A single link, no trailing path, and a question: how much actually sits behind it? Google Sites exposes no directory listing, so this is the reconstruction — search index, web archive, on-site links, then a direct HTTP probe of every candidate path to sort the living from the dead.
https://sites.google.com/online.island.edu.hk returns HTTP 404 with an
empty body. It is not a homepage — it is the namespace the school's Google Workspace
publishes Sites into. Every real site hangs one path segment below it, like
/maths or /questweek. So "how many" can only be answered by counting
those segments, and Google will not hand you the list.
No single source is complete, and each one lies in its own direction: the search index is shallow, the archive is stale, and on-site links only reach siblings someone bothered to link. So they were unioned and then every candidate was verified against the live host.
site:sites.google.com/online.island.edu.hk paginated
30 pages deep. Gave 12 slugs before Google served a CAPTCHA and the crawl stopped itself.
(595 indexed results reported)/online.island.edu.hk/* and the legacy /a/online.island.edu.hk/* form.
Gave 162 slugs and their page trees — the single biggest contribution.island.edu.hk, its sitemap, and the
live Sites' own raw HTML surfaced is-experiential-learning, which neither index had.accounts.google.com = gated. 11 vs 153. Zero false
positives: every public row was re-checked with a plain curl and returned 200.Live table, generated from this session's probe output. Filter by status, or type to search path and title. "Nav entries" is how many page links the site's own navigation HTML carries — a lower bound on real pages, since Google Sites only renders the nav it ships.
Showing 164 of 164 paths
| Path | Status | Title | Nav |
|---|
All verified in the final pass: 200, real <title>, no sign-in redirect.
| Path | Site title | Nav entries |
|---|---|---|
| /maths | IS Mathematics | 196 |
| /islandschoolfrench | Island School French | 135 |
| /island-school-spanish | Island School Spanish | 97 |
| /islandschooljapanese | Island School Japanese | 84 |
| /island-school-chinese | IS Chinese | 83 |
| /is-experiential-learning | IS – Elements | 77 |
| /islandschool-highereducation | Island School Higher Education | 56 |
| /ibcp | IBCP@Island School | 36 |
| /iscas | CAS at Island School | 16 |
| /questweek | Quest Week | 13 |
| /gpcookbook | GP Cookbook | 7 |
The five language sites alone carry roughly half the public surface. /maths is the
largest single site at ~196 navigation entries — larger than every other subject site combined
except French.
These all resolve, then bounce to a Google sign-in page: unpublished, restricted to the school
domain, or deleted-but-still-routed. The archive says what most of them were — student project
sites from roughly 2009–2013, back when Google Sites was the school's default publishing tool
and every class produced a dozen /james-games-style pages.
A sign-in redirect proves the path is still registered, not that the content survives behind it.
Some are genuinely private course sites; many are simply unpublished student work that Google
keeps routing. Reading them needs a @online.island.edu.hk session — which this
investigation deliberately did not use.
164 is the number of paths visible from outside, not the number that exists. A Site that was never published, never linked, never crawled and never archived is invisible to all four sources — and Google Sites has no API that lists a domain's sites without Workspace admin rights. The real figure is ≥164. The complete list lives behind the Sites dashboard, not the open web.
Two more bounded claims worth stating plainly: the Google leg stopped at a CAPTCHA (30 pages, not the reported 595 results), and the archive leg only sees what the Wayback Machine happened to crawl — its newest captures here are thin, so a site published in the last few months could hinge entirely on the on-site-link and probe legs.
| Check | Result | How |
|---|---|---|
| Namespace root | 404, 0 bytes | curl -sSI |
| Public sites reachable | 11 / 11 | curl + title grep |
| Public nav entries | ~800 | regex on nav HTML |
| Gated paths | 153 | 302 → accounts.google.com |
| Archive slugs seen | 162 | Wayback CDX, statuscode:200 |
| Credentials / auth attempts | 0 | public GETs only |
| Search leg stopped | CAPTCHA at page 8 | crawler self-abort |