RFC 1439
The Uniqueness of Unique Identifiers
RFC 1439: The Uniqueness of Unique Identifiers 🦋
Network Working Group 🌸 Author: C. Finseth, University of Minnesota 💅 Date: March 1993 ✨ Category: Informational. This memo is giving info for the Internet community. It is not an Internet standard 🎀 Obsoletes / Updates: The original doesn't list any RFCs that it obsoletes or updates 🧚
Status of This Memo 💜
Okay so, this memo just provides information for the Internet community. It is NOT a spec, so it doesn't define an Internet standard. You can share it as much as you want, distribution is unlimited 🌈
Abstract ✨
This RFC shares some info that could be helpful when you're picking a way to assign unique identifiers to people 🫶
1. The Issue 💖
Okay bestie, so computer systems need a way to identify the people using them. These are usually called "user names" or "account names" 🍓 The identifiers are typically short strings made of letters and numbers. And generally speaking, these identifiers have to be unique 🔥
The uniqueness is usually achieved in one of three ways:
- The identifiers get assigned in a unique way that doesn't use any info about the actual person. Some examples look like:
ax54tv
cs00034This method was super popular with big timesharing systems. It totally achieved uniqueness, but there was no way to guess someone's identifier unless you already knew it some other way 🙈
- The identifiers get assigned in a unique way, where most of the identifier is built from the person's name. Some examples look like:
Craig.A.Finseth-1
Finseth1
caf-1
fins0001- The identifiers are generally NOT assigned in a unique way. The identifier is built from the person's name, and if there's a duplicate, it gets handled in a kind of ad-hoc way. Some examples look like:
Craig.Finseth
cafNow that we have widespread email, an important feature of an identifier system is being able to predict someone's identifier from other info about them. That other info is usually their name 💌
Methods two and three make those predictions possible, especially if you have just one example of how a person's name maps to their identifier. Method two uses some or all of the name and changes it on purpose to make it unique (for example, by adding an integer at the end). Method three uses some or all of the name, and picks a different identifier when there's a duplicate 🦋
For both methods, it's really important to cut down on how often you need those uniqueness adjustments (an integer that isn't 1, or a totally different identifier). How likely an adjustment is depends on the identifier format and how big the organization is 👑
2. Identifier Formats 🎀
There are a bunch of popular identifier formats. This section lists some of them and gives both typical and maximum values for how many possible identifiers each one can make. A "typical" value is the number you're likely to run into in real life. A "maximum" value is the biggest number of possible values you can get, without going totally extreme about it 🧚 All the ranges are written as a number of bits 💜
2.1 Initials ✨
There are three popular formats based on initials: one, two, or three letters. (The number of people with more than three initials is assumed to be small.) Here are the values:
format typical maximum
I 4 5
II 8 10
III 12 15You can also think of these as first, middle, and last initials:
I 4 5
F L 8 10
F M L 12 152.2 Names 🌸
Again, there are three popular formats based on using names: the first name, the last name, and both first and last names together. Here are the values:
format typical maximum
First 8 14
Last 9 13
First Last 17 272.3 Combinations 💅
I've seen these combinations in use ("F" is the first initial, "M" is the middle initial, and "L" is the last initial):
format typical maximum
F Last 13 18
F M Last 17 23
First L 12 19
First M Last 21 322.4 Complete List 🦋
Here are all the possible combinations of nothing, initial, and full name for first, middle, and last. The number of Middle names is assumed to be the same as the number of First names. Here are the values:
format typical maximum
_ _ _ 0 0
_ _ L 4 5
_ _ Last 9 13
_ M _ 4 5
_ M L 5 10
_ M Last 13 18
_ Middle _ 8 14
_ Middle L 12 19
_ Middle Last 17 27
F _ _ 4 5
F _ L 5 10
F _ Last 13 18
F M _ 5 10
F M L 12 15
F M Last 17 23
F Middle _ 12 19
F Middle L 16 24
F Middle Last 21 32
First _ _ 8 14
First _ L 12 19
First _ Last 17 27
First M _ 12 19
First M L 16 24
First M Last 21 32
First Middle _ 16 28
First Middle L 20 33
First Middle Last 26 40And that's the first half of the tea! ☕💖 Section 3 is next, where we get into the math on how likely duplicates actually are, so stay tuned 🌈🧚
Appendix: Perl Program to Compute Probabilities 🧮✨
okay bestie, here's the actual code that did the math 💅 it's a little Perl script that loops over bit sizes from 6 to 31 and calls Compute for each one 🦋
#!/usr/local/bin/perl
for $bits (6..31) {
&Compute($bits);
}
exit(0);
# ------------------------------------------------------------
sub Compute {
$bits = $_[0];
$num = 1 << $bits;
$cnt = $num;
print "bits $bitsnumber $num:0;
for ($prob = 1; $prob > 0.99; ) {
$prob *= $cnt / $num;
$cnt--;
}
print "", $num - $cnt, "$prob0;
for (; $prob > 0.98; ) {
$prob *= $cnt / $num;
$cnt--;
}
print "", $num - $cnt, "$prob0;
for (; $prob > 0.95; ) {
$prob *= $cnt / $num;
$cnt--;
}
print "", $num - $cnt, "$prob0;
for (; $prob > 0.90; ) {
$prob *= $cnt / $num;
$cnt--;
}
print "", $num - $cnt, "$prob0;
for (; $prob > 0.80; ) {
$prob *= $cnt / $num;
$cnt--;
}
print "", $num - $cnt, "$prob0;
print "0;
}Appendix: Perl Program Output 📊🌸
here's what the script spit out 🫶 each block starts with the bit size and the number of possible values (number), and then each line pairs a count with a probability 🍓
bits 6 number 64:
2 0.984375
3 0.95361328125
4 0.90891265869140625
5 0.85210561752319335938
6 0.78553486615419387817
bits 7 number 128:
3 0.9766845703125
3 0.9766845703125
5 0.92398747801780700684
6 0.88789421715773642063
8 0.79999355674331695809
bits 8 number 256:
3 0.988311767578125
4 0.97672998905181884766
6 0.94268989971169503406
8 0.89542306910786462204
12 0.76969425214152431547
bits 9 number 512:
4 0.98832316696643829346
6 0.97102570187075798458
8 0.94652632751096643648
11 0.89748056780293572476
16 0.78916761796439427457
bits 10 number 1024:
6 0.98543241551841020964
7 0.97965839745873206645
11 0.94753115178840541244
16 0.88888866335604777014
22 0.79677613655632184564
bits 11 number 2048:
7 0.98978773152834598203
10 0.97823367137821537476
15 0.94990722378677450166
22 0.89298119682681720288
31 0.79597589885472519455
bits 12 number 4096:
10 0.98906539062491305447
14 0.97800426773009718762
21 0.94994111694430838355
30 0.89901365764115603874
44 0.79312138620093930452
bits 13 number 8192:
14 0.98894703242829806733
19 0.97932692503837115439
30 0.94822407309193512681
43 0.89545741661906652631
61 0.7993625840767998314
bits 14 number 16384:
19 0.98961337517641645434
27 0.97879319536756481668
42 0.94876352395820107155
60 0.89748107890372830209
86 0.79973683158771624591
bits 15 number 32768:
27 0.98934263776790121181
37 0.97987304880641035165
59 0.94909471808051404373
84 0.89899774209805793923
122 0.79809378598190949816
bits 16 number 65536:
37 0.98988724065590050216
52 0.97996496661944154649
83 0.94937874420413270737
118 0.89996948010355670711
172 0.79884228150816105618
bits 17 number 131072:
52 0.98993311138884398925
74 0.97960010416289267088
117 0.94952974978505377823
167 0.89960828942716541956
243 0.79894309171178368167
bits 18 number 262144:
74 0.98974844864797828503
104 0.97977315557223210174
165 0.94968621078621640041
236 0.8995926348279144058
343 0.7994422793765953994
bits 19 number 524288:
104 0.98983557888923057178
147 0.97973841652874515962
233 0.94974719445364064185
333 0.89991342619657743729
485 0.79936749144148444568
bits 20 number 1048576:
146 0.98995567500195758015
207 0.97987072919607220989
329 0.94983990872655321702
471 0.89980857451706741656
685 0.79974215234216872172
bits 21 number 2097152:
206 0.98998177463778547214
292 0.97994400939715686771
465 0.94985589918092261374
666 0.89978055267663470396
968 0.79994886751736571373
bits 22 number 4194304:
291 0.98999013137747737812
413 0.97991951242142538714
657 0.94991674892578203959
941 0.89991652739633254399
1369 0.79989205747440361716
bits 23 number 8388608:
412 0.98995762604049764022
583 0.97997846530691334888
929 0.94991024716640248826
1330 0.89999961063320443877
1936 0.79987028265451087794
bits 24 number 16777216:
582 0.98997307486745211857
824 0.97999203469417239809
1313 0.94995516684099989835
1881 0.89997049960675035152
2737 0.79996700222056416063
bits 25 number 33554432:
822 0.98999408609360783906
1165 0.9799956928177964155
1856 0.9499899669674316538
2660 0.8999664414095410736
3871 0.79992328289672998132
bits 26 number 67108864:
1162 0.98999884535478044345
1648 0.9799801637652703068
2625 0.94997437525354821997
3761 0.89999748465616635773
5474 0.79993922903192515861
bits 27 number 134217728:
1644 0.9899880636014986024
2330 0.97998730103356856969
3712 0.94997727934463771504
5319 0.89998552434244594167
7740 0.79999591580103557309
bits 28 number 268435456:
2324 0.98999458855588851058
3294 0.97999828329325222587
5249 0.94998397932368705554
7522 0.89998576049206902017
10946 0.79999058777500076101
bits 29 number 536870912:
3286 0.98999717306002099626
4659 0.97999160965267329004
7422 0.94999720388831232487
10637 0.89999506567702891591
15480 0.7999860979665908145
bits 30 number 1073741824:
4647 0.98999674474047760775
6588 0.97999531736215383937
10496 0.94999806770951356061
15043 0.89999250738244507275
21891 0.79999995570982085358
bits 31 number 2147483648:
6571 0.98999869761078929109
9316 0.97999801528523688976
14844 0.94999403283519279206
21273 0.89999983631135749285
30959 0.79999272222201334159References 📚🎀
Bruce Lansky (1984). The Best Baby Name Book. Deephaven, MN: Meadowbrook. ISBN 0-671-54463-2.
Lareina Rule (1988). Name Your Baby. Bantam. ISBN 0-553-27145-8.
Security Considerations 🔐
Security issues are not discussed in this memo. 🙈
Author's Address 💌
Craig A. Finseth
Networking Services
Computer and Information Services
University of Minnesota
130 Lind Hall
207 Church St. SE
Minneapolis, MN 55455-0134
EMail: Craig.A.Finseth-1@umn.edu or
fin@unet.umn.edu
Phone: +1 612 624 3375
Fax: +1 612 626 1002And that's the whole RFC, bestie! 🌈👑 Craig did the maths so we didn't have to 💜
📦 where this page comes from
- ietf: rfc1439 1993-03
- 💖 girlified by claude-haiku-5-5 on : a language model's rewrite of the original, under its terms