feat: add an enhancement for a Konveyor CLI

[HarikrishnanBalagopal]

Signed-off-by: Harikrishnan Balagopal harikrishmenon@gmail.com

[ashokponkumar]

@jwmatthews @rromannissen Here is the enhancement we had talked about. Let us know if more info needs to be put in.

[jwmatthews]

Thank you for capturing your thoughts in this enhancement, this greatly helps the community to understand and discuss.

I ask that we hold off on moving this to implementation/review, and wait until we have learned more with the Unified Experience and can properly assess the needs we want from a CLI.

The reason I ask to hold on this is because I feel this approach is not yet aligned to the vision of the Unified Experience. I think the technical elements of this can be iterated and we can find alignment in time, but I don't think now is the time to focus on the CLI.

I am not convinced that we want to introduce a CLI focused on unifying all the projects in Konveyor community, I think that is unrelated to the vision of creating a Unified Experience for Konveyor....though I understand this is not obvious, I will explain my thoughts below.

My reasoning for suggesting the delay is I think our primary next step in Konveyor is to establish a Unified Experience for the Konveyor project as a whole. I see the strategy for the Unified Experience being to create a user workflow centered on the Application Inventory delivered by the Hub. We want the Hub to be the central entity. We are not looking to form a collection of disparate tools and apply a light level of discovery/install/porcelain on top..we are looking to create a more unified workflow addressing the workflow of konveyor/methodology.

I see the Unified Experience being focused on:

1. Surface information about applications to a Decision Maker (static code analysis & assessments), storing info in the Application Inventory
2. Aid a Decision Maker to make informed decisions about their modernization/migration needs
3. Help a Decision Maker plan work via Migration Waves
4. Begin to facilitate the 'work' of modernizing/migration

I see 2 personas being involved for the Unified Experience

• Decision Maker (Architect): They will ultimately review information Konveyor surfaces and make informed decisions of what work to plan. Additionally, they will seed projects for 'Developers' to work to handle identified issues.
• Developer: A developer will handle the source code changes to satisfy issues identified and allow the application to take advantage of the given target technology

I see the CLI enhancement as proposed would most likely address the 'Developer' persona for 'Step 4'. i.e. it would help a developer doing the work to discovery tools, install, and use for aiding the work.

I see potential here to consider, but I don't think now is the time to focus on improving the developer persona, as I think the first priority is on the 'Decision Maker' persona. For the 'Decision Maker' in 'Step 4', they will want to execute tools in a prescribed manner to seed many projects for many developers. They are less concerned with a CLI tool for discovery and have a greater need for a non-interactive manner to execute tools as scale, likely via integration as an addon

[shawn-hurley]

Besides the points that John brings up, we would need to have a much better understanding of:

1. What a plugin is and how does it interact with the hub?
2. What is the released artifact, and how do we release them from a community perspective? How does X plugin get in and Y does not? How or who is hosting the repo for artifacts?
3. How does one upgrade? is there some dependency model that we need to build?

I do think that some of these can be solved after getting some initial thing done, but I think we have too many unanswered questions.

[shawn-hurley]

I don't know how this plugin is built/released and/or what the API contract is between the plugins(do we mean Addons?) and the hub is. I believe that we need to solve these sets of problems before writing porcelain commands.

The biggest concern would be designing the API's and interaction models based on a CLI that exists before we have the underlying pieces solved.

[HarikrishnanBalagopal]

Thanks for the feedback, I understand that this is a low priority right now while the underlying pieces are being solved.

I believe that we need to solve these sets of problems before writing porcelain commands.

Not entirely sure what you mean by this, can you please clarify the exact change that is required?
The enhancement simply mentions the vague goals of the commands that need to be implemented.

[shawn-hurley]

I’m sorry, that was a bad choice of words.So what I mean by porcelain is a nice UX around the plumbing we need to build. Hope that helps?