際際滷

際際滷Share a Scribd company logo

A Spot of TEI

Hugh Cayless
Hugh Cayless

Presentation on TEI (particularly as it relates to the Integrating Digital Papyrology project). Given at U. of South Carolina Center for Digital Humanities, 2/4/2013

1 of 17
Download to read offline
A spot of TEI Hugh Cayless, NYU philomousos@gmail.com follow me on Twitter: @hcayless February 4th, 2013
Who am I? Ph.D. in Classics, M.S. in Information Science Worked as a software engineer for the last 12 years or so the last 4 have been for NYU doing Digital Classics and similar cultural heritage digital access projects recently elected to the TEI Technical Council. One of the founders of EpiDoc, a TEI-based standard for encoding ancient inscriptions (and now papyri too).
What am I talking about? How we use TEI/XML in projects Why TEI? Current projects
Integrating Digital Papryology Uni鍖cation of several long-running projects: Duke Databank of Documentary Papyri (DDbDP) Heidelberg Gesamtverzeichnis (directory of Greek documentary papyri HGV) Advanced Papyrological Information System (APIS) Bibliographie Papyrologique Trismegistos
State of play at the beginning DDbDP: TEI SGML 鍖les HGV: Filemaker Pro database + web interface APIS: idiosyncratic text-based catalog + images + web interface BP: database only, published annually in print/on disk TM: database + web interface TM is a going concern, working with IDP, but with no plans to be subsumed by it
What we did DDbDP: converted TEI SGML to EpiDoc (TEI) XML HGV: converted to EpiDoc XML APIS: converted to EpiDoc XML BP: converted to TEI <bibl> fragments TM: inserted TM ids into IDP documents, generated linkages to TM site
Structure The core of the system is just TEI 鍖les in a Git repository. These are transformed, using XSLT, into RDF, HTML, plain text, and add documents for our search index. They are pulled into an editing work鍖ow system as needed, which allows editing the 鍖les using a web form or (for texts) a non-XML syntax based on papyrological/epigraphic editing conventions. An automated process syncs data from the editors repo and a Github repo, and publishes them to the site.
Or, visually Automated Document Sync Canonical Git Repo Numbers Papyri.info Server Editor Search Git Repo Database Github Repo Engine Git Repos Leiden+ Conversion API search Navigator SPARQL Github Editor API Interface API XSLT API
So why TEI? Lots of reasons: Granular control over records Attribution Multiple outputs Mixture of controlled and free-form data Relatively easy to obtain / create tools Engaged and responsive community
What Im working on now Fixing the TEI Pointer spec Annotation of documents to mark things like personal and place names Linguistic annotation Linking text and image
Some examples http://papyri.info/ddbdp/cpr;8;72 鍖ne-grained attribution / version control (click on Editorial History) and Detailed at the bottom of the text) http://papyri.info/ddbdp/c.ep.lat;;218 Whats going on underneath?
Beginning of a letter marked up according to the Leiden Conventions r 味[味味味]味c味[味味 Aelio Fel]ici plu畊[imam] 畊ａ此[lutem] opto deos 揃 ut mi[hi v]畉leas 揃 quod 畊畉[um votum est] ego enim 揃 vale畛 co畊p畛畊e 味味味[ -ca.?- ] te non vide畛 rog味畛 畊e 揃 fac味畛a畊 [ -ca.?- ] f味味味味味味味味味味[- ca.9 -]uma味[ -ca.?- ] 味味味味味味味味味味 [ -ca.?- ] v Aelio Felici
The same letter, diplomatic(ish) edition r 味[味味味]味C味[味味ca.9 ]ICIPLU畊[. . . .] 畊≡[. . . . .] OPTODEOS 揃 UTMI[. . . ]畉LEAS 揃 QUOD畊畉[ ca.10 ] EGOENIM 揃 VALE畛CO畊P畛畊E 味味味[ -ca.?- ] TENONVIDE畛ROG味畛畊E 揃 FAC味畛A畊 [ -ca.?- ] F味味味味味味味味味味[- ca.9 -]UMA味[ -ca.?- ] 味味味味味味味味味味 [ -ca.?- ] v AELIO FELICI
The same letter marked up in EpiDoc (TEI) XML <div xml:lang="la" type="edition" xml:space="preserve"><div n="r" type="textpart"><!-- milestone unit="4"--><ab> <lb n="1"/><gap reason="illegible" quantity="1" unit="character"/><gap reason="lost" quantity="3" unit="character"/><gap reason="illegible" quantity="1" unit="character"/>c<gap reason="illegible" quantity="1" unit="character"/><gap reason="lost" quantity="2" unit="character"/><supplied reason="lost"> Aelio Fel</supplied>ici plu<unclear>r</ unclear><supplied reason="lost">imam</supplied> <lb n="2"/><unclear>sa</unclear><supplied reason="lost">lutem</supplied> <lb n="3"/>opto deos <g type="middot"/> ut mi<supplied reason="lost">hi v</ supplied><unclear>a</unclear>leas <g type="middot"/> quod <unclear>me</unclear><supplied reason="lost">um votum est</supplied> <lb n="4"/>ego enim <g type="middot"/> vale<unclear>o</unclear> co<unclear>r</ unclear>p<unclear>or</unclear>e <gap reason="illegible" quantity="3" unit="character"/><gap reason="lost" extent="unknown" unit="character"/> <lb n="5"/>te non vide<unclear>o</unclear> ro<unclear>go</unclear> <unclear>n</unclear>e <g type="middot"/> fa<unclear>ci</unclear>a<unclear>s</unclear> <gap reason="lost" extent="unknown" unit="character"/> <lb n="6"/>f<gap reason="illegible" quantity="4" unit="character"/><gap reason="illegible" quantity="4" unit="character"/><gap reason="illegible" quantity="2" unit="character"/><gap reason="lost" quantity="9" unit="character"/>uma<gap reason="illegible" quantity="1" unit="character"/><gap reason="lost" extent="unknown" unit="character"/> <lb n="7"/><gap reason="illegible" quantity="4" unit="character"/><gap reason="illegible" quantity="4" unit="character"/><gap reason="illegible" quantity="2" unit="character"/> <gap reason="lost" extent="unknown" unit="character"/> </ab></div><div n="v" type="textpart"><!--milestone unit="4"--><ab> <lb n="1"/>Aelio Felici </ab></div></div>
The same letter, visualization of the tree structure of the XML
What is the text and what is the markup? There is no text, only readings. EpiDoc allows you to produce models of readings. Slicing the text up into bits isnt adulterating it, it just adds hooks for transforming the text in useful ways.
How to get involved Mailing list: TEI-L@LISTSERV.BROWN.EDU http://listserv.brown.edu/archives/cgi-bin/wa?SUBED1=tei-l&A=1 http://listserv.brown.edu/archives/cgi-bin/wa?A0=tei-l TEI Sourceforge: Report a bug: http://sourceforge.net/tracker/?func=add&group_id=106328&atid=644062 Make a feature request: http://sourceforge.net/tracker/?func=add&group_id=106328&atid=644065 IRC: #tei-c on http://freenode.net/

More Related Content

A Spot of TEI

  • 1. A spot of TEI Hugh Cayless, NYU philomousos@gmail.com follow me on Twitter: @hcayless February 4th, 2013
  • 2. Who am I? Ph.D. in Classics, M.S. in Information Science Worked as a software engineer for the last 12 years or so the last 4 have been for NYU doing Digital Classics and similar cultural heritage digital access projects recently elected to the TEI Technical Council. One of the founders of EpiDoc, a TEI-based standard for encoding ancient inscriptions (and now papyri too).
  • 3. What am I talking about? How we use TEI/XML in projects Why TEI? Current projects
  • 4. Integrating Digital Papryology Uni鍖cation of several long-running projects: Duke Databank of Documentary Papyri (DDbDP) Heidelberg Gesamtverzeichnis (directory of Greek documentary papyri HGV) Advanced Papyrological Information System (APIS) Bibliographie Papyrologique Trismegistos
  • 5. State of play at the beginning DDbDP: TEI SGML 鍖les HGV: Filemaker Pro database + web interface APIS: idiosyncratic text-based catalog + images + web interface BP: database only, published annually in print/on disk TM: database + web interface TM is a going concern, working with IDP, but with no plans to be subsumed by it
  • 6. What we did DDbDP: converted TEI SGML to EpiDoc (TEI) XML HGV: converted to EpiDoc XML APIS: converted to EpiDoc XML BP: converted to TEI <bibl> fragments TM: inserted TM ids into IDP documents, generated linkages to TM site
  • 7. Structure The core of the system is just TEI 鍖les in a Git repository. These are transformed, using XSLT, into RDF, HTML, plain text, and add documents for our search index. They are pulled into an editing work鍖ow system as needed, which allows editing the 鍖les using a web form or (for texts) a non-XML syntax based on papyrological/epigraphic editing conventions. An automated process syncs data from the editors repo and a Github repo, and publishes them to the site.
  • 8. Or, visually Automated Document Sync Canonical Git Repo Numbers Papyri.info Server Editor Search Git Repo Database Github Repo Engine Git Repos Leiden+ Conversion API search Navigator SPARQL Github Editor API Interface API XSLT API
  • 9. So why TEI? Lots of reasons: Granular control over records Attribution Multiple outputs Mixture of controlled and free-form data Relatively easy to obtain / create tools Engaged and responsive community
  • 10. What Im working on now Fixing the TEI Pointer spec Annotation of documents to mark things like personal and place names Linguistic annotation Linking text and image
  • 11. Some examples http://papyri.info/ddbdp/cpr;8;72 鍖ne-grained attribution / version control (click on Editorial History) and Detailed at the bottom of the text) http://papyri.info/ddbdp/c.ep.lat;;218 Whats going on underneath?
  • 12. Beginning of a letter marked up according to the Leiden Conventions r 味[味味味]味c味[味味 Aelio Fel]ici plu畊[imam] 畊ａ此[lutem] opto deos 揃 ut mi[hi v]畉leas 揃 quod 畊畉[um votum est] ego enim 揃 vale畛 co畊p畛畊e 味味味[ -ca.?- ] te non vide畛 rog味畛 畊e 揃 fac味畛a畊 [ -ca.?- ] f味味味味味味味味味味[- ca.9 -]uma味[ -ca.?- ] 味味味味味味味味味味 [ -ca.?- ] v Aelio Felici
  • 13. The same letter, diplomatic(ish) edition r 味[味味味]味C味[味味ca.9 ]ICIPLU畊[. . . .] 畊≡[. . . . .] OPTODEOS 揃 UTMI[. . . ]畉LEAS 揃 QUOD畊畉[ ca.10 ] EGOENIM 揃 VALE畛CO畊P畛畊E 味味味[ -ca.?- ] TENONVIDE畛ROG味畛畊E 揃 FAC味畛A畊 [ -ca.?- ] F味味味味味味味味味味[- ca.9 -]UMA味[ -ca.?- ] 味味味味味味味味味味 [ -ca.?- ] v AELIO FELICI
  • 14. The same letter marked up in EpiDoc (TEI) XML <div xml:lang="la" type="edition" xml:space="preserve"><div n="r" type="textpart"><!-- milestone unit="4"--><ab> <lb n="1"/><gap reason="illegible" quantity="1" unit="character"/><gap reason="lost" quantity="3" unit="character"/><gap reason="illegible" quantity="1" unit="character"/>c<gap reason="illegible" quantity="1" unit="character"/><gap reason="lost" quantity="2" unit="character"/><supplied reason="lost"> Aelio Fel</supplied>ici plu<unclear>r</ unclear><supplied reason="lost">imam</supplied> <lb n="2"/><unclear>sa</unclear><supplied reason="lost">lutem</supplied> <lb n="3"/>opto deos <g type="middot"/> ut mi<supplied reason="lost">hi v</ supplied><unclear>a</unclear>leas <g type="middot"/> quod <unclear>me</unclear><supplied reason="lost">um votum est</supplied> <lb n="4"/>ego enim <g type="middot"/> vale<unclear>o</unclear> co<unclear>r</ unclear>p<unclear>or</unclear>e <gap reason="illegible" quantity="3" unit="character"/><gap reason="lost" extent="unknown" unit="character"/> <lb n="5"/>te non vide<unclear>o</unclear> ro<unclear>go</unclear> <unclear>n</unclear>e <g type="middot"/> fa<unclear>ci</unclear>a<unclear>s</unclear> <gap reason="lost" extent="unknown" unit="character"/> <lb n="6"/>f<gap reason="illegible" quantity="4" unit="character"/><gap reason="illegible" quantity="4" unit="character"/><gap reason="illegible" quantity="2" unit="character"/><gap reason="lost" quantity="9" unit="character"/>uma<gap reason="illegible" quantity="1" unit="character"/><gap reason="lost" extent="unknown" unit="character"/> <lb n="7"/><gap reason="illegible" quantity="4" unit="character"/><gap reason="illegible" quantity="4" unit="character"/><gap reason="illegible" quantity="2" unit="character"/> <gap reason="lost" extent="unknown" unit="character"/> </ab></div><div n="v" type="textpart"><!--milestone unit="4"--><ab> <lb n="1"/>Aelio Felici </ab></div></div>
  • 15. The same letter, visualization of the tree structure of the XML
  • 16. What is the text and what is the markup? There is no text, only readings. EpiDoc allows you to produce models of readings. Slicing the text up into bits isnt adulterating it, it just adds hooks for transforming the text in useful ways.
  • 17. How to get involved Mailing list: TEI-L@LISTSERV.BROWN.EDU http://listserv.brown.edu/archives/cgi-bin/wa?SUBED1=tei-l&A=1 http://listserv.brown.edu/archives/cgi-bin/wa?A0=tei-l TEI Sourceforge: Report a bug: http://sourceforge.net/tracker/?func=add&group_id=106328&atid=644062 Make a feature request: http://sourceforge.net/tracker/?func=add&group_id=106328&atid=644065 IRC: #tei-c on http://freenode.net/