Last Known Copy

CENSUS

Fifty bookmarking services, one search box: what became of the sites this address indexed in 2007

In February 2007 this address published a custom search engine covering fifty social bookmarking services; nineteen years later eight of those names still offer somewhere to save a link, one survives read-only, and the largest single category is not shut down but nobody can now establish what happened.

Published 2026-08-06 · Updated — · 3,672 words · 28 sources · Status of facts checked 2026-08-06

On 2007-02-16 a page was posted at this address describing a Google Co-op custom search engine that queried a list of social bookmarking sites at once, together with a search plugin for Firefox. The Internet Archive holds no capture of the page before 2007-02-18 12:16:59 UTC, which is also its earliest capture returning HTTP 200.1 Thirty-five captures of that URL returned HTTP 200 between that date and 2016-03-13.2

The engine is not the interesting artefact. The list is. Enumerated in the body of the 2007 page, one per line, is a roster of the domains it searched — a dated, third-party-verifiable snapshot of which services a working practitioner considered to constitute the field in February 2007, taken before any of them had a history. That roster is treated here as a primary-source census. Each name is resolved to a status, a date, a statement of what became of the stored collections, and a source. None of the 2007 page’s prose is reproduced, here or anywhere on this site; the roster is a list of addresses, which is a fact about the document rather than a piece of writing in it.

Three findings carry the piece, and none of them is the one the shape of the subject suggests. Being acquired did not predict a service’s death. What happened to the company and what happened to its users’ collections are two different variables, and only the second one mattered to the people who saved things. And the largest category in the census is neither survival nor closure: it is thirty names about which nothing can now be established at all.

Fifty-three URLs, fifty services

The list in the 2007-02-18 capture holds fifty-three URLs.1 The same list, unchanged in membership and in order, appears in the capture of 2007-05-28.3

Fifty of the fifty-three are bookmarking, tagging or link-sharing services identifiable by name. Three are not, and are named here so the exclusion can be checked rather than taken on trust: library.upenn.edu, namakkal.co.uk and supr.c.ilio.us. What each of those three represented on the 2007 roster could not be verified, so each is excluded from the census rather than guessed at. One further name, lifehacker.com, appears on the page but only inside the comment thread beneath it, and is not part of the roster.

Two entries in the list are worth reading closely, because they are the kind of detail a retyped list loses. The roster gives clipclip.org, not the .com; on 2026-08-06 the .org did not resolve, while clipclip.com serves an unrelated Windows clipboard utility. It also gives smarking.com, which on 2026-08-06 redirected to a parking-analytics company that took the name later. Neither of the modern sites is the 2007 service, and a census that started from the names rather than from the captured URLs would have recorded two survivors that do not exist.

The census, in four numbers

Of the fifty names, as of 2026-08-06: eight still offer somewhere to save a link, one survives read-only, eleven no longer offer one and the record establishes what became of them, and for thirty the position could not be verified. Every row is carried, with its confidence grade and sources, in the Ledger; the twenty resolved rows are set out below.

Four of the eight survivors keep a per-account collection of saved links — Diigo, Netvouz, BibSonomy and Linkatopia. The other four — reddit, Slashdot, Digg and Fark — are public submission queues, where saving a link means publishing it. That distinction was already visible in 2007 and the roster made nothing of it, which is itself a finding: at the time these were one category.

The twenty roster names whose position is established — checked 2026-08-06
ServiceFirst seenStatus What happenedCollections on shutdown
BibSonomy2006KEPTAcademic bookmark and publication sharing, maintained by German universities.not applicable
del.icio.us / Delicious2003READ-ONLYChanged hands five times between 2005 and 2017; read-only from 2017-06-15, nothing deleted.login-walled
CiteULike2004GONEClosure announced February 2019; ceased March 2019, exact date disputed.unknown
Digg2004KEPTBroken up and sold in three parts in July 2012; reacquired by its founder on 2025-03-05.not applicable
Diigo2006KEPTBought Furl from LookSmart in March 2009 and absorbed its users.not applicable
Fark1999KEPTStill independently owned by its founder.not applicable
Furl2003GONESold to LookSmart September 2004, to Diigo March 2009; shut 2009-04-17.migrated
Linkatopia2006KEPTCreated 2006 and relaunched in 2020 by the same person, per its own site.not applicable
ma.gnoliacould not be verifiedeGONEDatabase and filesystem failure 2009-01-30; read-only 2010-09-29; closed end of 2010.irretrievable
NowPublic2005GONEAcquired 2009-09-02; closed effective 2013-12-27.unknown
reddit2005KEPTAcquired 2006-10-31, made an independent subsidiary 2011-09-06, listed 2024-03-21.not applicable
Scuttle2004GONESoftware rather than a service; last stable release 2013-04-22.held by whoever ran the install
Simpy2004GONECeased as a bookmarking service April 2010 per Wikipedia; acquirer unconfirmed.unknown
Slashdot1997KEPTOwners: Andover.net 1999, VA Linux 2000, Dice Holdings 2012, BIZX 2016; still publishing.not applicable
Squidoo2005GONEAcquisition by HubPages announced 2014-08-15; content moved there.migrated
StumbleUpon2001GONESold to eBay 2007-05-30, bought back April 2009, shut 2018-06-30.migrated
StyleFeeder2005GONEBought by Time Inc., announced 2010-01-19, and folded into commerce technology.unknown
Netvouz2005, per a 2008 interviewKEPTOne-person bookmark manager, serving the same product it served in 2007.not applicable
Yahoo! My Web2005GONEClosure announced February 2009; shut 2009-03-18.unknown
Zimbio2006GONEParent renamed Livingly Media 2012, acquired 2015; zimbio.com is an entertainment site, not a place to save links.unknown

Scroll →

The remaining thirty names are listed in the Ledger with the status UNVERIFIED and the cell text could not be verified. That is not a placeholder for research still to be done; it is the finding. For twenty-three of them — Jots, Listible, 30daytags, RawSugar, clipclip, Searchles, Gabbr, Wirefan, Taggly, Smarking, SiteTagger, IndiaMarks, Complore, Butterfly Project, BuddyMarks, BMAccess, Markaboo, MyBookmarkManager, i89.us, URLex, ListerLister, SyncOne and socialbookmarking.org — no source for a founding or a closing date could be found at all. For the other seven — Shadows, BlinkList, Clipmarks, eSnips, unalog, folkd and blogmarks.net — something is established about the company and nothing dated is established about the ending or about the current operating status. Earliest Internet Archive captures exist for many of the thirty and are recorded, but a first capture records when a crawler reached a domain, not when a service launched, and is not printed as a founding date.a What this publication declines to assert about each of them, and what evidence would settle it, is set out at the unverified list.

Being bought was not the variable

The tidy version of this history says acquisition killed social bookmarking. The roster does not support it. Of the eight names still accepting saved links, three have changed corporate hands repeatedly and are still running: reddit was acquired on 2006-10-31, became an independent subsidiary on 2011-09-06 and listed on the NYSE on 2024-03-21;4 Slashdot was sold to Andover.net on 1999-06-29, passed to VA Linux in February 2000, to Dice Holdings in 2012 and to BIZX in January 2016, with two renames of the holding company in between;5 Digg was broken into three lots in July 2012 — brand and technology, patents, and staff, each sold separately6 — and was bought back on 2025-03-05 by Kevin Rose, who founded it, with Alexis Ohanian.7

What the closures share is narrower and more useful: in each case the buyer wanted something other than the running service. Furl was bought by LookSmart in September 2004 and sold on to Diigo in March 2009, and the service stopped on 2009-04-17.8 Yahoo built My Web itself and launched it in June 2005,9 six months before it acquired del.icio.us in December 2005,10 and shut My Web on 2009-03-18. Time Inc. bought StyleFeeder, announced 2010-01-19, for commerce technology to put inside a magazine brand.11 HubPages announced the acquisition of Squidoo on 2014-08-15 and moved the content across; on 2026-08-06 http://www.squidoo.com/ answered a redirect to HTTPS and nothing beyond it.12

Against those, the five roster names still serving a bookmarking or link-sharing product for which the record shows no change of ownership at all: Diigo, Netvouz, Linkatopia, Fark and BibSonomy. Fark is still owned by the person who started it in 1999.13 Netvouz was described in a post by Mark Dykeman published on louisgray.com on 2008-08-26 as a one-person operation, in public since 2005, built by Henrik Sjöstrand because he wanted the tool;14 netvouz.com served the same bookmark manager on 2026-08-06.15 Linkatopia states on its own site that it was created by Robert Giordano in 2006 and relaunched in 2020, and it answered on 2026-08-06.16 BibSonomy is maintained by German universities rather than by a company.17

Stated as a rule about the object rather than as a lesson: on this roster, the services that outlived nineteen years are the ones whose continued operation never had to be justified to a buyer. That is a statement about who decides when a service stops, and it is checkable against every row above. It is not a claim that independence guarantees survival, and the census contains its own counter-example: Ma.gnolia was independent, and it is the one row in the table where the collections ceased to exist.

The company’s fate and the collection’s fate are different variables

The column that matters to anyone who saved things is not what happened to the company but what happened to the collections, and the two do not track each other. Three rows make the point on their own.

Furl was sold twice — to LookSmart in 2004, to Diigo in 2009 — and its users’ bookmarks survived both sales: on the 2009-04-17 shutdown they were migrated into Diigo.8 Delicious changed hands five times between 2005 and 2017 — Yahoo, AVOS Systems, Science Inc., Delicious Media, Pinboard10 — and ended with the outcome closest to preservation on the roster: when Pinboard bought it on 2017-06-01 it was put into read-only mode on 2017-06-15, with no new bookmarks and no API, and users were told there was no time pressure to move.18 Ma.gnolia was never sold, and on 2009-01-30 a database and filesystem failure destroyed the collections; on 2009-02-17 its founder stated the data was irretrievable, and what users recovered afterwards came from third-party caches and feeds — that is, from copies held outside the service.1920 The failure is examined separately in the Ma.gnolia article.

So, on the same roster: a service sold twice whose collections were carried across intact on the day it stopped; a service sold five times whose collections were never deleted, and which nine years later sits behind a login rather than behind nothing; and an independent service whose collections ceased to exist in a morning. The variable is not ownership. It is whether anyone other than the service held a copy at the moment it failed — a point the notice-and-window record takes further, since the amount of warning a service gives ranges from months to none.

One roster entry escapes the question entirely, and it is worth naming. Scuttle was not a hosted service but PHP software people installed on their own servers; its last stable release was 2013-04-22 and scuttle.org served a placeholder page on 2026-08-06.21 The project stopping did not take anybody’s bookmarks with it, because the database was already on the user’s own disk. That is the only row in the census where the software’s fate and the collection’s fate are genuinely independent.

Twenty-eight of the fifty domains answer. Six of them are the service

The roster is also a usable test set for something every large link collection eventually needs: an automated check of whether saved links are still good. On 2026-08-06 each of the fifty roster URLs, exactly as written in the 2007 capture, was requested once from a command-line client, following up to six redirects, with a desktop browser user-agent string and a 25-second timeout.b The result is a measurement of how far a status code goes.

One request per roster URL, 2026-08-06
What the request returnedDomainsOf which still a place to save a link
HTTP 200 after redirects286
HTTP 403 or 42952
HTTP 40410
HTTP 52210
301 to HTTPS, then no validated certificate chain20
Name did not resolve90
Connection timed out40
Total508

Scroll →

Twenty-eight of the fifty domains ended at HTTP 200. Six of those twenty-eight were the service the 2007 page listed. The other twenty-two answered with something else. Among them: a Boise embroidery and screen-printing business at shadows.com, a retail site at sitetagger.com, a domain-registrar placeholder at jots.com, a for-sale page at butterflyproject.nl, an entertainment site at zimbio.com, an unrelated content site at blinklist.com, a 114-byte body at unalog.com and another at nowpublic.com whose only content is a script redirecting to a parking lander, and three domains now serving adult or gambling content, which are named without links for that reason.c

One of the twenty-two is a borderline case, and the border is where the definition earns its keep. blogmarks.net returned 200 on 2026-08-06 and is still run by the people who started it; its own homepage states that it was founded in late 2003, that it is non-profit and independent, that new members are accepted only by invitation while the service is being relaunched, and that the public version is inaccessible.28 A domain that answers, is run by its original operators, and cannot be saved into by a member of the public is not counted among the eight; its position is carried as UNVERIFIED rather than as a closure.

Six of the twenty-two ended on a different registrable domain altogether: smarking.com at a parking-analytics company, 30daytags.com at an unrelated gambling site, clipmarks.com at salesforce.com, indiamarks.com at a white-label app builder, markaboo.com at a domain broker, and simpy.com at a registrar’s domain-inquiry form. Those six are catchable by any audit that records the final URL and compares it to the host it recorded at save time. The other sixteen kept their own hostname and changed what they serve, and no status code, redirect chain or DNS check can see that.

The error runs the other way too. Two of the eight survivors did not return 200: fark.com answered 403 and digg.com answered 429 to the automated request. Neither is closed — both are graded as operating in the sourced record — and a bulk checker that treats a non-200 as a dead link would have removed two working sites from a collection. Two further domains answered 403 without settling anything: folkd.com, behind a bot-management challenge, whose operating status therefore could not be verified in either direction, and socialbookmarking.org, which returned a bare 403 from its own server on three consecutive requests.

Counting both directions, a check that reads only the status code is wrong about twenty-four of these fifty rows — twenty-two false positives and two false negatives, just under half the census. The operational consequence is a rule rather than a recommendation: a collection audit that records only HTTP status measures whether a domain is answering, not whether the thing that was saved is still there. An audit that also records the final URL after redirects, the response size and a hash of the body can distinguish the two, and the link-rot audit sets out the stages in order.

A live-looking page is not a live service

One roster name deserves separating out, because it is the clearest case in the census of a page that reads as an operating service and is not one. citeulike.org returned HTTP 200 on 2026-08-06 and served what presents as the CiteULike homepage, including a counter reading “7,849,487 articles - 932 added today”.22 CiteULike announced its closure in February 2019 and ceased in March 2019.23

Two archived captures settle what that page is. The capture of 2026-05-25 03:23:25 UTC renders an online-casino page at the same URL.24 The capture of 2026-07-28 19:38:15 UTC renders the CiteULike homepage, carrying the same two figures — 7,849,487 and 932 — as the live request nine days later.25 A counter that says 932 articles were added today, and reads identically on two dates nine days apart, is not counting anything. This publication does not assert who put that page up or why; what is established is that the numbers do not move, and that the same address served unrelated commercial content in the capture of 2026-05-25.

Stated as a rule: a page that looks like a service is evidence of a page, not of a service. The test costs one request and one archived capture, and it is the same test that separates a working site from a soft 404.

How to repeat any row in this census

Every status line above is reproducible from public endpoints, and the queries are printed rather than described so that a reader checking this article does not have to trust it. As of 2026-08-06 the Internet Archive’s CDX Server API answers a request of this form, where the capture list for one address is returned as JSON:2

https://web.archive.org/cdx/search/cdx?url=EXAMPLE.COM&matchType=exact&output=json&filter=statuscode:200&limit=5

The first row returned is the earliest capture of that address that the archive holds with a 200 status. It bounds the period in which something existed at the domain. It is not a founding date and this census does not print it as one. Adding &limit=-5 in place of the positive limit returns the most recent captures instead, which is how the two CiteULike captures above were located; appending id_ to a capture timestamp in a web.archive.org URL returns the archived bytes without the archive’s own navigation wrapper.

For the roster itself, the query above run against this address returns the thirty-five 200-status captures of the 2007 page.2 The list of fifty-three URLs is in the body of the earliest of them.

The search engine the page was built on has its own history, and it is routinely stated wrongly. Google Programmable Search Engine still exists in 2026. What changed is that the Custom Search Site Restricted JSON API ceased serving traffic on 2025-01-08,26 and that on 2026-01-20 Google capped new engines at fifty domains, required all new engines to use the “sites to search” feature rather than the whole-web option, and set 2027-01-01 as the transition deadline for existing ones.27 Whether the specific engine built at this address in 2007 still functions was not tested and is not claimed either way; the companion article on index scope takes up that platform history alongside the question of what a bookmark search actually indexes.

Where to find it now

This address, /search-engine-all-social-bookmark-services
Thirty-five captures returning HTTP 200 between 2007-02-18 and 2016-03-13; the roster is legible in the earliest. No export ever existed, and the custom search engine behind it was hosted by a third party. Whether that engine still functions was not tested.
del.icio.us / Delicious
Read-only since 2017-06-15, with nothing deleted and no stated deadline. On 2026-08-06 http://del.icio.us/ answered 301 to the HTTPS form and the HTTPS request did not complete for a client that could not validate the certificate chain.d The most recent archived capture returning 200, 2026-08-04 19:02:48 UTC, is a login form and nothing else. Public profile pages are in the archive and can be tested per-user with the CDX query above. Fuller record here.
Furl
Shut 2009-04-17; bookmarks migrated into Diigo, which is still operating. furl.net did not resolve on 2026-08-06. The migrated records live inside Diigo accounts, not at any Furl address.
Ma.gnolia
Collections irretrievable from 2009-01-30, confirmed by its founder on 2009-02-17. No export, no window, no successor. What users recovered came from third-party caches and feeds. Captures of the site exist in the archive; the collections do not exist anywhere.
Yahoo! My Web
Announced February 2009, shut 2009-03-18. myweb.yahoo.com did not resolve on 2026-08-06. What became of stored bookmarks at shutdown could not be verified.
StumbleUpon
Closed 2018-06-30 with accounts migrated to Mix. stumbleupon.com returned 200 on 2026-08-06 with a page directing visitors to Mix; the address is a signpost, not a service.
CiteULike
Announced February 2019, ceased March 2019, exact date disputed. citeulike.org returns a page presenting the old homepage with counters that do not change; the domain served unrelated commercial content as recently as the capture of 2026-05-25. No export path is offered at that address.
Squidoo
Acquisition by HubPages announced 2014-08-15; content migrated there. On 2026-08-06 http://www.squidoo.com/ answered 301 to the HTTPS form and the request did not complete.
Simpy
Ceased as a bookmarking service in April 2010 per Wikipedia; the acquirer is asserted by Wikipedia alone and no contemporaneous report was found. simpy.com redirected to a registrar’s domain-inquiry form on 2026-08-06.
Scuttle
Self-hosted software, last stable release 2013-04-22; scuttle.org served a placeholder page on 2026-08-06. Any surviving data is on the disk of whoever ran the installation, in their own MySQL database, and was never held by the project.
The thirty unverified names
No closure announcements, no dated reports, and in most cases no surviving operator. Earliest archived captures were collected for many of them and are recorded in the Ledger as bounds, explicitly not as founding dates. The complete list of what this publication declines to assert about them, and what evidence would settle each, is at the unverified list.

References

  1. Internet Archive. “Capture of infopirate.org/search-engine-all-social-bookmark-services.” Wayback Machine, 2007-02-18. https://web.archive.org/web/20070218121659/http://infopirate.org/search-engine-all-social-bookmark-services — retrieved 2026-08-06. ↩
  2. Internet Archive. “Wayback CDX Server API.” internetarchive/wayback repository, n.d. https://github.com/internetarchive/wayback/tree/master/wayback-cdx-server — retrieved 2026-08-06. Query used for this article: url=infopirate.org/search-engine-all-social-bookmark-services&matchType=exact&output=json&filter=statuscode:200, returning 35 rows dated 2007-02-18 to 2016-03-13. ↩
  3. Internet Archive. “Capture of infopirate.org/search-engine-all-social-bookmark-services.” Wayback Machine, 2007-05-28. https://web.archive.org/web/20070528150354/http://infopirate.org/search-engine-all-social-bookmark-services — retrieved 2026-08-06. ↩
  4. Wikipedia contributors. “Reddit.” Wikipedia, n.d. https://en.wikipedia.org/wiki/Reddit — retrieved 2026-08-06. Cited as a secondary source. ↩
  5. Wikipedia contributors. “Slashdot.” Wikipedia, n.d. https://en.wikipedia.org/wiki/Slashdot — retrieved 2026-08-06. Cited as a secondary source. ↩
  6. Wikipedia contributors. “Digg.” Wikipedia, n.d. https://en.wikipedia.org/wiki/Digg — retrieved 2026-08-06. Cited as a secondary source. ↩
  7. Sawers, Paul. “Kevin Rose, Alexis Ohanian acquire Digg.” TechCrunch, 2025-03-05. https://techcrunch.com/2025/03/05/kevin-rose-and-alexis-ohanian-acquire-digg/ — retrieved 2026-08-06. ↩
  8. Wikipedia contributors. “Furl.” Wikipedia, n.d. https://en.wikipedia.org/wiki/Furl — retrieved 2026-08-06. Cited as a secondary source. ↩
  9. Wikipedia contributors. “My Web.” Wikipedia, n.d. https://en.wikipedia.org/wiki/My_Web — retrieved 2026-08-06. Cited as a secondary source. ↩
  10. Wikipedia contributors. “Delicious (website).” Wikipedia, n.d. https://en.wikipedia.org/wiki/Delicious_(website) — retrieved 2026-08-06. Cited as a secondary source. The price Yahoo paid in 2005 was never disclosed and is not stated here. ↩
  11. Wauters, Robin. “Confirmed: Time Inc. Buys Personal Shopping Engine StyleFeeder.” TechCrunch, 2010-01-19. https://techcrunch.com/2010/01/19/time-buys-stylefeeder/ — retrieved 2026-08-06. ↩
  12. Wikipedia contributors. “Squidoo.” Wikipedia, n.d. https://en.wikipedia.org/wiki/Squidoo — retrieved 2026-08-06. Cited as a secondary source; the live check of www.squidoo.com is dated 2026-08-06. ↩
  13. Wikipedia contributors. “Fark.” Wikipedia, n.d. https://en.wikipedia.org/wiki/Fark — retrieved 2026-08-06. Cited as a secondary source. ↩
  14. Dykeman, Mark. “Netvouz – A Quality One Man Bookmarking Show.” louisgray.com, 2008-08-26. Archived at https://web.archive.org/web/20230205035125/https://blog.louisgray.com/2008/08/netvouz-quality-one-man-bookmarking.html — retrieved 2026-08-06. The live host did not resolve on 2026-08-06; the archived capture of 2023-02-05 is the copy cited. ↩
  15. Netvouz. “Netvouz – Online Social Bookmark Manager.” netvouz.com, n.d. https://netvouz.com/ — retrieved 2026-08-06. ↩
  16. Linkatopia. “Linkatopia – Your Bookmarks and Social Links Manager.” linkatopia.com, n.d. https://linkatopia.com/ — retrieved 2026-08-06. ↩
  17. Wikipedia contributors. “BibSonomy.” Wikipedia, n.d. https://en.wikipedia.org/wiki/BibSonomy — retrieved 2026-08-06. Cited as a secondary source. ↩
  18. Pinboard. “Pinboard Acquires Delicious.” Pinboard Blog, 2017-06-01. https://blog.pinboard.in/2017/06/pinboard_acquires_delicious/ — retrieved 2026-08-06. ↩
  19. Backblaze. “The Story of Ma.gnolia: A Social Bookmarking Site That Wilted with No Backup.” Backblaze Blog, 2009-03-03. https://www.backblaze.com/blog/magnolia-wilts-with-no-backup/ — retrieved 2026-08-06. ↩
  20. Wikipedia contributors. “Ma.gnolia.” Wikipedia, n.d. https://en.wikipedia.org/wiki/Ma.gnolia — retrieved 2026-08-06. Cited as a secondary source. ↩
  21. Wikipedia contributors. “Scuttle (software).” Wikipedia, n.d. https://en.wikipedia.org/wiki/Scuttle_(software) — retrieved 2026-08-06. Cited as a secondary source; the live check of scuttle.org is dated 2026-08-06. ↩
  22. CiteULike. “CiteULike: Everyone’s library.” citeulike.org, n.d. https://citeulike.org/ — retrieved 2026-08-06. ↩
  23. Wikipedia contributors. “CiteULike.” Wikipedia, n.d. https://en.wikipedia.org/wiki/CiteULike — retrieved 2026-08-06. Cited as a secondary source; the exact closure date is disputed and is not asserted here. ↩
  24. Internet Archive. “Capture of citeulike.org.” Wayback Machine, 2026-05-25. https://web.archive.org/web/20260525032325/https://citeulike.org/ — retrieved 2026-08-06. ↩
  25. Internet Archive. “Capture of citeulike.org.” Wayback Machine, 2026-07-28. https://web.archive.org/web/20260728193815/https://citeulike.org/ — retrieved 2026-08-06. ↩
  26. Google. “Custom Search Site Restricted JSON API.” Google for Developers, n.d. https://developers.google.com/custom-search/v1/site_restricted_api — retrieved 2026-08-06. States the API will cease to serve traffic on January 8, 2025. ↩
  27. Google. “Updates to our Web Search Products & Programmable Search Engine Capabilities.” Programmable Search Engine Blog, 2026-01-20. https://programmablesearchengine.googleblog.com/2026/01/updates-to-our-web-search-products.html — retrieved 2026-08-06. ↩
  28. Blogmarks. “Blogmarks: Public marks.” blogmarks.net, n.d. http://blogmarks.net/ — retrieved 2026-08-06. ↩

Notes

  1. Earliest captures recorded for several unverified roster domains include 2005-05-26 for sitetagger.com, 2005-11-26 for buddymarks.com, 2005-12-12 for butterflyproject.nl, 2006-02-03 for listible.com, 2006-05-10 for complore.com, 2006-06-15 for markaboo.com, 2006-07-05 for searchles.com and taggly.com, and 2006-07-12 for wirefan.com. Each bounds the period in which a domain was reachable. None is a founding date, and none is printed as one.
  2. Method for the response table: one HTTP GET per URL as written in the 2007 capture, from a single network on 2026-08-06, following up to six redirects, 25-second timeout, desktop browser user-agent string. A single request from one vantage point is not a definitive statement about a site’s availability; a 403 or 429 is a refusal to answer that request, not evidence of closure, and the same domain can fall into a different row on a different day or from a different address. The eight survivors are graded from the sourced record rather than from this check. One domain in the 200 row answered this request while the sourced record held for it says otherwise; the discrepancy is logged in the Ledger rather than resolved here.
  3. Three roster domains serve adult or gambling content, unrelated to the 2007 services, at their own hostnames: urlex.info, listerlister.com and mybookmarkmanager.com. A fourth, 30daytags.com, redirects to cbcgames.org, which serves the same class of content. They are named because a census that silently dropped them would be less accurate, and none is linked, because this publication does not send readers to them.
  4. A certificate chain that a command-line client cannot validate does not always fail in a browser, which may fetch a missing intermediate certificate on its own. What is stated is what the request returned on 2026-08-06, not a claim about every client.
  5. Ma.gnolia’s founding year is given as 2006 by Wikipedia, while a 2009 interview implies about 2005. The year could not be verified, so the census prints none.

From early 2007 until roughly 2010 this address served a Drupal-based social bookmarking community whose users submitted links with tags, notes and votes; this page was posted there on 2007-02-16 and announced a custom search engine across the roster examined above. Captures of this URL returning HTTP 200 run from 2007-02-18 to 2016-03-13, and the site itself had been abandoned or repurposed well before the later ones. Posts at lifehacker.com and au.lifehacker.com still point at this address, as does u-g-h.com (checked 2026-08-06). Those citations are nineteen years old and outlived the thing they described, which is the subject of this page rather than a footnote to it.