Andreas Stuhlmüller

Some of our values can be used as keys as well, e.g. ``` `Optimize(model, {optMethod: 'sgd'});` ``` and ``` `Optimize(model, {optMethod: {sgd: {stepSize: 0.5}}});` ``` Would we allow case-insensitive keys...

This feels pretty radical. Personally, I'd choose consistency over convenience here, and improve error messages if necessary. But if Paul is ok with making everything case-insensitive, I'll be ok with...

I'm not sure the syntax has to change for a simple version of the proposal. You could still have a global pointer to the current state (including samplers), except that...

Any of these would be fine. I'd probably put it inline if it is just a sentence, and in doc/reference.md if it is more.

The infinite loop in the initial issue is a bug, but the later `ValueError: Action resulted in an infinite loop` is expected behavior that will be dealt with by decreasing...

It's not so much that I'm against it, but that I want general solutions instead of patching specific cases. Both the current cycle detection scheme and budgets are general, and...

There are valid cases where a context looks the same (when stringified) as one of its ancestors, but differs based on the values behind pointers (which depend on the workspace...

I don't think it's obviously unimportant. Identifying pointers (for the purpose of unlocking within a workspace) that have the same contents, but were created independently, feels like a symptom of...

Also want notifications for judges eventually (e.g. randomly selecting a judge to notify, and sending more notifications to other judges if the first one doesn't respond).

I agree! I think in practice this doesn't matter right now but good to do.