I have some updates about the Shion 2 release that I wanted to share. As regular readers of this weblog know, I’ve been aiming for an end-of-January release date for Shion 2. I want to get the product in people’s hands so that I can begin getting the necessary feedback to go through a few quick revisions in the next couple of months to solidify the software and begin building other products upon it.
Shion 2 has been in private beta testing for the last several weeks and this testing has been useful in ferreting out a variety of issues and bugs. However, I haven’t been able to assemble as large of group as I would have liked.
Back in the open-source days, I would have reached a stable point and pushed out a new release. In an open-source project that’s being given away for free, this seems like a fair arrangement. The users get free software, while the developer gets help testing it.
However, since I’ll be selling Shion 2 as a commercial product, I realize that I need to release something that is absolutely solid and robust before expecting people to write me checks (or some equivalent). In all honesty, the software’s been working great for me and most of my testers, with a manageable number of issues to address before the end of the month.
“It works great for me and my friends — let’s ship!” Famous last words. I’ve been writing and supporting software for long enough to realize that there’s all kinds of funky stuff that happens out in The Real World™ that testing simply fails to catch. Given the additional complexity introduced by interacting with external X10 & INSTEON hardware, it would be quite unexpected if the vast majority of new users did not experience one problem or another (assuming a Jan. 31 sale date).
Since it’s extremely important for the future success of the Shion platform that it make a good impression before anyone forks over $35, I’m changing the Shion release plan a bit. My original promise was to have something out to the mass market that anyone could begin using by the end of the month. That will not change. What will change is the process by which that happens.
Instead of putting the software on sale on Jan. 31 as originally planned, I will instead be launching an open beta for any interested users. This will consist of some full featured builds of Shion 2 (no 3 device limit) that expire at some future date. I will either release new betas before the existing betas expire, or finally launch Shion 2 as a paid application.
Everyone will be free to participate in this beta and begin using the software in their own setups. These will be the fundamental rules governing this beta test:
1. I welcome feature requests, but the focus at this point is on reliability and robustness. Please send me any feature requests, but be advised that most new features will not be included until a future release.
2. Any beta tester that reports a new specific problem that I fix in a beta release will be entitled to a free license. For each release, I’ll publish a blog post describing the known issues and testers will be free to publicly report issues in the comments to that post. (This helps everyone see what’s been reported before.)
3. As suggested by Joel Spolsky, the clock for a final release will reset when I release a new beta. New betas will be released no more frequently than once every two weeks. If four weeks elapse without any significant new problems being reported, the beta test will end and I’ll release Shion 2.0 proper.
4. After Shion 2.0 is released, I will begin working on new features and functionality. Consequently, if you want Shion to support your sprinkler, thermostat, or other hardware, it’s in your best interest to really give the application a hard time and to report any problems. The sooner the foundation is solid, the sooner we can begin adding wings to the house.
5. Similarly, the more people testing the application, the better. I am going to rely the current beta testers to get the word out to other potential users by way of forums, Usenet, or other channels. The more users that we can include in this test will decrease the overall beta schedule as problems are identified sooner.
At the moment, I have a few issues that I’ll resolve this week, but I’m planning on pushing out the first open beta this weekend. My goal is to go through enough cycles to have things ready for sale by the end of next month. If the testing finishes before then — awesome. If testing extends into March (or later), that’s fine by me as well. The important thing is that Shion 2.0 be rock-solid so that it can begin to support bigger and better things in the future.
Thank you for your patience and I look froward to your participation.