Hi Lonnie,
I agree on the points you made for making Genode more popular. The live system will certainly raise much more awareness of our project than the regular releases.
However, I see popularity as a two-sided sword. Of course, getting more attention increases the chance for more people joining our community and bringing forward the technology together with us, which is ultimately a good thing. Maybe it will even foster the consideration of Genode for products of companies that are unaware of Genode yet.
On the other hand, I am afraid of two things. First, the live system has to be really good and convincing to developers and end users alike. Otherwise, people will give it a try, it may leave a half-baked impression on them, and this impression will stick for a long time. Some people are quick with their judgments, in particular end users. In contrast people who spend the effort to download and compile source code are less likely to draw conclusions too quickly. The latter I would call our target "demographic", which I feel safe to address right now. The live system widens the audience. So we have to address the topics discussed in issue #1000. And we need to pay attention to thorough testing and little details, which is a lot of work in addition to the actual features to present.
The second point that worries me is related to Daniel's email. Once, the live system achieves to get people hooked on Genode, they may have high expectations on how easy it is to customize the live system. If such people download the source code but then hit a wall, this experience will leave a bad aftertaste as well. To see what I mean:
http://www.reddit.com/r/programming/comments/ubzu9/genode_os_framework_1205_...
Such a bad experience can do our project more harm than it helps us, especially when it becomes an universal truth through public postings like the first comment of the thread above. So we either have to clearly communicate that the live system is just an example to show off Genode's capabilities, or we have to make it really easily "hackable". I would prefer the latter. But making Genode easier to use is hard. In particular because "easy to use" for us developers has a different meaning to "easy to use" for, let's say, a GNU/Linux power user. The current "make prepare" and "make run/..." mechanisms are a good start but still far from being easy.
Please don't get me wrong. I don't want to shy away from the live system. But I regard it as a very sensitive topic. I certainly want to make Genode more popular, but we have to make sure to leave a sustainable good impression, not just on our immediate target demographic but on a broader audience. So we should give it the time needed to make us developers feel confident with it.
Best regards Norman