I certainly didn't mean to imply that you should stick with HS and only write plugins for it. I merely took exception to one of your statements about plugins. You said that you need to pay for plugins to get functionality. I do not agree with this (in general). Nor do I like the current state of plugins in HS and where things are headed.
Some of the issues that I think are important in any new development (and these have been mentioned before):
1) scripting - It should be extendable by the end user without requiring the purchase of a compiler (or even downloading and figuring out a free one from MS - these are not always the most straightforward for people with limited experience). There is no way you (or a plugin developer) can think of everything. Even if an end-user only writes simple scripts that are triggered by events, scripting is still a tremendously powerful tool. I have very few events in HS that do not run a script of some sort (i.e., just turning on a light and saying a canned statement just isn't enough).
2) client/server - separating the main functionality from the user interface is important. HS is moving toward this in 2.0, but they are emphasizing the web interface only. I actually do almost all my setup using the native interface in the current version of HS. I use plenty of web pages, but they are my own pages and not those created by HS. Something similar to the native interface, but running as a client and on any machine (include the server machine, of course) is very useful. The web interface can also be convenient at times. Allowing multiple links into the server is a good thing.
3) web server - I like idea of a built-in web server like HS uses. It is very simple to use and it's there without needing to install anything else. I am not a big fan of IIS. It has security problems (and plenty of malicious people looking for them), it only runs on/comes with certain versions of Windows, and installing it can cause funny things to happen to your machine. Downloading and installing something like Apache is beyond many (most?) potential users. While I am technically capable of handling it, I just don't have the time. I am able to do everything I need and most of what I want using the HS server and html/asp pages.
4) communications - Communications between processes, machines, and devices are critical. I think a well thought out and easy to use set of functions for TCP and UDP communications are critical, should be built into the core, and should be accessable to scripting. I think UDP is important because even though it's not as reliable, it is easy to implement in remote devices and many off-the-shelf interfaces use it.
I'm sure as I think about it more, I'll come up with more stuff.
As you can see, though, I've really emphasized scripting. I think this is vital to any open project like this. Not only does it allow end users to quickly implement things that the original developers or plugin authors don't think of, but it also encourages an open, friendly environment were people help each other. Anyone who has used HS for a while and been involved with the forum knows the way things used to be before the advent of for-pay plugins. I'm not saying that the plugins are directly responsible for the changes, but they did come about at the same time.
Certainly, with compiled but open-source plugins, you can have a similar community. But, it will probably be much smaller than one that can also develop based on scripts.