duplicative of https://github.com/PRX/prx-podagent ?

[kookster]

Seems like maybe we should be working together?
https://github.com/PRX/prx-podagent

[jamescridland]

I'd agree. Key differences are that the PRX work is designed to fit into their/your existing software, whereas the JSON that this project uses is designed to be extensible and usable for anyone, whatever their coding preferences. Do let me know - [email protected] - who to invite into the Slack group which should be a great place to discuss how we can work together.

[kookster]

Hey @jamescridland, the list of user agent info is in yaml format, and was not designed in any app or programming language specific way:
https://github.com/PRX/prx-podagent/blob/master/db/agents.yml

We chose this format (the yaml file format, and key names within it) based on other prominent user agent projects for web based identification, esp. BrowserScope's work:
https://github.com/ua-parser/uap-core/blob/master/regexes.yaml

I believe the format in the podagent project is extensible, and will work with whatever coding preference someone has, just as the yaml from BrowserScope has been used by many libraries in many languages (i.e. links to many of those libs: https://github.com/ua-parser).

That said, @cavis tells me he is adding json as a supported format that will be available in addition to the yaml for those that prefer that, as it is easy in the build scripts to output that format from the yaml.

I'll also note, that we released the podagent project under MIT license, which is also helpful for encouraging contribution and adoption, which we have already seen from several other publishers.

I'll write you re: the slack group for further discussion.

[cavis]

Happy to create whatever output formats are useful in prx-podagent. I've got an open PR that will convert to json, as well as building an html example page on every commit to master (with examples and notes). Something like:

I'd also be interested in other ways to taxonomize the data - we landed on the "name / type / os" classification as it was the most complete/useful for our metrics platform. But there are certainly other ways to slice and dice things.

[hellosteadman]

Hey @cavis. I for one would be more comfortable with a decentralised list, as opposed to one that’s part of PRX.

Not a value judgement — and I hope you don’t see it as such — but I feel like it’s easy to bring all parties on board when there isn’t a single identity seeming to “own” it, even if that obviously isn’t anyone’s intention.

[kookster]

I'd be happy to change the name from prx-podagent topodagent if that is objectionable, we just tend to namespace stuff we make.

I think an MIT license is good for opening a project up for sharing and inviting contribution, and that has been our experience with the project.

From my perspective, I don't know anything about opawg, so I don't know how to evaluate the identities or ownership of this repo, and that (in addition to it lacking a license) leaves me unsure about contributing. Perhaps you can tell me more, or point me to where I can learn more?

Perhaps another option is to make this a part of the https://www.syndicated.media/ (https://github.com/syndicated-media) org, as that was established for these kinds of shared tech/standards/infrastructure projects?

FWIW - I had previously corresponded with @jamescridland about the podagent project, and it had seemed like we were going to collaborate on improving it and combining our info, and then this popped up, so I feel like I must have missed something? @jamescridland did something change/evolve in your thinking?

[jamescridland]

@kookster I guess the absence of a response to PRX/prx-podagent#13 gave me a bit of a cause for concern. Mind, there are similar unresponded comments in the opawg project too. The ownership of this project by PRX gives me pause (it's not the name of the repo, it's the fact that it's clearly owned by PRX and has been driven by PRX's own requirements - which is fine, but).

You make a good point about a license. There's no formal setup for OPAWG yet - I think we're kicking the tyres and seeing whether there is anything wanted by the wider podcast community. There seems to be.

I know nothing about syndicated.media, and not heard about it until now. The Slack looks abandoned, though there are some good people in it from further up the list.

I must confess: the YAML format you're using is easier on the eye; PHP lacks a YAML parser out of the box, but it's easy enough to add one.

[jamescridland]

Since we're discussing this both here and in the Slack group, my intention is to ensure we discuss this in one, not two, places; and will close this here.