Question: motivation for by-index list initialization

[fandrei]

For example, in the docs:

var countdown = new TimerRemaining()
{
    buffer =
    {
        [^1] = 0,
        [^2] = 1,
        [^3] = 2,
        [^4] = 3,
        [^5] = 4,
        [^6] = 5,
        [^7] = 6,
        [^8] = 7,
        [^9] = 8,
        [^10] = 9
    }
};

(Let's set aside the fact that this code is awful and should never be used)
However, it's just unrealistic. What are real world examples where this feature is necessary, and why not achieve same results using other language features?

[HaloFour]

https://github.com/dotnet/csharplang/blob/main/meetings/2023/LDM-2023-11-15.md#evaluation-of-implicit-indexers-in-object-initializers

[Digitalio1]

We think the original index-init feature is a good thing. How many more times do you need to hear this? We think it's a good thing that you can type new Dictionary<string, int> { ["some key"] = some_value }.

What for? It's only new syntax that does the same thing as old syntax. It doesn't add any value. It doesn't let you do something that was difficult without it. It doesn't even save time. The only thing that it adds is the new type of bugs.

This also applies to types like ConcurrentDictionary

This makes somewhat more sense, but:

1. ConcurrentDictionary is used rarely
2. It's very easy to achieve the same result using features that already existed. No need to create the new language feature just for this rare use case.

[CyrusNajmabadi]

What for? It's only new syntax that does the same thing as old syntax.

It doesn't do the same thing as old syntax. You can see that in the spec.

[CyrusNajmabadi]

this rare use case.

Initializing indexable types using their indexers is not a rare case.

Regardless, we simply do not agree with you. You do not find this feature compelling. That's your right. We do. So we added it.

The requests for it were something we found reasonable. The use cases were compelling enough. And the costs were acceptable.

That's all that matters here. Universal acceptance of a feature is a non goal for us. You not liking it or thinking it is worthwhile literally doesn't impact the decision making. Especially the decision making from literally 2015.

No need to create the new language feature just for

Let me be as blunt as possible, as this discussion does not seem to be adding much value. This is not your call to make. If you would not take this feature for your own language that's your call. But this is our language, and we take in the data, requests, use cases, etc and we make determinations in what we think is worth investing in. 100% of the features we do have people who do not think it is worthwhile. That simply is not relevant to us. We do the features because we believe their value justifies their cost. You're welcome to disagree. But we're not relitigating decisions we made 10 years ago that we still think we're the right call.

[Digitalio1]

The use cases were compelling enough.

I've yet to see a single one that actually offers any improvement. I'm not even talking about necessity.
You've spent an hour beating around the bush, and it looks like "we found it useful, because reasons" is the only answer that you have.

[CyrusNajmabadi]

I've yet to see a single one that actually offers any improvement

So be it. It's not our job to provide you with cases you find compelling. As mentioned, with every release there are people like yourself that feel this way about every feature we do. It's unclear though then why you think this would be a concern for us. We have no goal around providing suitable examples that convince all users that a feature is worthwhile. It's a non goal. You keep expecting us to do something which we see no need or value in doing.

" "we found it useful" ... is the only answer that you have.

It's the only answer that we need. That's what language design is. We make determinations on what we think will be useful for the ecosystem and we design and implement that.

[CyrusNajmabadi]

You might prefer coming to our discord to just chat directly about your questions. You can directly interact live/async at: https://discord.com/invite/tGJvv88

[fandrei]

You might prefer coming to our discord to just chat directly about your questions. You can directly interact live/async at: https://discord.com/invite/tGJvv88

In this case, this info would be lost for the people who will have same question in the future.

[CyrusNajmabadi]

No it wouldn't. Discord is searchable and indexed and people find and comment on threads from well in teh past plenty of times. :)

It's much more conducive to this form of rapid-fire back and forth that you're doing here :)

[CyrusNajmabadi]

At this point i'm closing and locking this discussion. There appears to be no interest in engaging with this topic. The discussion is going in circles, with none of the information provided being accepted.