More actively discourage users from using Shuffleboard and SmartDashboard

[jasondaming]

With the amount of good dashboard options these days it really doesn't make sense to present these as viable options. Unless you insist on code based generation and are okay dealing with all of shuffleboards problems, I really can't make a logical case for either of them.

I think there should be more information here and on the first page for both Shuffleboard and SmartDashboard

We should also link to the AS section here https://github.com/wpilibsuite/frc-docs/pull/2418

[calcmogul]

I agree with this. Same reasoning goes for PathWeaver vs PathPlanner.

Going a step further, I guess the main requirement for dropping Shuffleboard, SmartDashboard, and PathWeaver entirely is having stable, long-lived alternatives for what most users need. Supporting a dozen users isn't enough reason to keep those buggy, unmaintained projects around, imo.

[Kevin-OConnor]

AdvantageScope is not intended to be a dashboard replacement. None of these other projects have been around longer than a year or so. Until there is something appropriate to refer people to, dissuading them from working, if flawed, tools that we are still packaging seems inappropriate.

[calcmogul]

That's the thing though. Shuffleboard doesn't work a lot of the time. It's laggy and slow, eats a lot of RAM, and has unusable plots. Glass is better for plots, and any of the existing thirdparty dashboards will serve their other needs better.

PathWeaver is painful to use, and the outputs don't even respect kinematic constraints (so the results can't be followed accurately by robots). Furthermore, PathWeaver only supports differential drives despite how popular swerve is at this point, and the diff drive heading constraint makes no sense for swerve. Much better thirdparty tools exist and have existed for a while, like PathPlanner. I would guess most teams use PathPlanner at this point anyway if they need to create trajectories, not PathWeaver (just speculating because we have no tool usage data).

In general, our users deserve better than what our official tools provide, and there are thirdparty tools we can direct them to.

[jasondaming]

AdvantageScope is not intended to be a dashboard replacement.

In my experience even though AdvantageScope isn't targeted as a in match dashboard it still works better than Shuffleboard in that respect. I agree it doesn't cover all that shuffleboard does, but do most users need custom plugins (completely unused) or many of the other features that aren't present.

I am not saying that we shut shuffleboard down only that we truly open the users eyes to all of the possibilities. There are constantly teams I meet that come and complain about shuffleboard (both in person and on places like discord) knowing nothing about any of the alternatives (and how shuffleboard is not being actively fixed). I like to think (and many teams do too) that when they see a program from WPILib they instantly assume it is the best available. In order for us to keep that reputation and respect it is critical that we are upfront about both our and 3rd party applications capabilities and if there is really any reason they should consider an alternative to WPILib we should help them make that informed choice.

[jwbonner]

In my experience even though AdvantageScope isn't targeted as a in match dashboard it still works better than Shuffleboard in that respect.

I think this really says more about the lack of mature alternatives, rather than the capabilities of AdvantageScope as a driver dashboard. The biggest reason that AdvantageScope isn't a good dashboard is the UI, not necessarily plugins or other niche features of Shuffleboard. Fundamentally, a tab-based layout like AdvantageScope is just severely limited for the primary use case of a dashboard; looking at data on a single screen in a way that's very easy to read during a match. There are also no capabilities to write data, like for selecting autos.

If we're trying to reduce confusion, I think the last thing we need is to recommend teams use AdvantageScope for something it was explicitly not designed to handle. I suspect it will only create more confusion about why it doesn't include the "basic" capabilities of a dashboard (e.g. displaying multiple widgets at once).

[jasondaming]

I agree completely. I also wouldn't recommend AS as a driver dashboard. I think it functions really well in a lot of other dashboard ways like glass. We put glass on the "dashboards" page and don't recommend it as a "driver dashboard".

Our team had great success with FRC-web-components. I understand Kevin's trepidation about solutions that haven't been around long and I worry about that too. I think I should put a PR together for this and hopefully the language I use would be fair and put everyone's fears in this thread to rest, but at least that will give us something concrete to talk about.