Draft a specification? #210
Comments
|
Presumably if we're interested in a stable standards doc, the goal is interoperability with other implementations, of which at the moment there is one, namely aws-event-ruler. So I think that if we were going to do this, we should file an issue over there to alert them and invite them to collaborate. Sound reasonable? While I would be happy to pitch in and help, I don't have the cycles to lead such an effort. |
|
I'd be happy to contribute in whatever form. In the early days here, I had started a crude IETF RFC draft somewhere. I didn't get very far, but I'll see if I can dig it up. |
|
yeah lets collaborate on this. Is https://github.com/timbray/quamina/blob/main/PATTERNS.md a good place to converge on for now? |
|
If by "converge on" you mean "start from a copy of and preserve the RFC-like style" then I agree. |
|
I've never written one of these before but would be keen to help. CC @jonessha as well :) |
|
@timbray, what do you suggest for the drafting format these days? I'm not current at all, but drafting in Markdown might work well enough for us here. My weakly held preference is to start with whatever real markup, boilerplate, etc. from the outset (with content from I'm happy to start a branch (or whatever) with a crude first draft. |
|
Standards-wise, these days I live in the IETF. There, more or less all RFCs are created in either Markdown or XML. The tooling is pretty good either way. Have a look at the setup in https://github.com/ietf-wg-jsonpath/iregexp, if you clone that you get a Makefile and everything. If we have an explicit goal to take this to a standards organization we should pick one sooner rather than later because they all have strong opinions about what input drafts should look like. My instinct would be to do it as an IETF internet-draft, because of that tooling, and anyone can post an Internet-draft for free, which gives at least a permanent URL. Hey AWS guys, is there any relationship with CNCF these days? I sort of did that for a while but ran out of patience with them. Oh, @jsmorph,d to answer your question: Yeah, markdown is fine. |
|
Yep, new repo feels appropriate to me, wonder what the AWS folks think.
…On Thu, Mar 23, 2023 at 4:11 PM Jamie Stephens ***@***.***> wrote:
@timbray <https://github.com/timbray> New repo for the internet-draft
since we have two implementations?
—
Reply to this email directly, view it on GitHub
<#210 (comment)>,
or unsubscribe
<https://github.com/notifications/unsubscribe-auth/AAAEJE7MJOOPLUGWJW7X3QLW5TKCFANCNFSM6AAAAAAWDWHSKQ>
.
You are receiving this because you were mentioned.Message ID:
***@***.***>
|
|
New repo makes sense. For full disclosure, so far I'm only speaking on my own behalf. I need to check if there's any due-process that needs to be followed before I work on behalf of AWS. From past experience, these things can take time so I wouldn't recommend waiting on it. I'd be keen to get started. Happy to create a repo w all permissions & scaffolding based on https://github.com/ietf-wg-jsonpath/iregexp over the weekend. |
|
on
Sorry didn't see this before. I'm not sure at the moment. Let me ask in our internal channels. |
I'm still a bit involved in the CloudEvents side (mostly for the Go and PowerShell SDKs). In general, we do have several streams with the CNCF. Which specific project is this question about? |
|
I've created this shell repo temporarily https://github.com/baldawar/draft-ietf-patterns-protocol/ based on https://github.com/ietf-wg-jsonpath/iregexp/ & https://github.com/martinthomson/i-d-template (neat tool, btw!) Everyone who I could think be interested should be invited to the repo (cc @longzhang-lz for context on the mysterious invite). I will move this to an organization as soon as someone can suggest a decent name. I could only come up w patterns / matchers / ruler patterns which are all felt too generic. Right now I'm heads down on few work-related gremlins, but I can follow up within AWS on if there's any prescribed guidelines the company requires. I also have a bit of backlog of clean up work around event-ruler that I'm desperate to wrap up first, so expecting to take a lot more active role. Until then (targeting Apr End) happy to take care of any minor tasks, administrative duties, and reviews. |
|
I'll be starting to push out some PRs on https://github.com/epml-spec (Event Pattern Matching Language) where I've moved the repo. |
I thought I'd create an issue for any on-going discussion:
As we've mentioned here and sort of in #57, it might be good to have some sort of specification for the pattern language. The current documentation is of course an excellent source, but adding something more formal is appealing (to me).
Ideally the specification is flexible enough to support something like feature flags ("
shellstylenot supported" or whatever). Good test cases: Evolution of capabilities here (for example: pre-existsand post-existsspecs) and similarly for Event Ruler.This table from the EventBridge docs is of course helpful:
A potentially interesting twist is having the specification(s) not tied to specific event rendering languages (like JSON or protobuf), but that goal seems like a stretch with questionable net worth at this point.
Thoughts? Any appetite for something like this?
References
The text was updated successfully, but these errors were encountered: