jsHash v0.3

Non-asynchronous hash function in Javascript.

Details

Takes an object of any size, and returns a pseudorandom 8-digit base62 string.

Usage

Vanilla Javascript

<script src="https://jbwx.github.io/jsHash/jsHash.js"></script>
<script>
      console.log(hash("Hello, world!"));
</script>

Node

const hash = require('./jsHash.js');
console.log(hash("Hello, world!"));

What's new?

Happy to announce some major improvements to the previous iteration!

The output domain has been changed. The output is now 8 characters of base62. A-Z, a-z, and 0-9. 62 characters total. 62^8 = ~218 trillion possibilities.

The output domain is much smaller than the previous, 16-byte hex output. If you notice, the characters are all URL-compatible. I'm trying to tailor this for a practical purpose – an IDing system for an enterprise application.

I've released an Integrity analysis. This has helped me iterate and improve the function a lot!

You can replicate the results below for yourself using this (hashTestSuite.js).

Integrity Analysis

Character Frequency (Total and Positional)

This analysis combines total character frequency with positional frequency.

• Total Frequency: 804,808 characters were analyzed. A good hash function should show all total counts close to the expected value of ~12,980.77.
• Positional Frequency: 100,601 hashes were analyzed. This shows character counts for each of the 8 output positions. The expected count per character at each position is ~1,622.60.

Character	Pos 0	Pos 1	Pos 2	Pos 3	Pos 4	Pos 5	Pos 6	Pos 7	Total
0	1637	1601	1573	1638	1609	1567	1623	1583	12831
1	1614	1577	1638	1615	1620	1623	1673	1610	12970
2	1618	1615	1630	1527	1558	1573	1567	1616	12704
3	1578	1698	1625	1651	1621	1668	1654	1619	13114
4	1544	1623	1697	1595	1637	1656	1650	1668	13070
5	1634	1599	1598	1613	1609	1599	1614	1588	12854
6	1595	1555	1616	1640	1656	1576	1607	1665	12910
7	1594	1592	1620	1650	1637	1584	1576	1604	12857
8	1614	1645	1560	1623	1587	1651	1685	1632	12997
9	1549	1620	1698	1618	1671	1631	1629	1594	13010
A	1644	1673	1678	1657	1684	1597	1639	1623	13195
B	1573	1684	1629	1657	1641	1661	1620	1627	13092
C	1589	1581	1602	1594	1650	1618	1608	1615	12857
D	1660	1603	1567	1602	1615	1610	1559	1544	12760
E	1591	1710	1552	1582	1628	1665	1615	1637	12980
F	1671	1607	1722	1635	1629	1626	1581	1641	13112
G	1683	1612	1659	1578	1569	1658	1587	1619	12965
H	1642	1624	1556	1583	1601	1647	1601	1675	12929
I	1654	1637	1603	1631	1628	1575	1591	1644	12963
J	1645	1591	1704	1563	1509	1709	1631	1548	12900
K	1618	1668	1617	1614	1573	1607	1642	1649	12988
L	1613	1651	1559	1579	1643	1582	1646	1656	12929
M	1646	1630	1578	1662	1683	1717	1580	1636	13132
N	1581	1575	1611	1660	1609	1677	1631	1672	13016
O	1579	1532	1570	1599	1574	1670	1684	1620	12828
P	1655	1620	1662	1657	1615	1663	1626	1654	13152
Q	1589	1627	1636	1734	1638	1604	1645	1575	13048
R	1683	1709	1652	1651	1610	1587	1669	1542	13103
S	1664	1602	1613	1650	1628	1598	1597	1719	13071
T	1643	1574	1645	1609	1630	1664	1657	1584	13006
U	1567	1635	1581	1672	1599	1669	1545	1627	12895
V	1651	1725	1651	1623	1607	1579	1657	1650	13143
W	1689	1711	1581	1620	1644	1612	1624	1625	13106
X	1674	1600	1624	1635	1587	1621	1692	1620	13053
Y	1626	1672	1555	1627	1608	1636	1716	1611	13051
Z	1566	1633	1574	1636	1647	1610	1590	1590	12846
a	1597	1589	1692	1600	1633	1608	1617	1632	12968
b	1618	1563	1679	1579	1669	1568	1611	1611	12898
c	1672	1726	1599	1677	1695	1664	1685	1640	13358
d	1616	1653	1628	1620	1676	1678	1622	1584	13077
e	1625	1638	1652	1610	1635	1619	1679	1650	13108
f	1656	1607	1564	1617	1654	1574	1630	1549	12851
g	1606	1635	1594	1637	1662	1639	1570	1622	12965
h	1581	1549	1660	1646	1557	1593	1569	1650	12805
i	1612	1644	1670	1596	1601	1667	1568	1586	12944
j	1536	1611	1655	1609	1659	1572	1596	1688	12926
k	1693	1574	1625	1617	1628	1625	1642	1645	13049
l	1641	1571	1543	1590	1593	1648	1626	1580	12792
m	1641	1627	1646	1613	1683	1575	1591	1600	12976
n	1633	1584	1637	1695	1561	1685	1626	1686	13107
o	1612	1585	1603	1591	1585	1598	1582	1611	12767
p	1634	1599	1591	1576	1603	1605	1657	1682	12947
q	1594	1648	1582	1647	1635	1667	1618	1551	12942
r	1631	1646	1709	1601	1554	1638	1664	1631	13074
s	1568	1582	1600	1609	1685	1569	1635	1584	12832
t	1631	1538	1625	1599	1604	1622	1577	1669	12865
u	1593	1659	1622	1568	1672	1585	1579	1601	12879
v	1622	1629	1623	1611	1599	1639	1627	1654	13004
w	1619	1635	1665	1633	1637	1592	1594	1657	13032
x	1640	1685	1644	1617	1643	1629	1636	1629	13123
y	1713	1608	1584	1630	1623	1581	1640	1609	12988
z	1644	1605	1673	1733	1601	1571	1649	1618	13094

Common Substring Analysis

This test looks for repeating patterns inside the hashes. The score is calculated as (Substring Length) * (Number of Occurrences). A good hash function should produce very few, low-scoring patterns.

Substring	Count	Score
'VB'	225	450
'RB'	224	448
'Vy'	223	446
'wQ'	223	446
'RZ'	222	444
'dA'	222	444
'1W'	222	444
'tA'	221	442
'xm'	220	440
'Hy'	220	440
'xI'	220	440
'mK'	220	440
'IN'	220	440
'Tj'	220	440
'nF'	219	438
'eL'	218	436
'Bg'	218	436
'NF'	218	436
'Ux'	217	434
'Z6'	217	434

Good that there isn't a string longer than two characters, but suspicious that the two highest strings both end with "B".

Collision probability

Since the output is an 8-character string from a 62-character set, the total number of possible hashes is 62^8, or 218 trillion. This is equivalent to a ~48-bit hash (log₂(62⁸) ≈ 47.6 bits).

Due to the Birthday Problem, you can expect a 50% chance of a collision after hashing only sqrt(62⁸) ≈ 14.8 million items.

Out of curiosity, I decided to actually test 15 million hashes. The results were quite promising – I did not find a collision.

---

As mentioned, this shouldn't be used for any cryptographic purpose. A full, professional cryptanalysis would likely reveal some vulnerabilities – so the previous disclaimer still applies.

This is not proven in any way to be mathematically sound. There is no guarantee that all outputs are equally likely, or that every output is possible. The output appears to be random from the limited tests I’ve done, although it’s unlikely there isn’t some sort of underlying pattern. Do not use this in any security-critical application.