Add/comment/multipleof

[akshat09867]

Documentation Update: multipleOf Validation Specification

Related to #1113

Summary

Updates multipleOf validation specification to address IEEE 754 floating-point precision issues, providing clearer guidance on implementation.

Documentation Changes

• Added warning about floating-point precision limitations
• Clarified validation requirements
• Added practical implementation guidance
• Updated examples to be more IEEE 754 friendly

[github-actions]

Welcome to the JSON Schema Community. Thanks a lot for creating your first pull request!! 🎉🎉 We are so excited you are here! We hope this is only the first of many! For more details check out README.md file.

[github-actions]

built with Refined Cloudflare Pages Action

⚡ Cloudflare Pages Deployment

Name	Status	Preview	Last Commit
website	✅ Ready (View Log)	Visit Preview	949c3fd

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 100.00%. Comparing base (8676fc2) to head (949c3fd).
Report is 10 commits behind head on main.

Additional details and impacted files

@@            Coverage Diff            @@
##              main     #1219   +/-   ##
=========================================
  Coverage   100.00%   100.00%           
=========================================
  Files           10        10           
  Lines          373       373           
  Branches        94        94           
=========================================
  Hits           373       373           

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

[gregsdennis]

Is this really the recommendation we want to make?

[mwadams]

I'd say "no".

Use of scaled integers is ok but is a choice not a recommendation. And we don't have direct support for a scaled integer - e.g. defining the scale. You push a whole new problem onto consumers.

When it comes to non-integers it isn't really meaningful; why would we be recommending that people only use "particular values". And then there is the problem of representation on all platforms (e.g. 1.3, 2.4, 5.6 cannot be represented precisely as a 32 bit float).

[gregsdennis]

A use case could be defining currency, for example dollars & cents. You'd want to ensure that your data only has a penny ($0.01) resolution.

But the point is that we don't want to force users into a scenario where they should be trying to figure out the nuances of binary float arithmetic while writing their schemas.

I think that the very most, we should be leaving a note that some implementations may use IEEE754 math.

I do think the way that the spec is worded is pretty clever, though.

6.2.1. multipleOf

The value of "multipleOf" MUST be a number, strictly greater than 0.
A numeric instance is valid only if division by this keyword's value results in an integer.

In C#,

double a = .0075;  // IEEE754
double b = .0001;
decimal c = .0075m;
decimal d = .0001m;

Console.WriteLine("double");
Console.WriteLine(a / b);
Console.WriteLine(a % b);
Console.WriteLine();
Console.WriteLine("decimal");
Console.WriteLine(c / d);
Console.WriteLine(c % d);

outputs

double
75
9.999999999999937E-05

decimal
75
0.0000

So following the method prescribed by the spec (using division instead of modulation) actually produces the correct result, even with IEEE754 math.