EODEM update 6

Work on the Exhibition Object Data Exchange Model, EODEM, continues, and we’ll be hosting two workshops during the CIDOC annual conference, in a few days’ time.

We invite developers and suppliers of museum collections management systems to a vendors’ workshop at 14:00-17:30 EEST on Sunday 22 May 2022. The workshop will be hybrid: both online, and in Tallinn, Estonia. It will introduce the EODEM standard, mark its formal release as a public beta, discuss the challenges likely to be met in its implementation – and hopefully identify some solutions to those challenges.

If you develop or supply collections management software for museums, we hope to see you at the workshop; if you know someone who develops or supplies such systems, please pass this information on to them. Attendance is free of charge, and is independent of attendance at the rest of the conference. To reserve a place, please contact me.

If you’re attending the conference, and would simply like to learn more about EODEM, then please come to the CIDOC Documentation Standards Working Group’s session at 16:00 EEST on Tuesday 24 May; again, this will be hybrid, both online and in Tallinn.

EODEM update 5

Development of the Exhibition Object Data Exchange Model (EODEM) reached an important milestone yesterday with the formal release of version 1.1 of the LIDO data-sharing standard. This is significant because EODEM is defined as a profile of LIDO – that is, an EODEM record comprises a fixed sub-set of LIDO data elements and values. And this in turn means that, if a collections management system vendor implements a LIDO 1.1 importer or exporter, they will have done the bulk of the work required to produce an EODEM importer or exporter.

Continue reading EODEM update 5

EODEM update 3

It’s a couple of months now since my last update on progress with EODEM (the Exhibition Object Data Exchange Model) – so what have we been doing? The short answer is: issued a further draft of the standard; and drawn up a stylesheet which demonstrates how XSLT can be used to transform a heavily-nested EODEM LIDO XML document into a flatter structure (actually CSV, as flat as they come), closer to many that used by many collections management systems.

Continue reading EODEM update 3

EODEM update 2

Since I last wrote about EODEM (the Exhibition Object Data Exchange Model), three months ago, we have been busy. The net result is a new draft of the standard (now available via CIDOC’s new document repository, by the way.) Whilst this is not yet final, it marks a significant evolution of the draft, and we don’t expect to make too many more changes.

Continue reading EODEM update 2

Towards a national collection

And no, I don’t mean TaNC …. I’ve been interested in Collection Trust’s plans in this area for some time; these are brought into sharper relief by their recent announcement that they’ll be turning Culture Grid off soon. This post is as much for my use as anything else: I’m using it to pull together resources related to Collections Trust’s work related to museum data aggregation. I expect I’ll keep it updated as I find more.

Continue reading Towards a national collection

EODEM update 1

I was very pleased to hear EODEM referred to last week,1 during the third Balboa Park Online Collaborative webinar on Dreaming of a New Collections Management System (a really interesting series, well worth watching if you haven’t already, and have the time). But then, as it was Richard Light speaking, and he’s been hard at work recently on building an EODEM importer / exporter for MODES, perhaps it wasn’t so surprising – but this made we realise it’s time to share where we are with the EODEM project.

Continue reading EODEM update 1
  1. EODEM is the CIDOC Documentation Standards Working Group’s project to create an Exhibition Object Data Exchange Model, and see it implemented in collections management systems worldwide – but then, you knew that. []

What is EODEM?

If you follow me on Twitter, you may have noticed that I’ve been very quiet over the last year: I haven’t really been able to face social media since the pandemic started. But when I have tweeted, it’s usually been about conferences – and I’ve been making much of something called EODEM. This post is to explain what EODEM is, and why I think it’s important.

Continue reading What is EODEM?

Collections Trust 2020: systems, interoperability and aggregators

Conferences have not been the same this year: I’ve particularly missed the opportunity to catch up with friends and colleagues, and the random conversations and encounters in queues that compensate for the quality of the coffee one is waiting for. We have been left with the formal proceedings, and I wanted to say something about the papers presented at the (comparatively) recent Collections Trust conference, held online over two half-days on 1 and 2 October 2020.

Continue reading Collections Trust 2020: systems, interoperability and aggregators

The problem with dates

This is prompted by a question that arose during a meeting for ArtUK’s project to pilot the automatic updating of data on their website by speaking directly to the databases of contributing organisations – but it’s not really about that, fascinating and welcome as the project is. At some point in the discussion the question of dates came up, and this reminded me of something that’s been bugging me for some time: have museums been recording their dates properly? It’s a matter I raised a couple of years ago in the users’ email list for the collections management system I use at the National Gallery, Gallery SystemsTMS – where it was met with a resounding silence.

Let me explain. Most museum collections management systems allow users to record dates in two ways: as free text (sometimes called a ‘display date’), which could be as vague as ‘early fifteenth century’ or as specific as ‘2 February 1976’; and as a properly machine-readable date (or dates), following a standard format. The format that they use is almost always that mandated by ISO 8601, Data elements and interchange formats – Information interchange – Representation of dates and times.

An engraving by Hans Sadeler of a Gregorian perpetual calendar
Johann Sadeler I (1550-1600), Kalendrivm Perpetvvm (Gregorian perpetual calendar, with the Four Seasons), 1595, engraving on paper, 317 × 242 mm; Amsterdam, Rijksmuseum, RP-P-OB-7498. Photo: Rijksmuseum CC0 Public domain.

The formatting element of ISO 8601 is quite simple; of the variants it offers, the one most often used in museum software is the basic YYYY-MM-DD (which can be reduced down to as little as YYYY for less precise dates). The problem is to do with calendars. ISO 8601 specifies that dates should be recorded using the Gregorian calendar – the calendar in use in much of the world today, first introduced in some countries on 15 October 1582 as a result of reforms instigated by Pope Gregory XIII, and designed to reduce the disjunction between astronomical time and calendar time that had crept into the previous Julian calendar because of the way that leap years were calculated: the Julian calendar was by then 10 days adrift of astronomical time. In other words, when the Gregorian calendar was first adopted, 4 October 1582 was, in some countries, immediately followed by 15 October 1582. (The days of the week, however, ran on in their normal sequence.) The calendar was adopted across the world gradually: it only came into use in the British Isles in 1752, when 2 September 1752 was followed by Thursday 14 September 1752 (after me, now: ‘Give us our eleven days!’)

Detail from William Hogarth's 'An Election Entrtainment', showing a bully and a placard reading 'Give us our Eleven Days'
Detail from William Hogarth (1697–1764), An Election Entertainment, Plate I, 1755-8, engraving on paper; New Haven, Yale Center for British Art, Paul Mellon Collection, B1981.25.1448. Photo: Yale Center for British Art, open access.

But what about dates before 15 October 1582? According to ISO 8601, these can only be recorded ‘by mutual agreement of the partners exchanging information’; they should be recorded using the proleptic Gregorian calendar, which simply extends the Gregorian calendar back in time before its creation. This means that the precise dates of events that took place in the past should be shifted accordingly when entered into museum databases, if they are actually to adhere to the standard they profess to use. For example, Titian’s Bacchus and Ariadne arrived at its destination, Ferrara, on the Julian calendar date of 20 January 1523 (or thereabouts) – or 30 January in the proleptic Gregorian calendar.

Titian, Bacchus and Ariadne
Titian (active about 1506; died 1576), Bacchus and Ariadne, 1520-3, oil on canvas, 176.5 × 191 cm; London, The National Gallery, NG35, bought, 1826. Photo © The National Gallery, London Creative Commons License This work is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License.

I’d like to ask anyone reading this who works in museum documentation: are you really recording pre-Gregorian dates in ISO 8601-formatted fields using the proleptic Gregorian calendar – or are you just using the Julian calendar, and formatting your dates to look like ISO ones?

I can understand why museums might not be implementing ISO 8601 properly. For one thing, it means that the proleptic Gregorian date needs to be calculated every time you enter a Julian date. For another, it leads to problems when we consider anniversaries. The Wikipedia article on the Gregorian calendar discusses the supposed coincidence of the deaths of both Cervantes and Shakespeare on 23 April 1616 – although, as Spain at that time used the Gregorian calendar, and England the Julian, Cervantes actually died ten days before Shakespeare, the interval by which the two calendars diverged in 1616. But in 2020, we will mark the anniversary of Shakespeare’s death on its calendar date, 23 April, and not 404 years (of 365/366 days) after it actually took place, on 3 May. In other words, our anniversaries are out of step with our calendars if the events being commemorated occurred under the Julian calendar. And museums, as memory institutions, do like to mark anniversaries – even more so when there are social media accounts to feed with regular content. To return to the example of the Titian painting I mentioned earlier: when should we tweet the anniversary of its arrival in Ferrara – on 20 January (Julian date) or 30 January (ISO 8601 date)?

But if we stick with Julian calendar dates entered using the ISO YYYY-MM-DD format, we’re setting ourselves up for further problems when we come to calendars where the new year doesn’t fall on 1 January. Many European states – including England and, before 1600, Scotland – marked the new year on Lady Day (the Feast of the Annunciation, 25 March). So in England, for example, 31 December 1602 was followed immediately by 1 January 1602, and 24 March 1602 by 25 March 1603. If we were to write these dates using the ISO format, and list them in chronological order, we would end up with the sequence:

  • 1602-12-31
  • 1602-01-01
  • 1602-03-24
  • 1603-03-25

Why does this matter? First of all, there’s the problem of putting chronological dates in order: YYYY-MM-DD dates should sort nicely, as they’re arranged in order from largest to smallest component. But they don’t, if the year doesn’t begin neatly at YYYY-01-01. Secondly, there’s ample scope for confusion: James VI of Scotland succeeded to the throne of England as James I, bringing about the Union of the Crowns, upon the death of his cousin Elizabeth I late on 24 March 1603 by the Scots calendar – but 24 March 1602 by the English. (To resolve the ambiguity, dates are sometimes written using the conventions ’24 March 1602/3′, or ’24 March 1602 Old Style (O.S.)’ / ’24 March 1603 New Style (N.S.)’.) In the United Kingdom, all this – the date of the new year and the divergence between the Julian and Gregorian calendars – was resolved by the Calendar (New Style) Act 1750.

Attributed to Paul van Somer, James VI and I
Attributed to Paul van Somer (circa 1576–1621), James VI of Scotland and I of England, circa 1618, oil on canvas; New Haven, Yale Center for British Art, Paul Mellon Fund, B1982.18. Photo: Yale Center for British Art, open access.

And – if you’re not already banging your head on your desk in despair – ISO 8601 also creates a discrepancy with dates BC. The Julian calendar, as conventionally reckoned (remember, the Romans didn’t know to mark the birth of Christ in the year in which it occurred) included no year 0: 1 BC was followed immediately by AD 1. This is the convention followed in most historical texts. But the proleptic Gregorian calendar as used by ISO 8601 includes a year 0, with dates before that given negative numbers, and dates after (dates AD) given positive ones. So Julian 1 BC is in fact proleptic Gregorian 0, Julian 2 BC is proleptic Gregorian -1, etc.

So I can understand why museums might not actually be adhering to the standard that they profess to be using when they record dates. And with that happy thought I’ll sign off. For those of you who like to do things the old-fashioned way (and have some Italian), I find Adriano Cappelli, Cronologia, cronografia e calendario perpetuo: dal principio dell’era cristiana ai nostri giorni, 7th edn, Manuali Hoepli (Milan: Ulrico Hoepli, 1988, and reissued subsequently), ISBN 88-203-2502-0, a useful compendium of calendar-related information. My copy came with a calendar conversion program on 3.5″ floppy disk (remember them?), which I can’t get to run on Windows 10; fortunately, there’s now a superfluity of calendar converters online – John Walker’s, available here, has the added advantage of downloadable code that you can install locally should you wish.

Now, back to working out how to record fifteenth-century dates from Florence (using the Julian calendar and starting the New Year on Lady Day) in my collections management system.


  • 17 February 2020: erroneous references to ISO 8901 corrected to ISO 8601 (thanks to @gndnet).
  • 17 February 2020: erroneous reference to a 3¼″ floppy disk corrected to 3.5″ (thanks to Michael Comiskey).
  • 18 February 2020: there’s a more extensive discussion of these issues in Miranda Lewis, Arno Bosse, Howard Hotson, Thomas Wallnig & Dirk van Miert, ‘II. Standards: Dimensions of Data. 3. Time’ in Howard Hotson & Thomas Wallnig (eds.), Reassembling the Republic of Letters in the Digital Age: Standards, Systems, Scholarship (Göttingen: Göttingen University Press, 2019) DOI:10.17875/gup2019-1146, pp. 97-117 (thanks to Arno Bosse).