Launch Feedback for the NEW MeltanoHub! 🚀

[aaronsteers]

We're excited to announce the launch of the new Meltano Hub!

https://hub.meltano.com

This discussion is to track feedback and ideas for improvements.

For any large issues, you may also use the Issue Tracker.

[aaronsteers]

Minor UX nit: it's not obvious that this has a scrollbar and the 'extractor' entry is below the visible border:

As a cue to the user, can we increase the base height so it shows a fraction of the next item - and/or add a visible scroll bar or indicator? Also might be worth opting towards 4 visible items if space allows.

[tayloramurphy]

@aaronsteers thoughts on making the search box bigger height wise?

Also, what if we cut the size of the search boxes by 2/3 and left justify them? All that white space feels weird.

[aaronsteers]

Agreed on both. 👍

[visch]

Which variant has the most usage? Seeing from the tap screen would be useful!

[visch]

Something simple like a "In the last week/month" this variant has 30% of the runs, this on 70% of the runs. As I type this then I wonder if percent of runs is what you want, maybe it's percent of unique project runs?

[aphethean1]

Suggest a small trend chart - plotting usage over time for each variant.

[aaronsteers]

Message received. 👍

The trend chart is probably a ways away but we can probable add more comparison-friendly details to the "Other Available Variants" section.

[visch]

Search for "click" , the results that come up are ordered alphabetically, not by search "strength". If you're trying to find clickup in the list it's a bit odd, not a huge one just slightly annoying.

[visch]

Other variant runs are not listed. For tap-clickup there's a number of folks that run meltano with their own forks (I think). It'd be nice to have a "other variant" list with the number of runs/projects, but maybe name matching isn't great? On the hub right now there's 11 projects that run clickup, on the new site it goes down to 6

[aaronsteers]

Hmmm... Fascinating. I hadn't thought through this before you called it out. I think we naively thought that if we have 3 variants, then we would get:

variant 1 usage + variant 2 usage + variant 3 usage = total usage

Now on thinking about this further, the math should probably be:

variant 1 usage + variant 2 usage + variant 3 usage + other forks usage = total usage

The challenge of registering the other forks usage is that we may not have data on which of the variants the "other forks" are derived from.

Those "other forks" usage would be relevant in terms of understanding interest but not usage, stability, or popularity of a specific implementation.

Curious what you think, @visch ...

[aaronsteers]

One more thought on this:

For privacy reasons, we don't report a clear-text pip_url back via telemetry, but we do report back a one way hash. So, if users are running under a public GitHub fork, and if the users are installing with a pip_url like git+https://github/<myorg>/<tap-something>.git, then we should be able to pair up the hashed pip_url against the possible hashed-fork-name values to identify specific variants which may be getting a lot of usage - and perhaps therefor would be worth reaching out the maintainer to see if they want their project to get promoted for other community members on the Hub. Even without formally hosting them on the hub, we might include them somewhere in 'other variants' list.

[aaronsteers]

Might be worth discussing more in a dedicated thread:

• Applying anonymized usage data to help the Singer and Meltano communities #827

[visch]

Yeah I'm not sure, I know one of the main use cases of different forks is security and stability so the maintainer can't just change things on you and next thing you know you're broken. I'm not sure how to deal with that, the "other" seems like a good way (and yes counting unique pip_url hashs would be interesting too with like a min,max or something)

[edgarrmondragon]

Would love to navigate the site with the keyboard 😄

[aaronsteers]

Message received. 👍

Tracking in:

• feat: support arrow keys in selecting search results #799
• Press 's' to start typing search #637

[aaronsteers]

@rwfeather has a pending fix landing soon in #871

🚀

@visch - thanks for raising.

[rwfeather]

Merged :)

Thanks @visch

[visch]

nice work!

[visch]

Search on mobile (Android pixel 6) doesn't populate until you hit enter

[aaronsteers]

Confirmed on my OnePlus Android.

Seems that hitting the space key (at least for me) also triggers the search action but it otherwise does not trigger while typing.

[aaronsteers]

Logged as issue:

• Mobile search experience doesn't render while typing #863