Read history

13 June 1990: systematic demolition begins

On 13 June 1990, the planned, official demolition of Berlin's remaining border installations began at Bernauer Strasse and Ackerstrasse, between Wedding in West Berlin and Mitte in the East. DDR Border Troops — the force that had guarded the Wall — operated the cranes, excavators and tracked machinery that now dismantled it.

The first operation did more than remove concrete: it reopened Ackerstrasse, restoring one of the street connections the Wall had severed. Demolition also proceeded that day along the Mitte–Kreuzberg and Treptow–Neukölln borders, and people gathered on the former border strip for a public celebration.

Not the day the Wall fell

This date is regularly confused with 9 November 1989, when the checkpoints opened. By June 1990 the Wall had long lost its border-control function: 'wall peckers' had chipped away pieces for months, and DDR pioneer units had already cleared selected installations.

What 13 June began was the formal, systematic removal program in and around Berlin — covering the complete border system of front wall, interior obstacles and rear 'hinterland' wall, not just the graffiti-covered face known from West Berlin. The scheduled work in and around Berlin encompassed about 184 kilometers of installations comprising roughly 45,000 wall segments.

Faster than planned

The urgency came from the calendar: German monetary, economic and social union was set for 1 July 1990, when remaining border controls were to disappear. The original plan called for 39 restored street links by that date.

Berlin beat the plan — by 1 July, border surveillance had ended and more than 100 cross-border streets were open. A section at Bernauer Strasse was later preserved, receiving protected monument status on 2 October 1990.

End of 1990: the Web works

By the end of 1990, Tim Berners-Lee had the first Web server and browser running together at CERN — a functional client-server system in which a browser could request and display linked documents over HTTP. CERN's own account says it was running by Christmas; the earliest known HTML document is dated 3 December 1990.

No exact completion day is securely established. The widely repeated 20 December date is not confirmed by the strongest sources — Berners-Lee's own FAQ says only 'end of 1990' — so the milestone is honestly dated to the month, not a day.

The project page on Berners-Lee's NeXT computer

Everything ran on Berners-Lee's NeXT computer in his CERN office. The first HTTP-served page lived at nxoc01.cern.ch/hypertext/WWW/TheProject.html — the famous info.cern.ch address was a later alias, not the machine's original hostname.

The page was the Web explaining itself: it described the WorldWideWeb project and linked to information about hypertext, technical instructions for setting up a Web server, and other servers as they appeared.

A browser that could write

The first browser was also an editor — users could edit pages inside it, a feature usually dropped from simplified accounts. The system grew from Berners-Lee's March 1989 proposal, formalized with Robert Cailliau in November 1990.

Two honest caveats close the story: no untouched 1990 copy of the first page survives online — the familiar reconstruction derives from later material — and access initially remained extremely limited, circulating within CERN through 1991 until the public announcement that August.

6 August 1991: a post on alt.hypertext

On 6 August 1991, Tim Berners-Lee posted information about the WorldWideWeb project to the public Usenet newsgroup alt.hypertext and invited broad collaboration. Alongside the post, he made project files available over the Internet by FTP — extending access beyond CERN and the high-energy-physics community.

The package included Web software and documentation for a system that was already real: a browser, a server and a project website were operating at CERN by August 1991, and the project website had existed there since the end of 1990.

What this date is — and the two it gets confused with

6 August was the public announcement, not the creation: the project page explaining the Web, hypertext and how to create a server had existed at CERN since the end of 1990. Nor was this the invention of the Internet — the Web is an information system built on it, and 'WorldWideWeb' named both the project and, initially, Berners-Lee's browser-editor.

What the day itself delivered

The day's concrete actions were the post, the invitation and the files: software and documentation sat on CERN's FTP site for anyone on the Internet to retrieve, and collaboration was explicitly invited.

Interest began spreading beyond CERN, and the first Web server outside Europe came online at SLAC in California in December 1991.

19 August 1991: the announcement

On 19 August 1991, the eight-member State Committee on the State of Emergency announced that Mikhail Gorbachev was allegedly ill and that it was taking power. The Soviet president had already been isolated a day earlier, on 18 August, at his Foros vacation residence in Crimea — where, his aide recorded, he refused the plotters' demand and insisted on facing the Supreme Soviet.

The timing was no accident: the move pre-empted the planned 20 August signing of a new, decentralized Union Treaty that would have sharply weakened central and Communist Party power.

Armor in Moscow, a parliament defended

Troops and armored vehicles entered Moscow that day. Outside the Russian parliament building — the White House — demonstrators gathered around the vehicles, built barricades and formed human chains.

Russian president Boris Yeltsin made the White House the center of resistance, denouncing the takeover and demanding Gorbachev's restoration while standing on an armored personnel carrier — the celebrated 'tank' of the retellings was not actually a tank. A reported rally the next day drew an estimated 100,000 people; that figure belongs to 20 August, not to a settled count of the first day.

Three days, then collapse

The army did not uniformly refuse the coup — some units initially supported or obeyed it — but decisive units later disobeyed, defected or withdrew. Overnight on 20–21 August, three protesters were killed near a Moscow barricade, shortly before the coup's military backing gave way.

Early on the third day, 21 August, the Defense Ministry ordered troops out of Moscow and the committee disintegrated. Gorbachev returned to Moscow in the early hours of 22 August — restored, and gravely weakened.

24 August 1991: 346 to 1

At an extraordinary session in Kyiv on 24 August 1991, the Ukrainian SSR's Supreme Soviet adopted the Act of Declaration of Independence — 346 votes for, one against — creating an independent state named Ukraine. The Act itself invoked the mortal danger revealed by the Moscow coup that had collapsed three days earlier.

The short text carried immediate legal force: Ukraine's territory was declared indivisible and inviolable, and Ukrainian constitutional and legislative acts became the only law in force within the republic.

More than a proclamation

The same sitting scheduled a republican referendum for 1 December 1991 to confirm the Act.

Parliament also resolved to create a Ukrainian defense ministry and subordinate military formations stationed in Ukraine to the Ukrainian parliament.

Belarus, one day later — by a different route

On 25 August, Belarus took a related but legally distinct step: rather than adopting a new declaration, it gave its July 1990 Declaration of State Sovereignty constitutional force — the act conventionally treated as its independence.

Neither move ended the USSR by itself. Two documents also shouldn't be conflated on the Ukrainian side: the 16 July 1990 sovereignty declaration and this 24 August 1991 Act are different instruments — and the Union's formal end came only in December, after Ukraine's referendum confirmed the Act.

13 March 1998: a manuscript arrives

On 13 March 1998, The Astronomical Journal received a manuscript from the High-Z Supernova Search Team, with Adam Riess as lead author: 'Observational Evidence from Supernovae for an Accelerating Universe and a Cosmological Constant.'

Nothing was published publicly that day — the paper was revised on 6 May, posted to arXiv on 15 May, and appeared in the September 1998 issue. The famous date is the journal's received stamp, not a publication.

Fifty stellar explosions, one reversal

The analysis compared 16 high-redshift Type Ia supernovae — including 10 newly presented at redshifts 0.16 to 0.62 — against 34 nearby ones. The distant supernovae averaged only 10 to 15 percent farther away than a low-density, no-cosmological-constant universe predicted.

That small discrepancy carried the argument: the team's two fitting methods supported present-day acceleration at 2.8 and 3.9 standard deviations, favoring a positive cosmological constant. A team that set out to measure how much the universe's expansion was slowing had found it speeding up.

Priority, carefully stated

The rival Supernova Cosmology Project had publicly presented acceleration evidence in January 1998, while NASA describes the High-Z September paper as the first published evidence — the two independent teams discovered the effect separately, and historical accounts distinguish announcement, preprint and refereed-publication priority.

The recognition was ultimately shared: Science named the accelerating universe its 1998 Breakthrough of the Year, and Saul Perlmutter, Brian Schmidt and Adam Riess shared the 2011 Nobel Prize in Physics. What the 1998 data did not do, by themselves, was identify the physical nature of the dark energy behind the acceleration.

20 November 1998: nine minutes to orbit

On 20 November 1998, a three-stage Russian Proton-K lifted off from Baikonur Cosmodrome in Kazakhstan carrying Zarya, the first component of the International Space Station. Roughly nine minutes later the module was in low Earth orbit — 220.4 by 339.6 kilometers — deploying its antennas and two solar arrays.

With that, the ISS existed: a single functioning module, 41.2 feet long and 13.5 feet across, massing 42,600 pounds (about 19.3 metric tonnes — the often-quoted '21 tons' is a rounding).

A station of one

Zarya was designed to run the embryonic station alone, supplying propulsion, electrical power, communications and orientation control while later modules waited on the ground. Its design descended from Soviet Almaz/TKS-derived orbital modules tested with the Salyut stations.

The funding arrangement defies simple flags: Zarya was Russian-built and Russian-launched, but U.S.-funded through a Boeing subcontract — neither an entirely Russian nor entirely American contribution.

Sixteen days alone

Zarya flew solo for 16 days. Endeavour launched with the Unity node on 4 December; on 6 December, astronaut Nancy Currie used the shuttle's robotic arm to capture Zarya and mate it with Unity, forming the station's first two-module configuration.

The rest came in stages: Zvezda docked in July 2000 and took over many control functions, Zarya settled into a storage and propellant-support role, and continuous human occupation began with Expedition 1 in November 2000.

February 2001: the draft meets peer review

The International Human Genome Sequencing Consortium's paper 'Initial sequencing and analysis of the human genome' appeared in the 15 February 2001 issue of Nature, reporting a freely available draft covering more than 90 percent of the human genome — with 2.91 billion base pairs of euchromatic consensus sequence — and its first genome-wide analysis.

The dates around the milestone are easily conflated: the working draft itself had been assembled and announced in June 2000, UCSC put an assembly online that July, NHGRI announced the published analysis on 12 February 2001, and Nature's webpage also lists a 1 February online date. The paper, not the draft, is what February 2001 contributed.

The unexpectedly low gene estimate

The analysis sharply cut the contemporary estimate of human gene count to approximately 30,000–35,000 — far below many earlier expectations. Behind it stood hundreds of researchers at 20 centers in six countries, and one graduate student's sprint: Jim Kent wrote the roughly 10,000-line GigAssembler program in about four weeks to help assemble the public working draft.

Demand was already enormous before the paper landed — UCSC reports half a trillion bytes of the draft downloaded in its first 24 hours online.

Not the only genome in town

Celera Genomics published a separate draft and analysis in Science at approximately the same time — so this was a milestone shared between two rival efforts, the public consortium's clone-based approach and Celera's whole-genome shotgun.

And 'the human genome' here meant a mosaic reference assembled from anonymous donors, not one person's complete genome — with the underlying public sequence remaining freely accessible without use restrictions.

17 May 2006: the biggest chromosome, last in line

On 17 May 2006, the Wellcome Trust Sanger Institute announced publication in Nature of the sequence and analysis of human chromosome 1 — the largest human chromosome, and the last to receive its dedicated analysis in the Human Genome Project's peer-reviewed series. (The Sanger announcement is dated the 17th; Nature's formal issue date is 18 May.)

The finished sequence ran to 223,569,564 bases — about 8 percent of the genome — at greater than 99.99 percent accuracy, covering 99.4 percent of the chromosome's amenable euchromatic sequence. The work was done principally at Sanger, with the University of Washington Genome Center contributing 13 percent of the finishing.

What 'final chromosome' actually means

The framing needs care: the Human Genome Project had ended in 2003, and its chromosome sequence data had already been released. This paper completed the series of peer-reviewed analyses covering the 22 autosomes plus X and Y — it did not newly create a gapless genome.

'Finished,' 2006-style, still left about 1.3 million base pairs of gaps inside the assembly, and the centromere and heterochromatic repeats — part of an estimated 237.6-million-base total — remained outside it.

What the sequence held

The annotated chromosome became a public resource for studying sequence variation, disease-related genes, recombination, replication and chromosome biology. The analysis identified almost 4,500 single-letter variants that could alter proteins, 90 of them predicted to shorten one.

In conventional chromosome-by-chromosome ordering, the human genome's first base is C — not G.

15 September 2008: three crises in one Monday

In the early hours of 15 September 2008, Lehman Brothers Holdings filed a voluntary Chapter 11 petition in the U.S. Bankruptcy Court for the Southern District of New York, reporting $639 billion in assets, after efforts to find a buyer or sufficient financing failed. It continued operating as a debtor-in-possession.

The same day, Bank of America announced an agreement to acquire Merrill Lynch — feared to be the next major investment bank to fail — in an all-stock transaction initially valued at $50 billion. And American International Group, unable to replace maturing commercial paper, met with Goldman Sachs, JPMorgan and the New York Fed about a secured lending facility.

The night the private rescue failed

Credit-rating downgrades triggered collateral calls AIG could not meet. One private plan discussed that day contemplated a $75 billion secured facility — but the New York Fed-convened syndication effort failed by nightfall, with AIG's stock closing at $4.76.

By the close, the Dow Jones Industrial Average had fallen 504.48 points, or 4.42 percent — the sharpest decline in years.

What September 15 was — and wasn't

The precise sequence matters. The Federal Reserve's AIG facility of up to $85 billion was authorized on 16 September, not the 15th. The Merrill deal was an agreement, not a closing — it completed on 1 January 2009 at a much-diminished market value. And Lehman's failure accelerated a crisis already underway; the Great Recession had officially begun in December 2007.

The aftershock came fast: Lehman losses helped push the Reserve Primary Fund below its $1 share value the next day, intensifying withdrawals from money-market funds — one fund received $10.8 billion in redemption requests on the morning of the bankruptcy itself.

25 August 2012: a boundary crossed in silence

On 25 August 2012, Voyager 1 crossed the heliopause — the boundary where solar plasma gives way to interstellar plasma — and became the first human-made object to enter interstellar space. It was about 121.6 astronomical units from the Sun, roughly 18.5 billion kilometers out.

The spacecraft's instruments recorded an abrupt, lasting transition that would later help fix the crossing date: particles originating inside the Sun's bubble dropped by more than a factor of 1,000, while galactic cosmic rays rose. The permanent transition happened within a window only about 13 hours wide.

Why nobody knew for a year

No one watched it happen. Voyager 1's direct plasma instrument had stopped working in 1980, so the crossing could not be confirmed in real time.

The proof arrived by improvisation: in April 2013, Voyager detected plasma oscillations excited by a March 2012 solar eruption, and the pitch of those vibrations — captured by the plasma-wave instrument — revealed interstellar densities. Working backward, and anchored by the durable particle changes, the team fixed 25 August 2012 as the crossing date. NASA announced it on 12 September 2013, more than a year after the fact.

Interstellar, not gone

Two boundaries get conflated in the retellings. The heliopause is not the termination shock, which Voyager 1 had crossed back in December 2004. And leaving the heliosphere is not leaving the Solar System by every defensible boundary — the probe remains far inside the region associated with the Sun's gravity and the hypothesized Oort Cloud.

Voyager 1 remains the most distant human-made object; as of April 2026 it still operates two science instruments, its plasma-wave subsystem and magnetometer.

3 May 2020: a frozen sample speaks

On 3 May 2020, intensive-care chief Yves Cohen publicly reported that retrospective testing had detected SARS-CoV-2 in a respiratory sample collected from a patient hospitalized near Paris on 27 December 2019 — nearly a month before France's first officially reported cases on 24 January.

The study behind the announcement, from clinicians at the Avicenne and Jean-Verdier hospitals in the northern Paris suburbs, had re-examined stored samples from intensive-care patients admitted with influenza-like illness between 2 December 2019 and 16 January 2020. Of 14 retested samples, exactly one came back positive.

The patient the headlines scrambled

The positive sample belonged to Amirouche Hammar, a 42-year-old Algerian-born man long resident in France, admitted with coughing blood and an otherwise unexplained illness — he had driven himself to hospital at about five in the morning after chest pain and breathing difficulty became unbearable.

The published paper described him as unemployed; some wire-derived reports called him a fishmonger, but later reporting identified his wife, not him, as the fish-stall worker. More significantly, the study reported no recent foreign travel or direct China link — suggesting, but not proving, local transmission in France in late December.

What one sample can and cannot show

The finding pushed France's evidence-based timeline back to at least 27 December, and within two days WHO encouraged other countries to examine stored samples from unexplained late-2019 pneumonia cases.

Its limits are just as important: a single stored sample with no reported viral genome sequence cannot establish France's first patient, the infection's source, or the size of any December transmission chain — and earlier unrecognized cases may have existed.