WebAssembly edits by nrllh · Pull Request #4401 · HTTPArchive/almanac.httparchive.org

nrllh · 2026-01-29T20:17:07Z

I have edited the chapter. Could you folks please take a look at it and take further adjustments?

@nimeshgit : we still need the tool "almanac-wasm". Can you please provide it to us?

Enhanced clarity and corrected minor inaccuracies in the text. It is not finalized yet.

Updated date filters and added new metrics for WebAssembly requests.

github-actions · 2026-01-29T20:18:50Z

Super-linter summary

Language	Validation result
EDITORCONFIG	Fail ❌
MARKDOWN	Pass ✅

Super-linter detected linting errors

For more information, see the GitHub Actions workflow run

Powered by Super-linter

EDITORCONFIG

�[33;1msql/2025/webassembly/counts.sql:�[33;0m
�[31;1m	10: Trailing whitespace�[33;0m
�[31;1m	41: Trailing whitespace�[33;0m
�[33;1msrc/content/en/2025/webassembly.md:�[33;0m
�[31;1m	39: Trailing whitespace�[33;0m
�[31;1m	53: Trailing whitespace�[33;0m
�[31;1m	97: Trailing whitespace�[33;0m
�[31;1m	97: Wrong amount of left-padding spaces(want multiple of 2)�[33;0m
�[31;1m	119: Trailing whitespace�[33;0m
�[31;1m	189: Trailing whitespace�[33;0m
�[31;1m
8 errors found�[33;0m

github-actions · 2026-01-29T20:22:55Z

Super-linter summary

Language	Validation result
EDITORCONFIG	Pass ✅
MARKDOWN	Pass ✅

All files and directories linted successfully

For more information, see the GitHub Actions workflow run

Powered by Super-linter

github-actions · 2026-01-29T20:24:49Z

Super-linter summary

Language	Validation result
EDITORCONFIG	Pass ✅
MARKDOWN	Pass ✅

All files and directories linted successfully

For more information, see the GitHub Actions workflow run

Powered by Super-linter

github-actions · 2026-01-29T20:26:27Z

Super-linter summary

Language	Validation result
EDITORCONFIG	Pass ✅
MARKDOWN	Pass ✅

All files and directories linted successfully

For more information, see the GitHub Actions workflow run

Powered by Super-linter

github-actions · 2026-01-29T20:27:15Z

Super-linter summary

Language	Validation result
EDITORCONFIG	Pass ✅
MARKDOWN	Pass ✅

All files and directories linted successfully

For more information, see the GitHub Actions workflow run

Powered by Super-linter

src/content/en/2025/webassembly.md

tunetheweb · 2026-01-29T20:49:14Z

src/content/en/2025/webassembly.md

 }}

-These WebAssembly modules differ considerably in size, with the smallest being just a few kilobytes, and the largest one is 228.102 MB in desktop's client and 166.415 MB for mobile client.
+When examining uncompressed sizes, we observe that while the median module remains lightweight at approximately 30 KB on both platforms, the largest binaries at the 90th percentile are significantly heavier on desktop (897 KB) than on mobile (756 KB).


Is Wasm not typically binary? What compression does it use? Gzip? Brotli? I thought those only really worked on text resources (maybe explaining the relatively low different between these two charts?) and they shouldn't be further compressed for sending on the wire?

Barry, The Wasm is fundamentally a binary instruction format, not typically a text format. the format sent over the wire is the compact binary format (wasm).

It has built in compression.

It is observed that applying Gzip or Brotli to an already highly optimized binary file yields minimal additional size savings (often less than 10%, sometimes even negligible), Because of the minimal benefit of compression, Barry the Wasm files are typically served with an identity encoding, meaning no further general purpose compression is applied

Is it worth commenting on this?

At the moment we show raw (which I think should be renamed "compress" as per other conversation to make it more obvious) and uncompressed, but never really explain that it's not really expected for them to be compressed. And that that explains why the charts aren't that different. By presenting both, it's almost like the opposite and we think it is something to be looked at, almost like we expect there to be interesting insights here?

So I went back to the 2021 chapter that we seem to be basing this on and it has this to say:

"...one of the benefits of Wasm bytecode is that it’s highly compressible, and size over the wire is what matters for download speed and billing reasons. Let’s check sizes of raw response bodies as sent by servers instead:"

That's a good narrative for why they included both.

But the 2025 chapter doesn't seem to have a similar narrative and just says "here's the raw bytes, here's the compressed bytes" and I'm not sure what the readers's supposed to take away from that.

Additionally it contradicts what you say about compression (and what I believed too).

Again, as a reader I'm confused as to what to take away from this section?

Hi Nurullah and Barry,
It seems to resolve this concern and to make it easy to understand,
It would be good to have the compression method chart along with its details.

What do you say, guys ?

I'm not sure what you mean by that? I'm still very confused here..

lets talk on slack tomorrow afternoon about this.

src/content/en/2025/webassembly.md

Co-authored-by: Barry Pollard <barrypollard@google.com>

Removed redundant sentence from introduction

github-actions · 2026-01-29T23:33:03Z

Super-linter summary

Language	Validation result
EDITORCONFIG	Pass ✅
MARKDOWN	Pass ✅

All files and directories linted successfully

For more information, see the GitHub Actions workflow run

Powered by Super-linter

github-actions · 2026-01-29T23:34:30Z

Super-linter summary

Language	Validation result
EDITORCONFIG	Pass ✅
MARKDOWN	Pass ✅

All files and directories linted successfully

For more information, see the GitHub Actions workflow run

Powered by Super-linter

github-actions · 2026-01-29T23:34:32Z

Super-linter summary

Language	Validation result
EDITORCONFIG	Pass ✅
MARKDOWN	Pass ✅

All files and directories linted successfully

For more information, see the GitHub Actions workflow run

Powered by Super-linter

Updated the data collection section to reflect the identification of WebAssembly modules on sites analyzed.

github-actions · 2026-01-29T23:36:15Z

Super-linter summary

Language	Validation result
EDITORCONFIG	Pass ✅
MARKDOWN	Pass ✅

All files and directories linted successfully

For more information, see the GitHub Actions workflow run

Powered by Super-linter

nimeshgit · 2026-01-30T13:54:10Z

I have edited the chapter. Could you folks please take a look at it and take further adjustments?

@nimeshgit : we still need the tool "almanac-wasm". Can you please provide it to us?

please, have the link almanac-wasm-stats
We are updating stack still that includes documentation and tests.

Later on We would be raising pull request and will merge the stack under HTTPArchive and enable in the pipeline with the help of Patrick and Barry.

github-actions · 2026-01-31T17:53:13Z

Super-linter summary

Language	Validation result
EDITORCONFIG	Pass ✅
MARKDOWN	Pass ✅

All files and directories linted successfully

For more information, see the GitHub Actions workflow run

Powered by Super-linter

tunetheweb

I'm still very confused and not sure this PR make this text any less confusing. Can we address the comments I've raised @nrllh and @nimeshgit ?

src/content/en/2025/webassembly.md

tunetheweb · 2026-02-08T13:07:17Z

src/content/en/2025/webassembly.md

+
+WebAssembly module sizes vary drastically based on their specific use cases. We observed that the bottom 50% of modules are quite small, ranging between 2 KB and 14 KB. These are typically "micro-utilities" like Base64 encoders or checksum calculators, often written in AssemblyScript or Rust to handle performance-critical tasks where JavaScript lacks precision.
+
+Conversely, at the 90th percentile, sizes increase significantly to 381 KB on desktop and 316 KB on mobile. These larger binaries usually represent full desktop-grade applications ported to the web—such as Adobe Photoshop or Google Earth—compiled from heavier languages like C++ or C# to handle complex 3D rendering and logic.


Can we add something like this to the text?

tunetheweb · 2026-02-08T13:07:48Z

src/content/en/2025/webassembly.md

 }}

-These WebAssembly modules differ considerably in size, with the smallest being just a few kilobytes, and the largest one is 228.102 MB in desktop's client and 166.415 MB for mobile client.
+When examining uncompressed sizes, we observe that while the median module remains lightweight at approximately 30 KB on both platforms, the largest binaries at the 90th percentile are significantly heavier on desktop (897 KB) than on mobile (756 KB).


I'm not sure what you mean by that? I'm still very confused here..

src/content/en/2025/webassembly.md

Removed mention of Microsoft's dominance in WebAssembly ecosystem.

As per the pull request (#4401), enhanced compression methods, data collection and analysis methods, revised statistics on WebAssembly module sizes and usage. Improved clarity on language usage and library dominance in WebAssembly + other minor edits as suggested.

revise language usage that counts duplicates in percentage

we have updated compression methods for both clients desktop and mobile + updated language usage chart with duplicate records counts percentage stats.

github-actions · 2026-02-10T08:05:41Z

Images automagically compressed by Calibre's image-actions ✨

Compression reduced images by 80.2%, saving 777.4 KB.

Filename	Before	After	Improvement	Visual comparison
`src/static/images/2025/webassembly/language-usage.png`	397.4 KB	70.0 KB	82.4%	View diff
`src/static/images/2025/webassembly/compression-methods-desktop-client.png`	286.4 KB	52.9 KB	81.5%	View diff
`src/static/images/2025/webassembly/compression-methods-mobile-client.png`	285.1 KB	68.6 KB	75.9%	View diff

nimeshgit · 2026-02-10T08:12:17Z

Barry, we have merged @nrllh updates plus your suggetions and compression methods for desktop and mobile clients along with couple of edits.

src/content/en/2025/webassembly.md

github-actions · 2026-02-10T10:54:15Z

Images automagically compressed by Calibre's image-actions ✨

Compression reduced images by 19%, saving 26.4 KB.

Filename	Before	After	Improvement	Visual comparison
`src/static/images/2025/webassembly/compression-methods-mobile-client.png`	68.6 KB	55.2 KB	19.5%	View diff
`src/static/images/2025/webassembly/language-usage.png`	70.0 KB	57.0 KB	18.6%	View diff

1 image did not require optimisation.

github-actions · 2026-02-10T14:36:07Z

Images automagically compressed by Calibre's image-actions ✨

Compression reduced images by 43%, saving 23.8 KB.

Filename	Before	After	Improvement	Visual comparison
`src/static/images/2025/webassembly/compression-methods-mobile-client.png`	55.2 KB	31.5 KB	43.0%	View diff

2 images did not require optimisation.

github-actions · 2026-02-10T14:42:04Z

Images automagically compressed by Calibre's image-actions ✨

Compression reduced images by 25.9%, saving 16.5 KB.

Filename	Before	After	Improvement	Visual comparison
`src/static/images/2025/webassembly/language-usage.png`	31.3 KB	22.9 KB	26.8%	View diff
`src/static/images/2025/webassembly/compression-methods-desktop-client.png`	16.3 KB	12.2 KB	25.1%	View diff
`src/static/images/2025/webassembly/compression-methods-mobile-client.png`	16.1 KB	12.1 KB	25.1%	View diff

tunetheweb

OK let's merge this.

Thanks for improving the text!

github-actions · 2026-02-10T15:18:07Z

Images automagically compressed by Calibre's image-actions ✨

Compression reduced images by 26%, saving 5.8 KB.

Filename	Before	After	Improvement	Visual comparison
`src/static/images/2025/webassembly/webassembly-by-rank.png`	22.3 KB	16.5 KB	26.0%	View diff

3 images did not require optimisation.

…parchive.org into webassembly-pass

nrllh added 3 commits January 28, 2026 23:16

Revise WebAssembly 2025

1d25a6c

Enhanced clarity and corrected minor inaccuracies in the text. It is not finalized yet.

Enhanced unique metrics

5403da9

Updated date filters and added new metrics for WebAssembly requests.

Final edits

3737f38

nrllh requested review from nimeshgit and tunetheweb January 29, 2026 20:17

tunetheweb added 2 commits January 29, 2026 20:21

Linting fixes

554a236

SQL JSON_VALUE preference

2db0755

Restore SQL comment

45c892f

Another comment

c4b261b

tunetheweb requested changes Jan 29, 2026

View reviewed changes

nrllh and others added 2 commits January 29, 2026 15:31

Apply suggestions from code review

712e381

Co-authored-by: Barry Pollard <barrypollard@google.com>

Further edits

bf892e5

Removed redundant sentence from introduction

Fix formatting in WebAssembly introduction section

a548ca1

Revise WebAssembly data collection statistics

b9f034b

Updated the data collection section to reflect the identification of WebAssembly modules on sites analyzed.

Merge branch 'main' into webassembly-pass

c277939

Merge branch 'main' into webassembly-pass

b39e948

tunetheweb added the editing Content excellence label Feb 8, 2026

tunetheweb requested changes Feb 8, 2026

View reviewed changes

Apply suggestions from code review

9b153cf

nimeshgit added a commit that referenced this pull request Feb 9, 2026

merge edits and suggestions of pull request #4401

fa27d47

nimeshgit mentioned this pull request Feb 9, 2026

suggestions of pull request #4401 #4409

Closed

nrllh and others added 5 commits February 9, 2026 15:05

Update WebAssembly libraries popularity statement

8776217

Removed mention of Microsoft's dominance in WebAssembly ecosystem.

Delete src/static/images/2025/webassembly/language-usage.png

617ad95

revise language usage that counts duplicates in percentage

added compression methods + updated language usage

7174150

we have updated compression methods for both clients desktop and mobile + updated language usage chart with duplicate records counts percentage stats.

Optimised images with calibre/image-actions

112e9d1

nimeshgit approved these changes Feb 10, 2026

View reviewed changes

src/content/en/2025/webassembly.md Show resolved Hide resolved

src/content/en/2025/webassembly.md Show resolved Hide resolved

src/content/en/2025/webassembly.md Show resolved Hide resolved

nimeshgit requested a review from tunetheweb February 10, 2026 10:49

nimeshgit and others added 2 commits February 10, 2026 16:23

Merge branch 'main' into webassembly-pass

a801ca0

Optimised images with calibre/image-actions

11a4fb9

tunetheweb and others added 2 commits February 10, 2026 14:35

Merge branch 'main' into webassembly-pass

26f645b

Optimised images with calibre/image-actions

ebef541

tunetheweb and others added 2 commits February 10, 2026 09:40

Retake images

2cbb3d2

Optimised images with calibre/image-actions

64388f0

Minor edits

201c768

tunetheweb approved these changes Feb 10, 2026

View reviewed changes

tunetheweb and others added 3 commits February 10, 2026 09:58

More edits

7ce58c3

Retake image

0fa7a63

Optimised images with calibre/image-actions

5d913fc

tunetheweb added 2 commits February 10, 2026 10:23

Edits

8a06e23

Merge branch 'webassembly-pass' of github.com:HTTPArchive/almanac.htt…

e239783

…parchive.org into webassembly-pass

tunetheweb merged commit 59ffa73 into main Feb 10, 2026
10 checks passed

tunetheweb deleted the webassembly-pass branch February 10, 2026 15:33


		WebAssembly module sizes vary drastically based on their specific use cases. We observed that the bottom 50% of modules are quite small, ranging between 2 KB and 14 KB. These are typically "micro-utilities" like Base64 encoders or checksum calculators, often written in AssemblyScript or Rust to handle performance-critical tasks where JavaScript lacks precision.

		Conversely, at the 90th percentile, sizes increase significantly to 381 KB on desktop and 316 KB on mobile. These larger binaries usually represent full desktop-grade applications ported to the web—such as Adobe Photoshop or Google Earth—compiled from heavier languages like C++ or C# to handle complex 3D rendering and logic.

Uh oh!

Comments

Conversation

nrllh commented Jan 29, 2026

Uh oh!

github-actions bot commented Jan 29, 2026

Super-linter summary

Uh oh!

github-actions bot commented Jan 29, 2026

Super-linter summary

Uh oh!

github-actions bot commented Jan 29, 2026

Super-linter summary

Uh oh!

github-actions bot commented Jan 29, 2026

Super-linter summary

Uh oh!

github-actions bot commented Jan 29, 2026

Super-linter summary

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

tunetheweb Jan 29, 2026

Choose a reason for hiding this comment

Uh oh!

nimeshgit Jan 30, 2026

Choose a reason for hiding this comment

Uh oh!

tunetheweb Jan 30, 2026

Choose a reason for hiding this comment

Uh oh!

nimeshgit Feb 7, 2026

Choose a reason for hiding this comment

Uh oh!

tunetheweb Feb 8, 2026

Choose a reason for hiding this comment

Uh oh!

nimeshgit Feb 8, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

github-actions bot commented Jan 29, 2026

Super-linter summary

Uh oh!

github-actions bot commented Jan 29, 2026

Super-linter summary

Uh oh!

github-actions bot commented Jan 29, 2026

Super-linter summary

Uh oh!

github-actions bot commented Jan 29, 2026

Super-linter summary

Uh oh!

nimeshgit commented Jan 30, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

github-actions bot commented Jan 31, 2026

Super-linter summary

Uh oh!

tunetheweb left a comment

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

tunetheweb Feb 8, 2026

Choose a reason for hiding this comment

Uh oh!

tunetheweb Feb 8, 2026

Choose a reason for hiding this comment

nimeshgit Feb 8, 2026 •

edited

Loading

nimeshgit commented Jan 30, 2026 •

edited

Loading