Hello,
thanks Norman for launching this enjoyable discussion (and keeping up the tradition). Reading the postings helped much to get in proper reflective mood to wrap up the past months and organize my ideas for 2019.
On Sun, Dec 23, 2018 at 08:46:48PM +0100, Norman Feske wrote:
The Year of Sculpt
Looking back at the past year I'm very proud of our achievements. Not only that our team at Genode Labs conducts its work on Sculpt every day, the OS is also mature enough to be useful for enthusiasts. Also, each step on the Sculpt timeline widened the horizon of what could be achieved with this approach in the future.
As the previous posters, I organize my ideas for 2019 into the suggested categories, because I'm confident they make up a perfect frame for our planning.
- Widening the audience of Sculpt OS
In this category I see two important aspects, which were mentioned in the discussion already.
Documentation
We should keep up the series of articles about Sculpt that we founded in 2018 and not only update the development progress but also document the Sculpt OS as a whole (starting from our talk slides). Ideally, the Sculpt documentation will evolve into an up-to-date book about this specific Genode OS accompanying the Foundations.
Applications
As we can't implement all required applications natively using the Genode API by ourselves, the porting effort of existing applications must be reduced. The first take of the Genode SDK was released in November and we should intensify our efforts in this regard. I see the SDK primarily as environment for POSIX applications that integrate seamlessly into the system by means of the libc and the VFS. Additional SDK modules could be added on top of the base SDK, for example, a Genode Qt SDK, which integrates with the Qt Creator/qmake/cmake easily.
Christian Prochaska also mentioned a tool-chain update with additional effort for better autotool support. Maybe we should merge the tool chain into the SDK? Also, we may consider musl as (Linux-compatible) C runtime for relevant platforms, namely x86-32/64, ARMv7/8, and RISC-V. The Linux compatibility of the libc may help to enable more applications which lack universal UNIX support or suffer from poor FreeBSD adaptation.
Further, I'd like to support the proposition that the enablement of more language runtimes attracts a broader diversity of developers.
- Fostering the community spirit around our project
I'll definitely have my part on genodians.org. Therefore, I plan to document progress on the network-appliance projects mentioned below and other stuff about Genode. Furthermore, I wonder if Genode could become more visible on open source gatherings beyond FOSDEM or even conferences/exhibitions and are all ears for suggestions how to achieve this.
- Marketing of Genode-based products
Even if "products" is a rather heavy term, I plan to push Genode-based deployments to address tasks in our networking environment at Genode Labs. I see the Sculpt runtime as a key to easily develop appliances for a broad range of applications like Sebastian's build server, secured NAS, IPv4 router/firewall, or even John Karcher's always-on mail/file/audio-streaming server. This goal involves many improvements and adaptions of components (e.g., NIC router) and Sculpt (e.g., support for head-less scenarios that even lack a framebuffer).
I'm looking forward to an exciting Year 2019!
Regards