P oceedings ExLing 2019: 10 h Tu o ial and Resea ch Wo kshop on Expe imen al
Linguis ics, 24-27 Sep embe , Lisbon, Po ugal
Expe imen al cus omisa ion o he Ve sioning
Machine
Elsa Pe ei a1, Roman Bleie 2
1CLUL, Uni e si y o Lisbon, Po ugal
2ZIM, Uni e si y o G az, Aus ia
DOI: 10.36505/ExLing-2019/10/0041
Abs ac
This pape deals wi h challenges in adap ing he XML-TEI publishing amewo k
Ve sioning Machine o composi ional d a s o 20 h-cen u y li e a y wo ks and
desc ibes he main cus omisa ions ha ha e been implemen ed o sui a gene ic
edi ion o poe y by Ped o Homem de Mello. The case s udy emphasises ha e en
minimal cus omisa ions equi e echnical wo k ha may go beyond an edi o ’s skill.
Keywo ds: ep esen a ion, d a s, TEI, Ve sioning Machine, cus omisa ion
Rep esen a ion o composi ional d a s
Ea ly 20 h-cen u y li e a y wo ks a e o en documen ed in a a ie y o
au ho ial wi nesses, including d a manusc ip s wi h se e al laye s o
addi ions, dele ions, and ea angemen s. A schola ly edi ion wi h a gene ic
o ien a ion – aimed a achie ing an insigh ul o e iew o he composi ional
and e isional de elopmen o hose wo ks – needs o ep esen he
o e w i ing laye s in each d a , while signalling he successi e a ian s ha
occu ac oss di e en wi nesses.
In digi al schola ly edi ing, he p ocess in ol es wo main s ages: da a
modelling and p esen a ion. The i s s age is achie ed using desc ip i e
ma kup o p oduce machine- eadable ansc ip ions, and while g aph and
hype g aph da a models such as TAG (Dekke & Bi nbaum 2017) a e
eme ging as an al e na i e o he XML pa adigm, he s anda d main ained by
he Tex Encoding Ini ia i e [TEI] p e ails as one o he mos obus
schemas a ailable. I allows o a ep esen a ion o bo h he ma e iali y and
he ex ual dimension o manusc ip s and makes i possible o combine in a-
and in e -documen a y a ia ion, using chap e s 11 and 12 o he TEI P5
guidelines. The second s age con e s he encoded ex s in o a isual laye o
be accessed by he eade and is achie ed by using XML ans o ma ion and
que y languages (XSLT, XQue y, XPa h), as well as HTML, CSS, and
Ja aSc ip , o build an in e ac i e g aphical use in e ace. Such a wide ange
o echnologies equi es ex ensi e echnical suppo , which is no always
a ailable o indi idual o disc e ely unded p ojec s.
The TEI communi y is awa e o his p oblem and has de eloped ools and
publishing solu ions o XML-TEI, sha ed as open-sou ce. Among he ligh -
Expe imen al cus omisa ion o he Ve sioning Machine
167
weigh solu ions a ailable, men ion should be made o he TEI Boile pla e
(Walsh e al.), he Ja aSc ip lib a y CETEIcean (Cayless & Viglian i 2018),
EVT (Di Pie o & Rosselli Del Tu co 2018), as well as he Ve sioning
Machine [VM] (Sch eibman e al. 2003; 2016) – one o he ea lies and mos
widely known. O iginally de eloped o ace he composi ion his o y o
poems w i en by Thomas MacG ee y, he ool was used in se e al
in e na ional p ojec s and has been adap ed o special publica ion needs
o e he yea s.1
Cus omisa ion o he Ve sioning Machine
VM is speci ically concei ed o display mul iple e sions o ex encoded
wi h he TEI c i ical appa a us ag-se and addi ionally allows a limi ed
numbe o elemen s o ep esen in a-documen a y a ia ion. By edi ing he
ool’s XSLT and CSS s yleshee s,2 we ha e achie ed a wide ange o
p esen a ional ea u es, o sui an edi ion o p in ma e ial and composi ion
d a s by he Po uguese poe Ped o Homem de Mello (1904-1984). The
al e a ions conce n he display o h ee main ca ego ies: me ada a, d a
laye s o e ision, and in e -documen a y a ia ion.
The de aul VM se up p o ides a basic p esen a ion o me ada a,
con ained by he s anda d elemen s o he ei: eiHeade , o be displayed in
he so-called “Bibliog aphic Panel”. Howe e , an edi ion o he gene ic
o ien a ion, ocused “on he documen s as sou ces o e idence o ex ual
de elopmen and change h ough ime” (Van Hulle & Shillingsbu g 2015:
36), equi es mo e de ailed desc ip ions. Fo his eason, we included a
“Composi ional Synopsis”, con aining he in o ma ion om he ei:c ea ion
elemen (encoded wi hin he p o ile desc ip ion) and modi ied he “O iginal
Sou ce” sec ion, by adding and s uc u ing he con en o ei:msDesc and
ei:biblS uc (wi hin he sou ce desc ip ion), as illus a ed by example 1.3
Se e al al e a ions we e made o sui he mul i-laye ed ansc ip ion o
complex documen a y sou ces. Al e na i e eadings, o ins ance, a e no
co e ed by he de aul VM se up, bu i was necessa y o include hem o
he p ojec , since d a manusc ip s occasionally ha e wo o mo e ph asings
a speci ic poin s o he ex , wi hou showing he au ho ’s p e e ence o any
o he op ions lis ed. In he XML-TEI his may be encoded using he ei:seg
and ei:add elemen s wi h an @ ype="al e na i e", o which we had o
c ea e a ans o ma ion ule, displaying one eading abo e he o he , as seen
in example 2.
Also impo an was ende ing encoded gaps and signi ican spaces, as
well as ex ending he display o au ho ial addi ions, h ough he p ocessing
o he alues " op", "bo om", "ma ginLe ", "ma ginRigh ", "o e lea ", and
"opposi e" o @place, since VM only deals wi h "abo e" and "below" by
de aul (see example 3). Howe e , e isions ha in ol e se e al lines o
E. Pe ei a, R. Bleie
168
s anzas p esen a challenge. I he ei:add and ei:del elemen s a e no
allowed, due o con lic ing XML hie a chies,4 he TEI guidelines sugges
using miles one o emp y elemen s: i.e. ei:addSpan and ei:delSpan wi h a
@spanTo a ibu e poin ing o he ID o a ei:ancho ha indica es how a
he addi ion o dele ion goes. The de aul s yleshee s o VM do no con ain
ins uc ions o p ocess hese miles one-delimi ed anges, so new
ans o ma ion ules we e equi ed (see example 4).
Ano he speci ici y o d a manusc ip s is he epe i ion o e ised
s anzas o lines o cla i y he wo ding. The TEI guidelines sugges encoding
such cla i ica ions wi h he ei: e ace elemen , which was in oduced in
2011, as pa o an encoding model o gene ic edi ing. VM was designed o
he c i ical appa a us ag-se and, he e o e, does no suppo his elemen by
de aul . A empla e ule had o be implemen ed o p ocess and display
ei: e ace and i s @ end a ibu es, as shown in example 5.
Fo he same eason, VM does no deal wi h displacemen s o ex
(indica ed by a ows o o he symbols in he au ho ’s d a s), and new
ans o ma ion ules we e necessa y o p ocess hese occu ences. In he
p ojec ’s encoding model, ei:di con aining displaced ex is speci ied by an
@ ype wi h he alue "displacedF agmen ", a ei:me ama k indica es he
place o inse ion, and bo h elemen s a e linked using an ID. Example 6
illus a es he ende ing achie ed by he modi ied s yleshee s.
Besides he p esen a ion o in e nal laye s o e ision, cus omisa ions
we e inally made o display he in icacies o in e -documen a y a ia ion in
wo ks wi h mul iple d a s. Acco ding o he TEI P5 guidelines, appa a us
en ies may nes , using pa allel segmen a ion wi h op ional loca ion-based
e e encing. While he VM is compa ible wi h his encoding me hod, he
de aul se up obscu es he display o sub a ian s, because o he indis inc
yellow highligh applied o all ei: dg elemen s. We changed his by applying
a di e en backg ound colou whene e ei:app elemen s nes wi h di e en
@loc ID, as illus a ed by example 7.
Conclusions
By looking a one speci ic so wa e applica ion ha was de eloped o he
publica ion o TEI da a – he Ve sioning Machine – his pape explo ed he
cus omisa ions necessa y o adap he so wa e o an edi ion’s needs and
encoding model. In his case, a se ies o al e a ions had o be made o
combine in a- and in e -documen a y a ia ion in poems whe e mul iple
d a s coexis .
The TEI ies o sui a di e se communi y o humani ies schola s, which
equi es g ea lexibili y o suppo di e en ex ual models and edi ing
s a egies. This ci cums ance means ha he s anda d is no pa icula ly well
sui ed o he de elopmen o gene ic publica ion solu ions. While some o
Expe imen al cus omisa ion o he Ve sioning Machine
169
he changes desc ibed in his pape could be achie ed wi h minimal changes
o he VM sou ce code, i s ill equi es an unde s anding o web echnologies
o implemen hem, and no e e y ex ual schola wi h basic XML
knowledge has hese skills o he esou ces o pay o hem. Hence, he
1S/1P/1DH (one schola , one p ojec , one digi al humanis ) pa adigm
desc ibed by Robinson (2013) ine i ably pe sis s. The ques ion ha emains
is how his can be o e come in he u u e.
No es
1. A selec ion o p ojec s is lis ed on he websi e: h p:// -machine.o g/ m-in-use.
The Wande ing Jew’s Ch onicle is a good example o special cus omisa ion.
2. See machine.xsl and machine.css in he “s c” olde o ou VM ins ance:
h ps://www.d opbox.com/sh/e habpwahu83j1/AAAE7nWx9woN dF5Hu_aU4
1na?dl=0
3. Examples men ioned in his pape a e in he “samples” olde o ou VM
ins ance.
4. See chap e 20 “Non-hie a chical S uc u es” o he TEI guidelines.
Acknowledgemen s
The au ho s wish o hank he VM eam o de eloping he Ve sioning
Machine and Gael Vaamonde o his help du ing cus omising expe imen s.
Re e ences
Cayless, H., Viglian i, R. 2018. CETEIcean: TEI in he B owse . In P oceedings o
Balisage: The Ma kup Con e ence 2018. Balisage Se ies on Ma kup
Technologies 21. h ps://doi.o g/10.4242/BalisageVol21.Cayless01.
Dekke , R.H., Bi nbaum, D. J. 2017. I ’s mo e han jus o e lap: Tex As G aph. In
P oceedings o Balisage: The Ma kup Con e ence 2017. Balisage Se ies on
Ma kup Technologies 19. h ps://doi.o g/10.4242/BalisageVol19.Dekke 01.
Di Pie o, C., Rosselli Del Tu co, R. 2018. Be ween Inno a ion and Conse a ion:
The Na ow Pa h o Use In e ace Design o Digi al Schola ly Edi ions. In
Bleie , R. e al. (eds.) 2018, Digi al Schola ly Edi ions as In e aces, 133-163.
No de s ed : Books on Demand.
Robinson, P. 2013. Why digi al humanis s should ge ou o ex ual schola ship.
h p://schola lydigi aledi ions.blogspo .com/2013/07/why-digi al-humanis s-
should-ge -ou -o .h ml.
Sch eibman, S., Kuma , A., McDonald, J. 2003. The Ve sioning Machine. Li e a y
and Linguis ic Compu ing 18-1, 101-107. h ps://doi.o g/10.1093/llc/18.1.101.
Sch eibman, S., e al. 2016. Ve sioning Machine 5.0. h p:// -machine.o g.
Tex Encoding Ini ia i e. 2019. P5: Guidelines o Elec onic Tex Encoding and
In e change. h ps://www. ei-c.o g/ elease/doc/ ei-p5-doc/en/h ml/index.h ml.
Van Hulle, D., Shillingsbu g, P. 2015. O ien a ions o ex , e isi ed. S udies in
Bibliog aphy 59, 27-44.
Walsh, J. e al. TEI Boile pla e. h p://dcl.ils.indiana.edu/ eibp/index.h ml.