Ancestors

Toot

Written by ⊕-2922 [Rox] on 2025-01-08 at 22:21

This one is going to do its own favourite number poll

It's going to ask for a IEEE 754 half-precision binary floating-point format: binary16 bit by bit, one poll for each bit sequentially

However, in what order should the bits be requested?

=> More informations about this toot | More toots from rox@catcatnya.com

Descendants

Written by ⊕-2922 [Rox] on 2025-01-09 at 07:41

Let's make a IEEE 754 half-precision binary floating-point format: binary16

Round 0

We're starting with the sign

Will the result be positive or negative?

=> More informations about this toot | More toots from rox@catcatnya.com

Written by ⊕-2922 [Rox] on 2025-01-09 at 08:47

Let's make a IEEE 754 half-precision binary floating-point format: binary16

Round 1

We have reached the exponent (0/4)

=> More informations about this toot | More toots from rox@catcatnya.com

Written by ⊕-2922 [Rox] on 2025-01-09 at 09:47

Let's make a IEEE 754 half-precision binary floating-point format: binary16

Round 2

So far: 01?_____________

We're at the exponent (1/4)

=> More informations about this toot | More toots from rox@catcatnya.com

Written by ⊕-2922 [Rox] on 2025-01-09 at 10:50

Let's make a IEEE 754 half-precision binary floating-point format: binary16

Round 3

So far: 011?____________

We're at the exponent (2/4)

=> More informations about this toot | More toots from rox@catcatnya.com

Written by ⊕-2922 [Rox] on 2025-01-09 at 11:50

Let's make a IEEE 754 half-precision binary floating-point format: binary16

Round 4

So far: 0111?___________

We're at the exponent (3/4)

=> More informations about this toot | More toots from rox@catcatnya.com

Written by ⊕-2922 [Rox] on 2025-01-09 at 12:50

Let's make a IEEE 754 half-precision binary floating-point format: binary16

Round 5

So far: 01111?__________

Last digit of the exponent (4/4)

=> More informations about this toot | More toots from rox@catcatnya.com

Written by ⊕-2922 [Rox] on 2025-01-09 at 13:54

Let's make a IEEE 754 half-precision binary floating-point format: binary16

Round 6

So far: 011111?_________

Time for the first digit of the mantissa (0/9) at least, it would be if you didn't vote the exponent to be all ones

We'll have to decide between infinity and NaN, where infinity needs only zeroes, NaN needs anything else

=> More informations about this toot | More toots from rox@catcatnya.com

Written by ⊕-2922 [Rox] on 2025-01-09 at 14:57

Let's make a IEEE 754 half-precision binary floating-point format: binary16

Round 7

So far: 0111110?________

Time for the second digit of the mantissa (2/9)

In the last poll it has not been stated, but the last bit being 0 means it could be a signal NaN

Any further bits of NaN have not been defined in IEEE 754, but if there's a definition for some, tell about it in the replies

=> More informations about this toot | More toots from rox@catcatnya.com

Written by ⊕-2922 [Rox] on 2025-01-09 at 15:07

The last post contains (at least) two typos:

Edit2: it seems to be moshidon

=> More informations about this toot | More toots from rox@catcatnya.com

Written by ⊕-2922 [Rox] on 2025-01-09 at 16:01

Let's make a IEEE 754 half-precision binary floating-point format: binary16

Round 8

So far: 01111101???_____

Let's continue with the mantissa (2-4/9)

Now that it's clear that the answer is NaN, things can be sped up, three bits this time

=> More informations about this toot | More toots from rox@catcatnya.com

Written by ⊕-2922 [Rox] on 2025-01-09 at 17:22

Let's make a IEEE 754 half-precision binary floating-point format: binary16

Round 9

So far: 01111101111???__

Let's continue with the mantissa (5-7/9)

As its clear that the answer is NaN, things can be sped up, three bits this time

=> More informations about this toot | More toots from rox@catcatnya.com

Written by ⊕-2922 [Rox] on 2025-01-09 at 18:25

Let's make a IEEE 754 half-precision binary floating-point format: binary16

Round 10

So far: 011111011111111??

Let's continue with the mantissa (8-9/9)

The final bits!

=> More informations about this toot | More toots from rox@catcatnya.com

Written by ⊕-2922 [Rox] on 2025-01-09 at 20:06

The result

01111101111111111

This is a signal NaN

This causes this one to throw an Invalid Operation Exception :neobot_sign_error:

Good job all making an invalid number your favourite number!

=> More informations about this toot | More toots from rox@catcatnya.com