About COMPUTEX.md

What this is, where the facts come from, and how to change them

本站是獨立專案，非外貿協會（TAITRA）或 COMPUTEX 官方網站，未經其授權或背書。所有資料整理自公開來源。

This is an independent project. It is not the official website of TAITRA or COMPUTEX, and carries no authorisation or endorsement from either. All data is compiled from publicly available sources.

COMPUTEX.md is an open archive of COMPUTEX exhibitors, products and editions. One file per company, in plain Markdown, accumulating year after year instead of resetting when the show closes.

The show runs for four days a year. The supply chain behind it does not. A directory that only exists during those four days cannot answer the question people actually ask the rest of the year: who makes this, and how long have they been doing it.

So the unit here is the company, not the event. A page opened this year is the same page that gets added to next year, and the year after.

Why the .md

Every page here is a Markdown file in a public Git repository. That is not a technical detail, it is the product. A Markdown file can be diffed, so every change has an author and a date. It can be quoted by a language model without a rendering step. It can be forked by anyone if this project ever stops being maintained.

Each page also serves a plain-text twin at the same address plus a .md suffix, and the whole archive is indexed in llms.txt. Machines read the same thing you do, not a stripped-down version of it.

Browse the vendor pages

The editorial charter

Multiple parties with commercial interests co-edit this archive. That only works if it is clear, in advance, who may write what. These three rules are enforced in review, not by good intentions.

  1. The fact layer

    Company name, venue, booth number, official URL, product tags, and the year-by-year exhibiting record. Extracted from the official exhibitor directory by a script, with no language model rewriting anything in between. Every row carries its source URL and the date it was checked.

    Who may write it: Anyone, including the company itself, through a pull request. A change is accepted when it carries a source that a third party can check. A correction without a source is declined, however obviously true it looks.

  2. The curation layer

    Where a company sits in the supply chain, what changed since last year, how one edition compares to the last, which technologies clustered on the floor. This is judgement, not record.

    Who may write it: Neutral editors only. A company can dispute a curation passage in a pull request and the dispute is answered in public, but it cannot edit the passage directly. Without this line an AI engine has no reason to treat this archive as anything other than a brochure.

  3. The blank

    When the official directory does not publish a field, the page leaves it empty and says so.

    Who may write it: Nobody fills a blank from a company website, a press release, last year’s pattern, or a third-party aggregator. Filling it takes a verifiable source. A page that looks complete is worth nothing if a single row of it was inferred.

Where the facts come from

What is the source?
The official COMPUTEX exhibitor directory. Each vendor page links to the exact directory entry it was built from, and every row in the fact table repeats that link.
How is it extracted?
By a script in this repository, which you can read and run. No language model touches the fact layer, because a model that paraphrases a booth number is indistinguishable from a model that invents one.
How current is it?
Every page carries the date it was last checked, in the frontmatter and next to each fact row. When you are reading a page, that date is the honest answer, not the date at the top of this paragraph.
What if it is wrong?
Open an issue or send a pull request. Corrections to the fact layer need a source; that is the only requirement, and it applies equally to us and to you.
86
exhibitors on file
64
with 2+ COMPUTEX editions
2010-2026
years covered
87
pages published

Where this came from

COMPUTEX.md is a derivative of Taiwan.md, an open-source knowledge base built by Frank Wu that demonstrated something worth copying: a plain-Markdown archive, structured for machine reading, can become what a language model quotes when someone asks about a subject.

The site scaffolding, the quality-gate architecture and the AI-readable layer come from there. The taxonomy, the editorial charter and every word in knowledge/ are written here. The parent project explicitly invites derivatives and asks only that lineage stay visible, which is what this section is for.

Questions people actually ask

Is this the official COMPUTEX website?
No. The official site is computextaipei.com.tw, run by TAITRA. This is an independent archive with no authorisation or endorsement from them. If that ever changes, it will say so here in plain language and the change will be in the Git history.
My company has a page here and nobody asked us. Why?
Because the page contains only facts your company already published through the official exhibitor directory: name, booth, venue, official URL and exhibiting history. Nothing on it came from anywhere else. The page names its source on every row so you can check that claim rather than take our word for it.
We want the page changed or taken down.
Open an issue and say which. Corrections are usually faster and better for you than removal, but removal is a legitimate request and is handled the same way, in public, with the reason recorded.
Is this AI-generated?
The fact layer is not. It is extracted mechanically and never passes through a language model. The curation layer is written by people, with AI assistance in drafting, and every page records whether a human has reviewed it. Where that flag is false, it is visible rather than hidden.
Can I reuse this?
Yes. Content is CC BY 4.0, so commercial reuse is fine as long as you attribute. Code is MIT. Attribution is the only thing asked for, deliberately: the point of this archive is to be quoted.
Why does it look like Taiwan.md?
Because it is a derivative of it, by design and with the parent project’s blessing. See the lineage section above.

Claim, correct, or ask

Every change to this archive goes through a public pull request, including ours. If your company is in here, that is the same door you use.

Open an issue How to contribute hans@groupg.org