[92779] in North American Network Operators' Group
Re: that 4byte ASN you were considering...
daemon@ATHENA.MIT.EDU (Ian Mason)
Wed Oct 11 09:41:37 2006
In-Reply-To: <200610102154.k9ALs4G00710@wraith.qbfox.com>
Cc: nanog@nanog.org
From: Ian Mason <nanog@ian.co.uk>
Date: Wed, 11 Oct 2006 14:40:45 +0100
To: Per Gregers Bilse <bilse@networksignature.com>
Errors-To: owner-nanog@merit.edu
On 10 Oct 2006, at 22:54, Per Gregers Bilse wrote:
>
> [This isn't meant to be flippant or anything else of the kind, it's
> a genuinely heartfelt thing, albeit maybe a bit off topic.]
>
> What all things computer related has needed from day one is a way
> of pronouncing ("reading out loud") hexadecimal. My first computer
> was a 6502, and I've resented numbers larger than FF since then
> (been working with AMD Opterons for a couple of years now, =20
> disturbing).
>
> If you print and read in hex, you don't need dots or any other =20
> syntactic
> aids, the human eye/brain can easily group the requisite number of =20
> digits,
> at least for the time being.
>
> The problem is that from and including A we can't talk about the
> damned things any more -- we resort to spelling out each number, with
> no inherent and natural feel for what we're taling about.
>
> An A380 has a maximum take-off weight of around 24E (two-four-E) =20
> tonnes.
> An A380 has a maximum take-off weight of around 590 (five hundred =20
> and ninety)
> tonnes.
>
> Solve that, and we don't need any new notations beyond subtle =20
> groupings,
> just like we group thousands and millions in decimal notation.
>
> - Per
This is so, so off topic it's not true. I started this as an off-list =20=
reply
to Per but I'm so pleased with my solution that I can't help sharing it.
Take the solution from natural languages. Most languages I speak (or =20
have
a smattering of) have a regular or semi-regular way of pronouncing =20
numbers.
Single digit numbers have a unique name.
10 (the base) has a unique name.
Numbers from 11 to 19 have a name with a suffix and a sound similar =20
to the terminating
digit usually with a break from the rule for 11 and 12. (nine, =20
nineteen) (f=FCnf, f=FCnfzehn)
We'd regularize that and not have
Two digit numbers with a zero in the lowest position have a name =20
using, again, suffix and a similar
sound to the name of the single significant digit involved. (four, =20
fourty) (vier, vierzig)
100 has a unique name. 1000 has a unique name. Multiples of either =20
are said
<digit name> <multiplier name>.
That's enough rules apart from the rules for combining all the above =20
rules.
So, we just need:-
1) Unique names for all the single digit numbers.
2) A unique name for the base.
3) A suffix sound for 1x form numbers.
4) A suffix sound for x0 form numbers.
5) As many unique names for x00000... form numbers as we feel we =
need.
6) A combining rule(s).
So:
1) Use the english names for 0..9. A..F may need new names if
combined versions sound too similar to the compound forms.
2) 0x10 =3D hen
3) Use the suffix -heen for 0x11 .. 0x1f
4) Use the suffix -he for 0xX0
5) 0x100 =3D hexdred, 0x1000 =3D hexdrend
6) use the english combining rules
7) Try lots of combinations and then revisit 1. e.g
0xA0 becomes 'Aye'-he which sounds too much like eighty for
comfort; so A may need a new name.
So:
0x5432 =3D five hexdrend, four hexdred and thirhe two.
0x1017 =3D one hexdrend and sevenheen
0x10000 =3D hen hexdrend
Happy counting,
Ian
