I disagree with most of what Kirk said and stood for.
But you have to be a special kind of sick person that on this day, when two daughters lost their father, in a gruesome shooting in broad daylight in front of thousands and you have -1,000,000 empathy.
You should feel shame in every cell.

Before I post anything I “code-review” myself to make sure I only address the point in a technical and respectful manner.

Removing anything that can be considered offensive or personal.

Social media doesn’t HAVE to be toxic.

We make it so.

Adults, especially professionals, such as yourself are supposed to act professionally and address issues technical, scientifically and not ad-hominem.

A person who gets upset about a simple disagreement in a respectful inquisitive discussion and resorts to personal attacks is acting like a child

I’m not calling you hostile for disagreeing with me.

On the contrary.

I’m calling you hostile for attacking me personally.

It’s not a nice thing to say, it’s condescending and belittling and there’s no reason for that.

One thing I really hopes twitter would do is a tree graph visualization of a tweet.

Where you can navigate all threads as branches of the tree from the original tweet.

So, I’m being hostile because I don’t just take your word for it?

and instead am just curious and intrigued to dig deeper into what’s happening here?

Let’s take your advice and actually “do better”!

Let’s be professional and respectful.

Put a link to the code up so we can investigate.

I have jumped to exactly 0 conclusions.

I only asked questions and was interested to dig deeper and investigate why such extreme differences in response times.

I kindly asked you to provide the code so I can do exactly that.

No assumptions and conclusions.

Just curiosity.

I am?
How am I being hostile?

That’s why I asked if you could provide the code.

It’s interesting to dig deeper and profile it to see what’s going on.

Also, why such hostility?

I thought that was why we left the other app and came here to avoid 😔

1 millisecond =1,000 microseconds

30 milliseconds =30,000 microseconds

In my given example:
Db query - 30 milliseconds.
Routing - 600 microseconds.

30,000 (30 * 1,000) microseconds of a db query + 600 microseconds of routing = 30,600.

You don’t have to be hostile.

It matters very little how fast your routing is.

If your code is doing any real work - that’s the bottleneck and not your router.

If your endpoints do no I/O and are extremely fast and optimized, the fast routing becomes important.

My point is - routing is negligible with real work.

I think you got my point wrong.

My point was:

* Hono routing takes 60 nanoseconds
* Express routing takes 600 nanoseconds.

If my endpoint does a simple db query that take 30 milliseconds, then:
* Hono takes 30,060 ms
* Express takes 30,600 ms.

Not that big of a difference now.

>>>

How did you come to this conclusion?
What question did I ask?
And what was the answer?

So, if it takes 60 or 600 nanoseconds to find the route, it is still negligible if your endpoint is doing real work that takes time.

If you check the details of the benchmark - they’re all in the nano/micro second ranges. No real test of actual real world scenarios - just a test for how fast the router is (impressive and important still).

No doubt!

I’ve already read that page before.

What makes Hono fast is the router - How quickly can it find the route?

It doesn’t magically make all your code go faster.

The code inside the route handler is the same regardless of the router.

>

Can you post the code?
The increase is orders of magnitude and can’t possibly be related to a simple framework change.

Only of something extreme is happening there.