Chapter 15. Linking to external resources
Version 3.0 (final publication expected in December 2019)
by Tarrin Wills
This Chapter defines how to refer to phenomena in texts that have some kind of existence outside the text — both electronically naming and locating the thing referenced. These include the following phenomena and their associated TEI elements:
- people <name> <handDesc> <person>;
- places <name> <placeName> <place>;
- events <name> <event>;
- texts <text> <name> <msItem>;
- parts and versions of texts <div>, <lg>;
- words and their corresponding headwords (lemmas) <w>;
- scholarly books and articles <bibl>;
- manuscripts <msDesc> <msIdentifier> <name>
(The Names and Dates module gives elements for <person>, <place> and <event>, but these are generally unnecessary for the purposes of projects using this handbook: use <name> with the @type attribute.)
In this chapter an external phenomenon falling under one of these categories (person, word, place, etc.) is referred to simply as the thing and can be distinguished from the words or features in a particular text which in some way reference or mention the phenomenon.
There are various reasons to link to external resources. The first is simply so that the end-user of the encoded text can find more information about the thing that appears in the text: biographical information about an author or character, for example. This type of linking is very much of the traditional Web variety.
The Semantic Web introduces principles that can greatly enhance resources by promoting two-way, meaningful linking of resources. For our purposes, that means that any element that contains information referencing external resources has a meaningful link, implied in the naming of the element and/or in the form of the reference, to the relevant authority that defines or curates the thing to which it relates. Unlike linking using, for example, the <ref> element (or <a> element in HTML), these links belong in the attribute part of an element, defining an implied equivalence between the thing encoded by the element and the external authority referenced.
There are many advantages to this approach: a standardised referencing system means that across the texts in a repository, for example, one can find all the examples of the same thing: all the references to the author and historical figure Snorri Sturluson; all the texts that represent versions of Egils saga (or a particular chapter in Egils saga); all the references to the places Helgafell or Lejre; and so on.
Additionally, if the application or repository includes metadata such as dates or geolocations (possibly generated from the linked resources) to generate a timeline or map, there is a link between the information in the application or repository and the authority that supplied the metadata. This means that the metadata can be updated if the authority updates its information.
This Chapter does not define how these links might be processed, but ideally the repository would generate authority files that incorporate the metadata generated from the linked resources, and provide a central index to the links in the corpus.
Note that the methods outlined in this chapter involve a large amount of searching external resources. To assist in this process, the author has developed both a human user interface and API to assist in some of these searches and in inserting the generated links into the XML file. The location of these tools will be published on the main Menota site.
15.2 Linking methods
15.2.1 Attributes used for linking
The methods defined here utilise the TEI-defined attributes for linking to external resources: @ref, @key for elements for which they are valid (e.g. <name>), and the additional Menota-defined equivalents for other elements, @me:ref, @me:key.
The content of the two differs: @key contains information defined by the project and can only be resolved (i.e. the referent found) within the project; @ref contains universal identifiers (URIs) which can be resolved according to the general principles of URIs.
15.2.2 Internally defined references: @key and @me:key
The TEI @key attribute is intended to contain an identifier such as a database key which is unique for the particular type of resource represented by the element which contains it. For example, a <name> might contain a number that represents the database key for the person or place the element refers to.
The advantage of using a @key over a @ref is that the notation can be much simpler and easier to process, with less chance of alternative references to the same thing. But using @ref gives universal references without the need for an internal authority file (although use of one is nevertheless recommended).
15.2.3 Universal references: URIs, URLs and URNs
Uniform Resource Identifier (URI) is a standard for identifying things or resources universally. There are two main subsets: URLs (Uniform Resource Locators), for locating resources on the network (i.e. the Internet), and URNs (Uniform Resource Names), for consistent naming of resources. These standards are well documented — this section is for discussing particular issues for Menota texts and applications.
This chapter uses both methods — names and locations — for identifying resources identifying and associated with the phenomena listed above. Naming methods provide a simpler and more consistent method for identifying things in the encoded text, but locating methods provide a means for finding more information about the thing, including metadata that can be processed by the application — for example, by mapping places in a text.
In the URI string certain characters are reserved for identifying parts of the string, including colon, slash, dot/period and so on. Other characters can be encoded using % (percent) and a two digit hexadecimal representation of the character. Because this chapter normally only describes variation in the last part of the URI, where there is normally no ambiguity with respect to periods and slashes, in most cases the only characters needing encoding will be the following:
URLs are normally found in the web browser location bar and can be copied and pasted into an XML file for use in relevant attributes. Normally the required characters are automatically encoded by the application and/or operating system using the hexadecimal notation above when the copy and paste is performed, and no further intervention is therefore needed from the encoder. There are some exceptions, however.
The standard delimiter of key/value pairs in the query string section of URLs is
(ampersand), which is normally shown in its unencoded form in a web browser location bar, and
therefore where most users are likely to find the URLs documented in this section. However,
using ampersand on its own is not valid XML (& should be used) but many
interface languages such as PHP do not recognise this notation if it is passed unparsed to a
web application. You will need to convert
& in URIs to & but also
ensure that the application converts them back for linking for end-users (web browsers
normally do this automatically).
Some of the URLs for external resources are longer in the browser than what is needed to resolve the link, because they contain additional information that is used, for example, to identify particular browser sessions. Some of the recommendations below include instructions on how to modify URLs in order to reduce ambiguity and simplify the attribute value, without affecting the ability to link to the resource.
220.127.116.11 URNs: official URNs and Menota names
The URN standard is a less-used notation that forms part of the URI standard. It does not provide a way of locating resources but rather provides a means of identifying them consistently and universally. The current standard for URNs is limited to a few types of resources that might be of use to a Menota user, including book identifiers (ISBNs, e.g. urn:isbn:8270992615) and journal identifiers (ISSNs, e.g. urn:issn:0305-9219).
Earlier versions of the standard included experimental namespaces that could have been used for Menota projects, but their use has now been deprecated.
Instead, an unofficial notation should be used, using menota: as the protocol section of the URI, followed by the type of thing/resource, followed by the authority (abbreviation) that curates it, followed by the standardised, unique identifier for the authority. The syntax is as follows:
menota:[type of resource]:[authority abbreviation]:[unique name or identifier]
Examples of Menota-defined names:
|menota:person:sbi:Snorri%20Sturluson||person||Name in the form found in the Scandinavian Biographical Index.|
|menota:place:ds:167334||place||Identifier ‘ID’ from Danmarks Stednavne data.|
|menota:lemma:onp:23709||word||Identifier from Ordbog over det norrøne prosasprog.|
These are discussed in detail below. Where URNs are used, it is recommended that a URL be included in addition. This allows the thing to be both named and located.
15.3 Principles of naming
The resources linked by these methods should:
- Give a stable URL or URN which is unambiguous (i.e. one URI/key per unique thing, without duplicates) and unlikely to change in a way that cannot be converted later (i.e. base host name may change but could be converted to a new domain);
- Contain sufficient information to identify and verify the thing referenced (i.e. following the link or looking up the resource will give a resource that names and describes the thing so that the reference can be understood);
- Contain additional information (metadata) in the linked resource to determine the date, location and relationships of the thing referenced, depending on the type of thing;
- Ideally, it should be curated by a professional body that has sole or primary responsibility for the thing in question (e.g. dictionaries, archives, repositories, gazetteers), but otherwise it should be a well-established resource which satisfies the other requirements.
Open resources (e.g. Wikipedia, Open Street Map) allow users to enter and edit information, meaning that a gazetteer, for example, can be built alongside the the encoded text if no other authority is available. But there are a few things to note if you are to link to open resources, and especially if you are contributing data yourself:
- URLs for open resources are generally quite stable for longer, older, more detailed entries, but newer and shorter contributions may undergo subsequent editing in a way that affects the reference, for example, because the article is split into different articles, or the title changes.
- Some resources, notably Wikipedia, contain often differing information and detail for the same item, depending on the language of the entry. This means that there may be more than one entry for the same thing. Wikipedia and related resources normally automatically use the Wikidata service to provide a way of referencing the thing itself rather than the various pages which document it. If you find a relevant page on, e.g. Helgafell, you should look up Wikidata (wikidata.org) and use the URI given there (https://www.wikidata.org/wiki/Q972971 — which also gives a geolocation, for example).
- Sometimes it can be difficult to find the desired reference in Wikidata, even in cases where a Wikipedia page has been identified. In such cases the API can help, e.g. https://www.wikidata.org/w/api.php?action=wbgetentities&sites=enwiki& titles=Battle_of_Stiklestad&normalize=1 (JSON output). Here sites=enwiki identifies that this corresponds to a page in the English language version of Wikipedia (the first part of the Wikipedia hostname in the URL (en.wikipedia.org) and titles=Battle_of_Stiklestad uses the exact format of the page in the Wikipedia URL (https://en.wikipedia.org/wiki/Battle_of_Stiklestad). The resulting identifier (Q83078) can be used to construct the Wikidata URI: https://www.wikidata.org/wiki/Q83078, and should be the same regardless of the language of the entry.
- Where Wikidata is not available, or in other resources (such as Handrit.is) which contain different language versions of the same item, you should link to the article according to the following language priority: English, Swedish, Danish, Norwegian (bokmål), Norwegian (nynorsk), Icelandic — that is, according to the size of the language community represented by each language, restricted to the languages of the Menota project. In other words, if there is an article on the thing in English, that should be used, regardless of whether the English article is longer or more accurate than the others. (Different language versions for the entry are linked by Wikipedia and can be found in the left-hand side on the desktop version of Wikipedia.)
- In some interfaces, other uses may delete contributions easily and altogether (e.g. Geonames.org). If you use such resources, you should save the metadata separately from the resource in case this happens (e.g. dates, coordinates, discursive text). This applies particularly if you contribute data yourself — you should save all your data separately in case someone else deletes it.
You can use web.archive.org to archive and reference a particular version of an open website such as Wikipedia (either use the link at the site or open the URL preceded by 'http://web.archive.org/save/'). This should be used if you have contributed information yourself to the resource, or you have reason to think that the metadata or information may be changed at a later point in a way problematic to your project. Using archive.org’s service to archive the page ensures that the page can always be accessed and may be used, depending on the resource, to discover newer versions of the page and access udpated metadata. Using this method, however, does not guarantee that the user or application can recover newer versions of the referenced page, or metadata not directly associated with the linked page; nor does it provide a uniform way of referencing a particular thing (because the date is included in the archive URL). So links to this site should be included in addition to the formats supplied below: they have little use as a naming scheme, only as a locating scheme.
15.4 Types of reference
15.4.1 Bibliographic items
Bibliographic items are normally marked up with the <bibl> element, which should contain a @me:ref value to identify the work
Bibliographic items with a DOI, ISBN, or journals with an ISSN, should be referenced using the official URN or URI in the first instance:
- doi:10.1093/llc/fqt045 (Wills, ‘Relational Data Modelling of Textual Corpora’ 2013)
- urn:isbn:8270992615 (Krömmelbein, T. Dritte grammatische Abhandlung 1998)
- urn:issn:03059219 (Saga-Book of the Viking Society)
But you should also include the URL to locate the item, e.g.
<bibl me:ref="doi:10.1093/llc/fqt045 https://doi.org/10.1093/llc/fqt045"> (Wills, ‘Relational Data Modelling of Textual Corpora’ 2013)</bibl>
In other words, the name (URN) comes first, followed by the location (URL), separated by spaces.
Where individual page(s) are referenced, the page number (Arabic or Roman) can be included in the fragment section (at the end, following ‘#’), e.g. urn:isbn:8270992615#page=155 — this format is taken from Adobe’s PDF referencing format and will not be processed with other file formats.
Items without these official URNs (e.g. because they predate the ISBN system) should be found on stable bibliographic resources, in the following order of priority:
- ONP (https://onp.ku.dk/b, about 4500 items) select the item and use the permalink (e.g. hhttps://onp.ku.dk/b1608.
- Skaldic Project (http://skaldic.abdn.ac.uk/m.php?p=bibliographyskp, about 3700 items) — find the individual item and use the URL (e.g. http://skaldic.abdn.ac.uk/m.php?p=bibl&i=1157). You should also use the Menota-defined URN based on this identifier: menota:bibl:skp:1157).
- Handrit.is (https://handrit.is/en/manuscript/browse/bibliography) — browse and use the URL (e.g. https://handrit.is/en/bibliography/view/Smastykker11). You should also use the Menota-defined URN based on this identifier: menota:bibl:handrit:Smastykker11).
- Worldcat (e.g. http://www.worldcat.org/oclc/925766862, http://www.worldcat.org/oclc/13481601 — remove all parts of the path part of the URL except 'oclc/[number]'). Worldcat may include multiple entries for the same work. To prioritise, follow the link ‘view all editions/formats’ and choose the first item in the list that is in fact the same item. Then select the ‘permalink’ button to copy the URL.
- Haithi Trust (https://catalog.hathitrust.org/Record/008550198)
More than one URL can be included, but include in order of priority, e.g.:
Translated <q>die notwendige Verkürzung von Wörten</q> in <bibl me:ref="urn:isbn:8270992615#page=155 http://skaldic.abdn.ac.uk/m.php?p=bibl&i=1168 http://www.worldcat.org/oclc/925766862">Krömmelbein’s (1998, 155) edition of the treatise</bibl>.
In this instance, the name (ISBN URN) is given first, followed by the first resource that contains the item, followed by the first Worldcat item (which may have more information, but may be more ambiguous).
Avoid Google Books, as there are often multiple records for the same item and no easy way to distinguish or link between them.
15.4.2 Word lemmas
18.104.22.168 Linking of lemmas
Word lemmas should be referenced as dictionary headwords, using additional information to distinguish between homographs (e.g. database keys or numbering of homographs, depending on the wordlist used). Ideally the lemma referenced would belong to a universal wordlist of all languages covered by Menota, but no such list is available, at least in a form suitable for unambiguous linking. The referencing system will therefore be dependent on the particular language of the text.
The reference should be included in the @me:ref attribute as part of the <w> element. Note that in this type of resource, the reference is not in fact to the element itself (the word in a particular text), but to the headword to which it belongs. There may be instances where the word in question may require a link to an equivalent word (e.g. to facilitate collation and automatic textual variants). In such cases, the link to the actual word should come first in the attribute and the link to the lemma later.
See also ch. 10 above.
22.214.171.124 West Norse
West Norse words should be referenced using the Ordbog over det norrøne prosasprog (ONP) at onp.ku.dk. To find the URL for a particular word, look it up under ‘Word’ / ‘Ord’ and click on the word. The URL can be used to reference the lemma.
e.g. forskáli https://onp.ku.dk/onp/onp.php?o22938 — in this item the identifier (database key) is the integer at the end, namely ‘22938’. The permalink (click on the link icon - it is simply the URL with ‘onp/onp.php?’ part removed) can be used to reference the lemma.
<w lemma="forskáli" me:ref="https://onp.ku.dk/o22938">forskali</w>
ONP’s wordlist includes a large number of words not covered by its corpus, including poetic and younger words. These may be useful and can be found and linked using the unofficial interface: skaldic.abdn.ac.uk/m.php?p=onpadv (search for ‘all words’).
126.96.36.199 Old Danish
The primary lexicographic resource for Old Danish is the Gammeldansk Ordbog (gammeldanskordbog.dk), which should be used to find headwords. The interface uses a search mechanism that will give the resulting word if there is no ambiguity in the search (e.g. http://gammeldanskordbog.dk/ordbog?query=forskatte). However, this is not a direct link to the word, so you should click on the word in the right-hand column to find the direct link (here http://gammeldanskordbog.dk/ordbog?aselect=forskatte&query=forskatte). The query= section of the URL is used to fill out the query box on the page, but may otherwise be ambiguous. Therefore replace this part of the url with query=%25 (Note that the query ‘%’ has been converted to its ‘escaped’ equivalent ‘%25’ here). In other words, the resulting URL should be: http://gammeldanskordbog.dk/ordbog?aselect=forskatte&query=%25.
Words with homographs in GDO are distinguished by a number, which should be included in the URL. E.g. skin2 (for mod. Dan. ‘skin’ rather than ‘skind’) http://gammeldanskordbog.dk/ordbog?aselect=skin,2&query=%25.
An internal URN should look like: menota:lemma:gdo:forskatte, menota:lemma:gdo:skin,2.
188.8.131.52 Old Swedish
The authoritative source for the Old Swedish lexicon, the Fornsvensk lexikalisk databas (spraakbanken.gu.se/fsvldb), does not have an interface for finding database keys. Unambiguous headwords should use the Menota URN format, e.g. menota:lemma:fsld:forskilia. To distinguish between homographs, word class and other morphosyntactic information should be included using the @me:msa attribute for the word.
15.4.3 Literary works and their parts
184.108.40.206 Pre-1540 West Norse prose works
ONP’s database (https://onp.ku.dk/v gives a comprehensive and unambiguous overview of this group of prose works. E.g. Egils saga Skallagrímssonar: use ONP’s permalink URL (https://onp.ku.dk/v109) or key (menota:work:onp:109 — the ‘109’ is the key, taken from the previous URL).
Sometimes it will be more useful to use a particular version of the text listed under the sigla above, particularly for charters and other texts where the sub-categorisation in ONP is in fact a distinct document rather than an alternative version. To find the relevant URI, select the relevant version/text from the list under the work and use the permalink there. E.g. the theta fragment of Egils saga (Eg162θ(2001)) can be referenced by the URL: https://onp.ku.dk/r11195 or the name: menota:workversion:onp:11195.
220.127.116.11 Post-1540 West Norse prose works
(E.g. Hrómundar saga): use the ‘uniform’ titles from handrit.is, but note that the Handrit titles are not standardised. You should choose the most commonly used form, e.g. https://handrit.is/en/manuscript/list/uniform?v=Hrómundar+saga+Greipssonar (name format: menota:work:handrit:Hrómundar+saga+Greipssonar (Not https://handrit.is/en/manuscript/list/uniform?v=Hrómundar+saga+Gripssonar or https://handrit.is/en/manuscript/list/uniform?v=Hrómunar+saga+Greipssonar [sic]).
See also ch. 18.104.22.168 below.
22.214.171.124 Old Danish texts
GDO has a comprehensive collection of references to the Old Danish corpus, but it is quite difficult to find the database keys. You need to find a word from the text you need to reference from the citation slips (seddelsamlingen) and then click on the work’s siglum. The most reliable way is to choose an uncommon word from the text, search for that in the dictionary, then browse the slips until you find the text you require. When you click on the siglum, you will get a URL like: http://gammeldanskseddelsamling.dk/forside/kilde?qid=463079&id=251&portlet_word_or_source=source&portlet_alphabetical=1 but only the id= part of the query string is required to look up the reference (http://gammeldanskseddelsamling.dk/forside/kilde?id=251 — DiplDan). The integer following id= is the database code which can be used in the name: menota:work:gdo:251.
126.96.36.199 Chapters and sections of prose works
Different versions and different editions use different numbering for divisions within a text and the primary numbering (as indicated in the @n attribute) for the encoded edition should reflect the needs of the project. However, it is very useful to be able to align the sections in a text with other versions of the same text, and a reference to an established external set of divisions will allow different versions across different projects to be aligned and compared. Unfortunately very few online resources can be linked at this level by URI. The numbering of chapters from sites such as heimskringla.no can nevertheless be used using # notation to indicate the section on the page: http://heimskringla.no/wiki/Hrómundar_saga_Gripssonar#chapter=3 (ch. 3; name format: menota:work:handrit:Hrómundar+saga+Greipssonar#chapter=3); or http://heimskringla.no/wiki/Egils_saga_Skalla-Gr%C3%ADmssonar#chapter=60 (Eg ch. 60; menota:work:onp:109#chapter=60.
For resources that have named HTML or XML divisions (using @xml:id or @name), these should be used in the fragment section of the URL and will be processed when the application accesses the page. Other references will not be processed by the referenced resources, but can be used internally by Menota projects to reference parts of prose works: ‘chapter’ is used in the example above, and this or other divisions, if found in the most authoritative edition (i.e. the edition referenced by the relevant dictionary), can be used in the same format in the fragment section of the URI (e.g. section, stanza, volume, page).
188.8.131.52 Poems belonging to the Poetic Edda
At the time of writing, no authoritative digital edition of the eddic corpus is available, but there will likely be one soon, which may supercede the following.
A list of Eddic poems is available here: skaldic.abdn.ac.uk/db.php?id=425&if=default&table=skalds and the links provided there can be used to reference, for example, a poem: http://skaldic.abdn.ac.uk/db.php?id=2331&if=default&table=text and/or menota:work:skp:2331 (Hymiskviða); or a stanza http://skaldic.abdn.ac.uk/db.php?id=12722&if=default&table=verses and/or menota:stanza:skp:12722 (Hym 4)
184.108.40.206 Poems belonging to the skaldic corpus
All non-eddic, non-rímur poetry up to c. 1400 can be found on the Skaldic Project database at skaldic.abdn.ac.uk. These are organised by poet / speaker and prose work, with each skald, poem and stanza having an identifier.
The interface at skaldic.abdn.ac.uk/m.php?p=skaldic has simple URLS and comprehensive metadata. Browse poets, poems, and prose works to find individual stanzas and poems, and use the resulting URLs for references (e.g. http://skaldic.abdn.ac.uk/m.php?p=verse&i=2036 for Einarr Skúlason, Eysteinsdrápa 1)
220.127.116.11 Other works which are ambiguous or unavailable in the above
Some works, particularly postmedieval ones, may be found in Simek and Pálsson’s Lexicon der altnordisches Literatur (2007). In cases where there is ambiguity in the names above (such as in Handrit.is’s ‘uniform’ titles, the entry head form of the work may be used in the following name format (only the bold part of the head should be used, and cross references should be resolved — use only the forms with a full article, not the form followed by a cross-reference arrow):
menota:work:lanl:Hrómundar%20saga%20Gripssonar for Hrómundar saga Gripssonar
This particular work includes almost all West Norse prose works, Eddic poems and major skaldic poems.
Authoritative online resources are available for naming and referencing manuscripts relevant to the Old West Norse textual corpus. They should be used in the following way, and in the following order of priority — that is, only use lower priority sources where you are unable to find the manuscript in a higher priority source.
Handrit.is contains records on a large number of manuscripts in the field, and is the only online catalogue of a very large number of younger Icelandic manuscripts. The manuscripts are catalogued as TEI/XML files, and the @xml:id value in the <msDesc> element for each of these will contain a unique identifier, which is also used in the URL on the site, e.g. https://handrit.is/da/manuscript/view/is/JS04-0041. This produces the name menota:ms:handrit:JS04-0041.
The XML files may be in one or more of English, Danish or Icelandic, and the URL also includes the interface language (‘da’ for Danish above). The content of a particular entry may unfortunately differ according to the language, but any entry should contain sufficient information to identify the item and to establish basic metadata and relations.
Modify the URL so that the interface language is always English (handrit.is/en — not handrit.is/da or handrit.is/is), and choose the language of the catalogue entry according to the priority shown in 15.3 (English, Danish, Icelandic). E.g.
<msDesc me:ref="menota:ms:handrit:JS04-0041 https://handrit.is/en/manuscript/view/is/JS04-0041"> <msIdentifier> <idno>JS 41 4to</idno> </msIdentifer> </msDesc>
This identifies the particular Menota edition as belonging to the manuscript referenced, and allows an application to find all texts from a particular manuscript.
Do not use the mirror site at Handrit.org, as this does not produce URLs that link to individual manuscripts.
ONP’s Registre contains information on almost 5000 manuscripts, fragments and charters. This information is based on the published editions of the corpus used by the dictionary and is supplemented by more recent research on the dating of manuscripts.
The website now includes all these (https://onp.ku.dk/m. Use the permalink there to reference the ms. e.g. me:ref="https://onp.ku.dk/m8330"
FASNL (fasnl.ku.dk) is a fork of the Handrit.is catalogue relating to the fornaldarsaga corpus, with manuscripts supplemented from other sources. The XML files use the same format as Handrit.is but in a number of cases the XML identifiers for manuscripts have been modified, making some identifiers ambiguous. Additionally, the XML identifiers are not available through the web interface, although URLs are generated for each manuscript entry (e.g. for http://fasnl.ku.dk/browse-manuscripts/manuscript.aspx?sid=TQBTACAAMgAxAC4AMgAuADkA0 (Adv 21 2 9)). The name format should instead use the content of the main page heading, e.g. menota:ms:fasnl:MS%2021.2.9 for Adv 21 2 9). The future of the resource is not as secure as for Handrit.is, so it should be used only if the other resources here do not contain the manuscript.
18.104.22.168 Swedish manuscripts
The Manuscripta project (manuscripta.se) is an ongoing project to catalogue manuscripts in Swedish libraries. It has a simple interface and once a manuscript is found the resulting link is suitable for both naming the manuscript and locating the resource. E.g. Kungliga biblioteket, A 5 b (Birgitta, Uppenbarelser): https://www.manuscripta.se/ms/100206.
This resource should be used for manuscripts in Swedish libraries if they are documented in the resource. As it is currently in progress, you should check periodically for updates relevant to your project.
22.214.171.124 Manuscript pages
A consistent naming system for divisions of manuscripts allows for two-way semantic linking of pages across different documents, corpora and image repositories, for example, to link a transcription of a page in the repository to an edition of a verse stanza on the same page in the Skaldic project’s corpus and an image of the page on Handrit.is or the forthcoming digitised photograph library from AMS in Copenhagen. Numbering of manuscript pages across catalogues and editions is not very consistent for a large number of documents, and most systems use the numbering as the only reference point to the page in the manuscript — that is, there is rarely a machine-readable referencing system that would translate between different numbering systems for a particular manuscript.
We recommend using the pagination/foliation of the manuscript according to the priority in ch. 126.96.36.199–4, namely the pagination found in the chosen catalogue. For the naming system, the fragment part of the URI should be used to specify the page with the page= format (regardless of whether the numbering uses folio or page numbers) also used above for bibliographic references (ch. 15.4.1). E.g.
Handrit.is or other sources may be used to additionally specify an image URL to help the end-user check what is in fact on the page in question, although they do not supply in themselves any machine-readable metadata or other information about the page:
In this case, the image URL is extracted from the Handrit.is link by right-clicking on the relevant page and opening the image in a new tab/window.
The main reference point should be the Scandinavian Biographical Index https://www.degruyter.com/viewbooktoc/product/31428 (ISBN 978-3-11-091414-6) Use the full name there, with additional dates or information if necessary for disambiguation:
(Not: ‘Snorri Sturluson yngri’, unless that person is meant.)
Handrit.is includes a list of names associated with Icelandic manuscripts and texts which are standardised to a reasonable extent, e.g.:
https://handrit.is/da/biography/view/SnoStu001 (Snorri Sturluson) menota:person:handrit:SnoStu001 https://handrit.is/da/biography/view/SaeSig002 (Sæmundur Sigfússon ; fróði) menota:person:handrit:SaeSig002
Other figures, such as many people appearing in saga narratives, are not documented systematically in these external resources. In such cases, you can use or create a Wikipedia or Wikidata entry and use the resulting Wikidata URI to reference the person, e.g. (Lbs Frag 82, 1v/7):
<persName ref="https://www.wikidata.org/wiki/Q940404"> <forename> <w lemma="Valþjófr" me:ref="menota:lemma:onp:83796" me:msa="xNP gM nS cN sI"> <choice> <me:facs>Valþıoꝼr</me:facs> <me:dipl>Valþiofr</me:dipl> <me:norm>Valþjófr</me:norm> </choice> </w> </forename> <roleName type="political"> <w lemma="jarl" me:ref="menota:lemma:onp:41394" me:msa="xNC gM nS cN sI"> <choice> <me:facs><am>.</am>&jnodot;<am>.</am></me:facs> <me:dipl>j<ex>arl</ex></me:dipl> <me:norm>jarl</me:norm> </choice> </w> </roleName> </persName>
In this case, the Wikidata reference includes all relevant information about the person, including dates, parents and role, with an API that can be used to incorporate this information into an application. Wikidata references can therefore be used to supplement the above references.
The reference here is to places rather than place-names, but the tagging will generally be done when a place-name is encountered. Ideally the place reference will be to a digital resource that contains disambiguating information (e.g. district names) and geodata coordinates so that the reference can be used to map the places and process geographical relationships with other places. Unfortunately there are very few resources that provide this function, apart from less authoritative open resources.
The primary point of reference for current places should be download.kortforsyningen.dk/content/stednavne. Unfortunately there is no web interface to this resource where keys / URIs can be extracted but by searching the dataset available at this link (GIS shape file) one can find the entries for current place names (the FEAT_ID column in the data table).
menota:place:sdfe:159733 (Gammel Lejre / Hleiðagarðr)
Danmarks Stednavne (danmarksstednavne.navneforskning.ku.dk) should be used where the previous does not contain the place name, for example, where the settlement is no longer present. Unfortunately the service does not currently provide URIs for individual places, but may do so in the future. Use the link to search for the place name and once found, click ‘Vis link til søgningen’ and ‘Hent søgeresultat i Excel-format’. The resulting xls file will contain the placename ids as the first column, along with coordinate and other data. The place can be referenced:
menota:place:dsn:167334 (Gammel Lejre / Hleiðagarðr)
You may wish to save the other data, including coordinates, in an authority file or database table for use by the project or repository.
Landmælingar Íslands (lmi.is) is the central mapping and place-names authority in Iceland. Their interactive web map unfortunately does not reveal unique identifiers for their place-name database, but place-name data can be downloaded from their site. Register for the download service and download the zip file for the IS 50V dataset’s place names (under 'IS 50V 17/06 2017 - Örnefni. GDB og SHP. ISN2004').
You will find the GIS shape files in the IS50V_ORNEFNI_17062017_ISN2004.zip file under the IS50V_ORNEFNI_SHP folder. There are separate files for polygons (is50v_ornefni_flakar_17062017.shp), lines (is50v_ornefni_linur_17062017.shp) and points (is50v_ornefni_punktar_17062017.shp). The numeric identifiers (NAMEID) appear to be unique across the different tables, which means that they can be used on their own.
Search for the place name (NAFNFITJU column) in the tables associated with the shape files and use the resulting NAMEID value to identify the place in question, e.g.:
menota:place:lmi:18513 (Helgafell (Fjall), Helgafellssveit)
Kartverket is the central Norwegian mapping authority and its interactive web map (www.norgeskart.no) can be used to identify current places which correspond to historical ones. Search the map and select the desired place. Click the link from the left-hand menu ‘Se fakta om stedsnavnet’ and use either or both the resulting URI or name based on the ‘Stedsnummer’:
menota:place:kartverket:558690 (Trondheim / Niðarós)
Additional historical names can be found online in O. Rygh’s Norske Gaardnavne (). Search for the place using the form in the link given here and then reference the name using the information given in the results. The name (Gård) should be substituted with the number (Gårdsnr) in case more than one farm appears in the same parish, and the parish, district and region (Sogn, Herred, Amt) should all be included separated by commas:
menota:place:ngn:70,Røken,Røken,Buskeruds%20amt (Auvi søndre, Røken sn., Buskerud)
TORA - Topografiskt register på Riksarkivet (riksarkivet.se/tora) should eventually be able to provide the necessary services for linking Swedish historical places, but at this stage it does not offer public resources for linking.
Lantmäteriet (lantmateriet.se) is otherwise the relevant authority for geographical data in Sweden, including place names. Unlike similar bodies in the Nordic countries, the geographical data is only licensed for free for educational and research institutions within Sweden (through geodata.se). The web map service at kso.etjanster.lantmateriet.se allows for place-name searches, but there are no unique identifiers available through this service. Searches on this service show the name (namn), municipality (kommun), district (län) and name type (namntyp), which can be taken together in this order to identify the place where unambiguous (use the Swedish version of any name), e.g.
menota:place:lm:Torsbro,Uppsala,Uppsala,Bebyggelse (Torsbro, Uppland)
But other names may be ambiguous by this method, e.g. Nöbbele, Växjö, Kronoborgs län (Bebyggelse) matches six distinct places, in which case you should use the resources documented in the next section to identify the actual place.
188.8.131.52 Unofficial and historical places: sources
OSM is the main open geographical resource in use today. It allows places to be indicated by points, lines, areas (polygons) or groups of areas, and referenced with a numerical key. Where a place exists today and can be identified with the place in the text, this should be used, e.g. https://www.openstreetmap.org/node/4333948261 (Helgafell, Snæfellsnes, Iceland). OSM excludes historical information, that is, places which no longer exist. In some cases these might be identifiable with archaeological sites (e.g. https://www.openstreetmap.org/way/219229751 — hall site at Gammel Lejre), but there is some uncertainty in such attributions. Open Historical Map appears to be a similar project but with a historical focus that is deliberately excluded from OSM. There appears to be very limited data for the Nordic countries at this stage.
This site predates Open Street Map and has a similar user-contributed data set. It is easy to edit and has a simple API, but only allows places to be indicated by points or squares on a map (rather than points, lines, areas (polygons) or groups of areas or lines as in OSM). Unlike OSM, it allows historical names and places to be entered, but with an easier interface than Openhistoricalmap. Only the numeric part of the URL is needed, e.g.: http://www.geonames.org/3133880 (Trondheim / Nidaros); http://www.geonames.org/3416474 (Helgafell); http://www.geonames.org/10242884 (Dritsker). Where names are missing, they should be entered and referenced using the site’s interface.
NB save separately any entered information including coordinates as other users can easily delete your contributions to this service.
Wikidata (check Wikipedia first) can be used as a last resort as many places, including historical and legendary places are often documented there. Wikidata also has a consistent georeferencing system that can be extracted through an API. There are problems, however. A reference to Bjarmaland could be linked to: https://www.wikidata.org/wiki/Q524769 or https://en.wikipedia.org/wiki/Bjarmaland but the Wikipedia / Wikidata entry there has no geodata, even though the location of the historical place can be roughly identified. A different problem comes from the example of Niðarós: it is functionally identical as a settlement with modern Trondheim, but Wikipedia has entries for both: https://en.wikipedia.org/wiki/Nidaros and https://en.wikipedia.org/wiki/Trondheim. While it would be tempting to use the entry closest to the old name, only the latter has geodata attached to it, meaning that the application or archive could not map the place if the former reference is given. (This example is just to illustrate the point: the place is found in its modern form in the official gazetteers given above, and the two names are linked under the Wikidata entry, together with geolocation, at https://www.wikidata.org/wiki/Q25804.)
Events such as battles and journeys may be documented in various encyclopediae, but are highly dependent on the context of the event to find a reliable source. We therefore recommend using Wikidata to identify the event. Where there is no current documentation for the event, an entry in Wikidata/Wikipedia should be created in order to provide a common reference point for different projects (but before making a new entry always check very carefully all possible names and naming schemes to ensure that no current entry exists).
E.g. https://www.wikidata.org/wiki/Q83078 (cf. https://en.wikipedia.org/wiki/Battle_of_Stiklestad — Battle of Stiklarstaðir) — see also ch. 15.3 above.
15.4.8 Mythological and legendary figures, or anything else
As with ch. 15.4.7 Events, this category (of named or sometimes even unnamed beings, places or objects which are mentioned in Menota texts, but which do not seem to have an existence in objective reality) is documented in a range of encyclopediae and lexica, but is very dependent on the context of an individual text or reference to work out what may be authoritative or otherwise. We recommend using Wikidata to identify the figure, place or object. Where there is no current entry, a new entry in Wikidata/Wikipedia should be created in order to provide a common reference point for different projects (but before making a new entry always check very carefully all possible names and naming schemes to ensure that no current entry exists).
Such names or identifications may be the subject of controversy, such as the identification of the mythological figure Rígr with Heimdallr in AM 242 fol. The Wikipedia page for ‘Ríg (Norse deity)’ redirects to the entry for the poem Rígsþula, but an entry for the apparent god himself exists on Wikidata (https://www.wikidata.org/wiki/Q15843446). When beginning this chapter, there was no reference to Heimdallr here, but using the facility to define relations on Wikidata, the author of the present chapter has added a new statement to the Wikidata entry: ‘said to be the same as’ > ‘Heimdall [Norse god]’, with qualifiers: ‘criterion used’ > ‘Codex Wormianus’ (i.e. attribution is made here), ‘statement disputed by’ > ‘Rudolf Simek’, ‘statement disputed by’ > ‘Finnur Jónsson’. In other words, the reference to the Wikidata entry for Rígr can be used to accurately reference the god in the manifestation he takes in AM 242 fol., but with a machine-readable statement concerning the possible identification of the god with Heimdallr, including the origin of the identification and those that dispute it (with links to all). The principle here is that a more specific reference is preferable, but the reference should include possible relationships to other items and an indication of controversy, where it legitimately exists.