On 06/24/2012 04:43 PM, Shu Ming wrote:
On 2012-6-23 1:25, Dave Neary wrote:
Hi all,
With a number of colleagues and members of the oVirt community, we have been thinking about how we can help oVirt's adoption and increase developer participation and diversity in the project.
What we've come up with is a broad definition of the target audience for the project, and a three-phase set of initiatives we believe will help the project be more successful.
I'm bringing this to the board list because I would really like your feedback and assistance in helping make this a reality - it's important for me to know that we're proposing things which the project & community think are important.
The sample user we discussed (based on feedback from Users and the persona I sent previously was a cost-sensitive sysadmin, potentially looking to move to a KVM-based virtualisation solution, working on a small virtualisation set-up in a test lab environment. He loves open source because it means he can test and deploy it without approval, adoption is bottom-up, and he can tweak it to suit his needs.
We want to focus on people who want to download and try it in ~30 minutes - running an engine and a node on the same machine (potentially a laptop or desktop machine) and on getting people started with small set-ups - perhaps two nodes and a laptop running the engine.
From my experience, it is non-trivial and time-consuming to setup an working engine with physical nodes attached to it. In order to shorten the time for people, I would suggest we can release some images for engine, virtual nodes, self-configuring scripts in one package. The people who want to try oVirt only need to run a shell command. Then the command will deploy the images into right places and start the engine and nodes in different virtual machines. A well known IP is assigned to the IP for them to access the engine, to create VM in the virtual nodes, &etc.
that's the purpose of the all-in-one module. have you tried it?
Given that target audience and positioning, the three phases we propose are:
Phase 1: Web presence
We want to review the website and wiki to ensure that the content is appropriate for our target audience, that we're making it easy to adopt oVirt. This will cover the top level navigation, the organisation of information in the website and wiki to make it easy to find, development of new content to focus on the beginner experience (including videos, tutorials, and screenshots). At the end of this, when someone from our core audience comes to the website it should be immediately clear to him that we are offering a solution to a problem he has.
We plan to start working on this through the Infra, Users and engine-devel mailing lists (depending on which one is most appropriate) immediately, and complete this work in the next 3 months.
Phase 2: Adoption
We need packages for all the main Linux distributions. Packaging oVirt for OSes other than Fedora is tricky because of various differences in config file placement, low-level tools and so on. But this must get done, and we'd like for it to be a priority for the project.
We also want to help the project with its promotion - both in terms of content on the site, and driving traffic here through conference outreach, articles in tech press, social media and blogs pringing people here via the nice to consume content (videos, tutorials, etc) which we would like to see produced.
In addition, we would like to ask for your help in running a set of oVirt meet-ups around the world, with the co-operation of board members, to show oVirt to the people who are interested in it, and prefer a high-touch approach.
This will take longer, but we expect to see some movement on this front in the next 6 to 9 months.
Phase 3: Expanding the target audience
While the low-end user with small virtualisation will help us drive adoption, we also want to help oVirt appeal to a broader audience. We'd like to help do a UX review of the installation and configuration process, and the Engine UI, to make the oVirt experience more pleasant for both limited-hardware users (people using all-in-one in 3.1) and for more sophisticated users. This is a longer term goal, of cours,e but we would like to see the results of this UX review in a future release of oVirt in ~12 to 15 months.
Looking forward to your feedback!
Thanks, Dave.