basis of real apps. - They will be cut and pasted and form the basis of bug reports, email posts, articles, etc. (they come back to haunt you) - Newbies, not only to Pivot, but to Java, will learn about both Pivot and Java from examples and demos.
For example, in the EventBus project (http://eventbus.dev.java.net), someone was very happy about using the EventBus publish/subscribe API in a desktop music player and posted some examples. One example was something like this class that was responsible for fetching the files in a directory: public MusicFileFetcher(...) { public void getFiles(File dir[]) { .... EventBus.publish(new StatusBarEvent(...)); } } The strange thing is that main purpose of the EventBus is to decouple objects so that they don't know about each other, yet, here was a MusicFileFetcher that knew about the StatusBar (or at least its events). Of course, the first example on the main page of the EventBus project has a StatusBarEvent! I led this user down a dark road with an example that was too simple... I used a StatusBarEvent because I was intending to be real simple, real quick and real clear, but it turns out that people often don't see when code "cuts to the chase." So may I suggest a series of simple to complex demos that are reviewed for their correctness - "what would a good Pivot developer do?" Yes, I'm volunteering to help because Greg asked me to look into what pub/sub in Pivot would look like and having good demos would be most helpful to working it out. The demos I'm thinking of are of the "Pet Store" variety - real full blown apps. However, I tend to think that doing something real would be more pedantic. A full demo would exercise all the concerns of a UI - preferences, persistence, live updates, menus, status bars, skins, windowing, forms, etc., etc. I strongly believe that Swing would still be a major player if the Swing team had been tasked with creating apps beyond bare bones Paint and smiley faces. Real apps that I think might be useful: - A Pivot bug tracker - a domain everyone understands. - A stock tracker - most people get it, and it's suitable for liveliness - System visualization/management app - since it lends itself to richness. I'm thinking there has to be some real data out there that would look interesting - whois breakdowns, whatever. Thoughts? Michael -----Original Message----- From: Michael Bushe [mailto:[email protected]] Sent: Monday, August 24, 2009 8:03 AM To: [email protected] Subject: RE: Preferences Demo Demos are wicked important, and how you write them live with you forever. When users kick Pivot's tires, they will cut and paste large chunks of into their applications. How they understand and write Pivot will be most influenced by that work. For example, this post to an EventBus mailing list May I suggest a series of simple to complex demos (Yes, I'm volunteering to help.) Michael -----Original Message----- From: Sandro Martini [mailto:[email protected]] Sent: Friday, August 21, 2009 10:49 AM To: [email protected] Subject: Re: Preferences Demo Mhhh Ok, I have so to think at a simpler demo ... but you know, reuse (of what I've already written) is better for me :-) . Bye
