Recogito

Original song: The Major-General’s song (Gilbert and Sullivan)

YouTube link: https://www.youtube.com/watch?v=nqmWljOnFao

Download: TEI

Semantic annotation’s very easy with Recogito 1
You just upload a text file – how about the Visigothic Code? 2
To tag a person, place, event, or judgement editorial
Just follow the instructions in the ten minute tutorial 3

It’s very well adapted too for annotating images
The “linked box” tool is excellent for marking where each village is 4
The interface for drawing polygons is simply glorious
To use it standalone download the library Annotorious 5

Another thing that you can do is annotate relationships
A pair of clicks and you can mark Iolcos home to Jason’s ship 6
This software simplifies things, like the early Visigothic Code
Semantic annotation’s very easy with Recogito

But Geoannotation is the area it works the best 7
With gazetteers and NER to auto-tag but just suggest 8
It will not overwhelm you like the Emperor Heliogabalus 9
The manual confirmation means you pick the right Tarabulus 10

Now Recogito online is free but you will find a quota set
200 megs is fine for text, but not archival photo sets 11
You could self-host an instance if you’re rich like John Minet Fector 12
But this is just the situation IIIF’s perfect for 13

If your research involves translating Babylonic cuneiform
You can import a manifest from CUNE-IIIF-ORM 14
This software simplifies things, like the early Visigothic Code
Semantic annotation’s very easy with Recogito

In fact if you don’t know what’s meant by “mamelon” or “ravelin” 15
Recogito’s not single player, you can invite Pavel in
And grant him write permission to update the page you’re staring at 16
By entering his username “Wallachian Hospodariat” 17

When you have annotated everything that you are able to,
And you desire to have your annotations in a table too,
Some exports are in beta but you can download a CSV
Or GeoJSON, Markdown, Spacy, RDF, or IOB 18

While my own Linked Data knowledge is alas yet rudimentary
(I don’t do pointy brackets although I was born this century) 19
This software simplifies things, like the early Visigothic Code
Semantic annotation’s very easy with Recogito

Notes

  1. Recogito (https://recogito.pelagios.org/) is an open-source tool that allows people to mark and annotate texts and images, and export their annotations in machine-readable format without needing to learn XML or other languages. Back to text
  2. The Visigothic Code, promulgated by Recceswinth in 654 AD, was the first law code to apply equally to the Roman and Gothic subjects of the Visigothic kingdom. Previously they were governed by separate sets of laws. Back to text
  3. https://recogito.pelagios.org/help/tutorial, also available in languages other than English. Back to text
  4. The “Linked box” tool allows one to highlight a place name on a map, and link it to a marker representing the physical location. See https://recogito.pelagios.org/help/tutorial#linked-box. Back to text
  5. Recogito developer Rainer Simon has extracted the image annotation functionality into a stand-alone Javascript library called Annotorious (https://recogito.github.io/annotorious/). Back to text
  6. Recogito allows annotation of relationships between two words/phrases in a text. E.g. if Jason's ship the Argo was mentioned in the same text as her home port Iolcos, one could highlight both words and add a line between them with the label “Home”. See https://recogito.pelagios.org/help/relations. Back to text
  7. See 5:53 of this introductory video to Recogito https://youtu.be/Qdb8Ds9ayts?t=353 Back to text
  8. Recogito supports Named Entity Recognition (NER) in several languages for automatically flagging names of people, places, etc. For places it further tries to match them against several gazetteers with different temporal or geographic scope. Back to text
  9. In one infamous incident the Roman Emperor Heliogabalus is alleged to have buried guests at a dinner party under flowers dropped from a false ceiling, smothering some of them. See e.g. https://en.wikipedia.org/wiki/The_Roses_of_Heliogabalus. Back to text
  10. Tarabulus (Arabic: طرابلس) is the Arabic name of Tripoli and can refer to the city in either Libya or Lebanon. Back to text
  11. A standard account on the Recogito server has a quota of 200MB. See https://recogito.pelagios.org/help/faq#user-quota This is likely to be insufficient for projects annotating or transcribing text from many high-resolution images. Back to text
  12. Because Recogito is open-source software, one way of getting a larger quota is to run it on your own server. This has ongoing costs both for the server and for maintenance/system administration.
    John Minet Fector was an 19th century English banker. See https://en.wikipedia.org/wiki/John_Minet_Fector Back to text
  13. The International Image Interoperability Framework (IIIF, pronounced triple eye eff) defines a set of standards for sharing images online. In particular it allows high-resolution images to be used online without needing to host your own copy. Back to text
  14. CUNE-IIIF-ORM is a research project with the goal, among other things, of making the cuneiform collection of the Royal Museum of Arts and History available via IIIF. See https://www.kmkg-mrah.be/en/scientific-research/cune-iiif-orm. Back to text
  15. Mamelon: “A rounded eminence or hillock.” (OED mamelon, n. 1). Ravelin: “Fortification. A detached outwork, constructed beyond the main ditch and in front of the curtain, and consisting of two faces forming a salient angle.” (OED ravelin n.1, 1) Back to text
  16. Recogito allows you to grant other users permission to read or collaborate on annotating your documents. See https://recogito.pelagios.org/help/tutorial#invite. Back to text
  17. The governors of the provinces of Wallachia and Moldavia under the Ottoman Porte were known as “Hospodars”, and their jurisdiction as “Hospodariats” (OED, hospodar, n). Back to text
  18. Recogito supports exports of annotations in several formats, including CSV (a spreadsheet), GeoJSON (structured key-value pairs of geographic information), Markdown (a form of rich text), Spacy and IOB (formats for Named Entity Recognition algorithms), and RDF (linked data). Back to text
  19. The tagline for Recogito is “Semantic annotation without the pointy brackets.” The reference to being born this century alludes to the problematic idea that “digital natives” have a greater level of technical skill than their elders. Back to text