I don't think I'd find it frustrating. I think I would enjoy the puzzle part of it.

But indirectly I see that difficulty in placing tiles impacting my enjoyment: because it means that there are no/few meaningful decisions to be had in terms of where to place your tile (because there's often only 1 place you can put it, and it may sometimes benefit your opponent more than yourself) or which tile to place (because you don't get any choice -- unless you can't play the first one, and then you can play a previously unplayable one or draw blind).

Nothing wrong with interactive novels being sold in the same store as games... as long as it's clear what it is (no false advertising).

Somewhat agree with some of the other points...

But if you intend to save to a model after the form validates, then you can't escape the models' validations:

either you check that the models pass their own validations ahead of time (like I want to do, and I think @mattheworiordan was wanting to do), or you have to accept that one of the following outcomes is possible/inevitable if the models' own validations fail:

1. if you use object.save then it may silently fail to save
2. if you use object.save then it will fail to save and raise an error

Are either of those outcomes acceptable to you? To me, they seem not to be. Hence we must also check for / handle the models' validations. Hence we need a way to aggregate errors from both the form object (context-specific validations) and from the models (unconditional/invariant validations that should always be checked by the model), and present them to the user.

What do you guys find to be the best way to accomplish that?

I am interested to know what best practices you use / still use today after all these years. I keep finding myself running into this same problem/need, which is how I ended up looking for what the current options are for form objects today...

Wrong direction: I'd recommend migrate from Angular to React.

Does "evil" refer to religion? Or perhaps they meant "evil" in a more general way, as a more extreme version of "bad".

I think it actually looks great/better because it intersects descenders of the characters.

Eh? I didn't know links were considered buttons. I'm not sure I totally agree understand, but it's not outrageous either...

Update: Okay, I guess when you put an outline around it (like they directly below this paragraph), and even more if you put an icon with it (like they did further down; https://hyp.is/DZTZzi6fEeuu65uvQJ9W1Q/uxdesign.cc/ui-cheat-sheets-buttons-7329ed9d6112), the link looks like more like a button.

But (and I think this is their point) it is what it is because of how it's used and not how it's styled: it should be the same thing (a button) whether or not it has an outline.

Why on earth not? There's nothing wrong with reusing the same answer if it can work for multiple questions. That's called being efficient. It would be stupid to write a new answer from scratch when you already have one that can work very well and fits the question very well.

I'm not quite sure I understand this (so until I do, I'm not sure I agree)...

Why does he think copying and pasting is the only way to make it readable? Like he pointed out, you can extract JSX snippets and use loops within JSX. But maybe he means (his previous point), that people often don't do that. Hmm. 

Not really. How?

... which have good names.

This implies that those names where the inverse_of cannot automatically be inferred are bad names. I disagree that a "good name" is at all related/dependent on that ability.

What they should say here instead is:

... which have names that allow the relationship to be easily inferred.

Or refer to these names as the "default" or "Rails conventional" names for these associations.

But it is not necessarily a better name. A better name is, quite often, one that is more descriptive and specific.

For example, just because by default if you use rails generate with a User model, it might (I don't remember; can it even generate associations?) create a belongs_to :user association doesn't mean that's the best name for it. belongs_to :author or belongs_to :owner, for example, being more specific, are likely better names. The model still needs a generic name like User because it may be used in various relationships, but the relationships themselves should pretty much never be called user because there's almost always a more specific name that better reveals/describes the relationship.

This is better than creating a duplicate new thread.

There is no better place to respond to an existing topic than in the existing thread for that topic.

What's so bad about this solution? If it works, it works. And it only requires wrapping in 1 additional block. Pretty simple.

I'm not sure I agree with this:

[the technical cookies] are explicitly requested by the user when they indicate that they would like to add the item to the cart.

The only thing they requested was that the item be held in a cart for them. They didn't explicitly request that cookies be used to store information about items in the cart. They most likely don't understand all of the options for how to store data like this, and certainly wouldn't know or expect specifically that cookies be used for this.

In fact, localStorage could be used instead. If it's a single-page app, then even that would be necessary; it could all be kept in page-local variables until they checked out (all on the same page); such that reloading the page would cause the cart data held in those variables to be lost.

It's not an abuse of power. It's free speech. It's protesting against an awful proposed law.

IMHO, re-opening classes is okay. Certainly better than duplicating an entire core Ruby class and giving it a silly, less-meaningful name. (Though I'm not sure he actually intended people to use Lax instead of Lazy. I think he was just showing how easy it is to implement Lazy from scratch in Ruby.)