Participation

count

all entrants
only those with a work

chart

matrix · countries × editions
line chart · one line per country

The Project

get all data

Coverage — this chart is knowingly incomplete, and the strip above each edition square says on what authority each edition is counted. For the twenty-two editions of 1973–1994 the bailiff's record of deposit — the original is at the Bibliothèque nationale de France — has been transcribed into the database in full: those counts are attested entry by entry, and they are exact. 1993 and 1994 joined that group in August 2026, the last two of the run. From 1996 on the count rests mainly on a recapitulative list of entrants kept by the IMEB itself — a second-hand document, and a fallible one: it omits 269 deposits that the bailiff’s records attest, and every one of those people is listed elsewhere in the same document under the same name — so a blank in that list proves nothing at all. Where a first-hand record exists, that record wins and the list is never allowed to contradict it. Against the minutes themselves, the database holds roughly three quarters of the entrants they list: 348 of 500 for 1996, 474 of 633 for 2005, 364 of 506 for 2008. A selection is flagged complete data up to 1994 and incomplete data from 1996 on, next to the page title. No competition was held in 1995 (36 editions in all), so the timeline skips that year and the lines join 1994 directly to 1996. Two low editions are not gaps in the record: 2004 (137), because the competition itself ran at a third of its size — the first three categories, the Trivium A, were not opened, and the IMEB’s own chronicle records 154 works from only 132 composers, fewer than the database holds; and 1992 (112), because that year the competition was open for one degree only — the Residence — and the bailiff’s record says so twice: « only the section Residence is open to candidates », and « 102 compositions were submitted ». The 20th-anniversary retrospective — the Euphonies d’Or, honouring the finest works of 1973–1991 — was held the same year, but it was a separate event: this page long said 1992 was that retrospective instead of a competition, and the bailiff’s record, read in August 2026, says otherwise. The 33 people whose only trace in 1992 is a work replayed at the retrospective are counted here as a presence, not a candidacy.

Most of the people counted here have no work in the collection at all. The two figures beside the page title say how many of each: composers with a work in the IMEB capsules, and everyone the database records as an entrant. The gap between them is made of candidacies and nothing else — a name read off an entry list, with no music attached. That is why most names in the list below are withheld, and why the strip above each square matters more than the height of the line.

Timeline (top strip)

one square per edition of the competition — the first square, all, charts every edition from 1973 to 2009
transcribed in full from the bailiff's record — every entry is attested (1973–1994)
counted from the entrants list and from archived works — second-hand, and short of the minutes (1996–2009)
within that last group, six editions have minutes that were transcribed but never imported (1996, 1999, 2005–2008), and eight have no transcription at all (1997, 1998, 2000–2004, 2009). That difference says what is left to do, not what the count is worth — so it is written here and not in the strip
selected edition(s)
editions inside the selected time span
span toggle on: pick two years to chart the period between them — with only the first one picked the chart waits and says so · off: pick a single year to get a bar chart of that edition. In both of those cases the chart switch greys out: a single edition has neither a matrix nor a line to show, and a half-made span has nothing at all

Count — what the three charts count

all entrants — everyone the database records as having entered that edition, whether or not any music of theirs is in the collection. This is what the charts have always counted, and it is the honest measure of the competition
only those with a work — the same people, minus those the database knows by a candidacy alone. This is the measure of the collection, and it is a good deal smaller: of the 4 262 people the database records as entrants, 2 773 have no archived work at all — nearly two in three, and the share grows with every edition transcribed, because a bailiff’s record names entrants, not works
the switch acts on all three charts at once, and on the list of composers below them — it is the same question put to the same data, so the figures change together. That last part had been promised here and was not done until 2026-08-11: the charts dropped the people known by a candidacy alone, and the list went on naming them just underneath. With the switch on, the list holds only composers with an archived work — so no name in it is withheld any more, and every line is clickable: a name published with a work is a published name. The c/t counts beside each country do not move: they state a fact about that country across all editions, not about the current selection

Countries — whose map is this

81 countries are charted. The country shown is the one a person is attached to today, not the state named on the envelope they posted in — so this is the map of the present laid over a corpus that runs from 1973 to 2009
the consequence is worth knowing before reading any line: Czechoslovakia, the German Democratic Republic and the Soviet Union never appear, though the database holds 62, 22 and 9 addresses that name them on the deposit itself. The 29 Czechoslovak entrants of 1973–1991 are counted under Czech Republic (13) and Slovakia (16); the 11 East German entrants of 1977–1990 under Germany; the 7 Soviet entrants of 1989–1991 under Russia. Yugoslavia is the one vanished state still carried by a few people (3); the twelve others whose address says Yugoslavia are spread over its five successor states
this is a decision, not an oversight. The address keeps the wording of the document — that is what a bailiff’s record is for — while a person outlives the state they lived in and is looked up under the name that state has now. Both answers are true and they are not the same; a chart drawn on addresses instead of people would show the other one
Unknown is a country row like any other (25 people): the source names no state, and a blank is not a zero here either

Matrix — the default chart

one row per country, one column per edition; the colour of a cell counts the entrants recorded for that country in that edition — on the authority the strip above the column names
an empty cell is not a zero: that country entered nothing that year. The grid keeps the hole, so a country's first entry, its single-edition appearances and the 1995 gap are read directly — none of which a line could show
the scale is square-root, as it was on the old vertical axis: it now runs through the colour, so one entrant stays visible next to a hundred. The key sits top right
rows: total · first entry · A–Z reorders the matrix. Three orders, three questions: how large, how early, where is a given country
click a country's name to isolate it: it takes a colour, and its share is stacked inside the band above. The other rows fade back but keep their place, so isolating a second and a third country is the same gesture again, and the ones you chose are read among the others rather than instead of them. Eight at most carry a colour; beyond that they stay in the grey mass, because a ninth band is one nobody can read
click a cell to list, below, all the composers of that country, and click the same cell again to drop it. Clicking a column that is crossed out (1995), or the key, or the band, does nothing. reset all clears everything at once

Band on top — entrants per edition

the outline is the total for each edition, every country together — the one figure sixty lines never allowed anyone to read
the scale here is linear, not square-root: an area that is stacked must sum, and √a + √b is not √(a+b). The square-root of this page lives in the colour of the cells, where it sums nothing
the right-hand column totals each country over the period on screen — its heading says this period when a span is selected, total when the whole run is. The c/t figure to the left of the grid is a different reckoning: c composers with archived works out of t entrants from that country, across all editions, whatever the period shown
1994 joins 1996 directly, in the band as on the lines: there was no 1995 edition, and a fall to zero would read as an empty one

Bar chart — one edition on its own

turn span off and pick a single year: one bar per country, and the bar is read in two. Its full height counts the entrants recorded for that country — the same quantity as a point on a line or the colour of a cell. Its emerald foot counts how many of them have a work in the collection
the gap between the two is the subject of this database: of the people the index counts, about half left nothing but a candidacy — a name read off an entry list, with no music attached. Two countries can send the same number of entrants and leave very different amounts of music behind, and it is the height of the emerald foot that says so
hover a column for the exact figures; click it to list that country's composers below, exactly as clicking a line or a cell does — same orange bar, same list, same works panel. Click it again to clear

Line chart — the other view

each line is a country; the vertical axis counts the entrants recorded for that edition, and the scale is square-root (gridlines at 1, 2, 5, 10, 20, 50, 100…)
this chart carries no provenance strip of its own: the strip on the edition squares sits immediately above it and is read with it
it is kept because it reads a single trajectory better than any matrix once a handful of countries are isolated — and because two views answering the same click on the same data is what makes a regression something anyone can check rather than something to be argued about
in the chart legend, c/t has the same meaning; the square next to each country shows or hides its line. The legend now lays itself out from the width of the canvas and the number of countries, and every country charted is named in it — until 2026-08-11 the last six alphabetically (Unknown, Uruguay, Venezuela, Vietnam, Yugoslavia and Zimbabwe, 58 people between them) were painted just past the right edge: invisible, and their square unclickable. A canvas has no edge that protests, so nothing said so
click a point on a line to list, below, all the composers of that country

Composers list

Name (n) n works archived in the IMEB capsules — click the name to list them
Name took part in the selected edition
J****** D***** recorded as an entrant, but no archived work — the name is withheld and the entry is not clickable: what is recorded here is an application to the competition, not a published work
the orange bar sums up the current selection — click it to switch between all composers and those of the selected edition only
the violet box of works and the ISNI record are both capped at about a third of the window height and scroll inside themselves, their headers staying put: a composer with a dozen archived works no longer pushes the rest of the column off the screen
clicking a name opens a panel on the right: the composer's name, then the number of their archived works — click that count to unfold the list. For a composer who has an ISNI, a third box appears between the two, headed by that identifier: click the header to unfold the international identity record. Nothing is requested until you unfold it

no selection