From cwd3@COLUMBIA.EDU Tue Oct 26 10:13:46 1999
Date: Thu, 13 May 1999 08:00:59 -0400
From: Consuelo W Dutschke <cwd3@COLUMBIA.EDU>
Reply-To: TEI Medieval Manuscripts Description Work Group
    <TEI-MMSS@LISTSERV.UIC.EDU>
To: TEI-MMSS@LISTSERV.UIC.EDU
Subject: draft-Rome minutes

TEI Work Group for Description of Medieval Manuscripts
MINUTES

Vatican City, Biblioteca Apostolica Vaticana, Sala della Biblioteca
Barberini
5-7 March 1999
Under the auspices of Don Raffaele Farina, Prefect, and with welcome also
by Don Jorge Maria Neja

Present (at one time or another):

Marco Buonocore (BAV)
Lou Burnard (TEI)
Francesco D'Aiuto (BAV)
Matthew Driscoll (MASTER)
Consuelo Dutschke (TEI)
Marco Jennarelli (Universit dell'Insubria)
Salvatore Lilla (BAV)
Antonio Manfredi (BAV)
Eva Nylander (TEI)
Don Cesare Pasini (Biblioteca Ambrosiana)
Ambrogio Piazzoni (TEI)
Merrilee Proffitt (TEI)
Paolo Vian (BAV)
Paul Weston (BAV)

Agenda
1.  Welcome
2.  Introductions
3.  Discussion of the proposed DTD
        3.1.  Decoration
        3.2.  Other asterisked points in order of occurrence
4.  Plans for writing documentation
5.  Coordination with MASTER
6.  Business
        6.1.  Reimbursement forms
        6.2.  Next TEI WG meeting


Minutes are a slightly edited version of the careful notes kept

 by MP during the discussions:  thank you, MP!!

5 March 1999

Decoration

Start with decoration because it's been so problematic.  PK lays out five
major breakdowns:
--miniatures
--initials
--borders
--secondary decoration
--diagrams
each of which can carry four attributes (discussion of their values
postponed):
--type
--size
--technique
--quality
with two optional subelements:
--figurative
--non-figurative

Discussion of quality as an attribute, with people perhaps feeling uneasy
about having a non-fixed list.  The same is true with "size" since it
could be measurements, or free text like "full page miniatures," or number
of lines for initials.

AP: divide between illustration and ornamentation.  Illustration is
ancient way of decorating, is part of the text, it says something.  The
other, ornamentation (such as initials) comes later.  Another way would be
to say two colors, or three colors, etc.

LB: "deep confusion between text of description and the thing being
described." "How to organize the prose." Dividing the prose, or
subdividing it, is a different problem than identifying components within
this (such as <artist>).  Identify decorative components <decorative> with
type = "miniatures" "initials" "borders" "secondaryDec" "diagrams" (in
other words, rather than the five elements above, treat them all as
attributes to the element <decoration>).

LB: bring back <iconTerm> from Oxford DTD (which can refer to a typology
such as Iconclass); general agreement to this.

AM: a list provided to cataloguer would be helpful.  LB: yes, we can have
that, but at what level.  So instead of subdividing at a higher level, one
would mark elements within the prose itself.

Separate list of things that could appear everywhere internal to <p> in
the "manuscript broth":
[this list in fact compiled gradually over the days of discussion but
assembled here at one point for convenience:]
<placeName> or subdivisions within this which we've talked about before?
Country, region, city, etc.
<date>
<dateRange>
<responsibility> who is responsible for something intellectually?
attribution by?
<bibl>
<msIdentifier>
<marginalia>
<dim>
<q>
<term>
<p>
<persName> with "role" attribute as to scribe/artist/binder
<orgName>

LB: <attribution> element which could contain <artist> <place> <date>
<certainty> BUT only if these things are ALWAYS together.

More on decoration: the attribute "technique" for example:  allow use
limited to one's own list?  But then what good is it for researchers in
general if everyone has their own list of attributes?

Come up with a list of things which should be access points.  These should
be either elements or attributes.

AP: stay close to the surface, because the deeper we go, the further we
can lead people astray.

What to do with rubrication within decoration?  This is only the material
or execution of the rubrication.  Other information about the rubrication
goes into information about the text.

95% of manuscripts are not decorated.

Rubrication will be dealt with as secondary decoration.

Decoration contains all decoration.  Decorated catchwords, for example, if
they are worthy of note, should be described here.


Physical Description

EN handed out her prepared list (see Appendix):
1.  Support
2.  Extent and foliation
3.  Dimensions
4.  Collation
5.  Catchwords
6.  Quire/leaf signatures
7.  Layout
8.  Binding

This was followed in the discussion in essential accord with her work
conceptually and structurally.  Some of the comments included:
--re n. 1, Support:
        1.  type and watermarks as attributes
2.  palimpsest should be listed under support (but their texts should be
discussed as separate parts)
--re n. 2, Foliation:
1.  separate from Extent and place under <Paratext> along with items such
as column numbers, Oxford line numbers, text finders, running heads
--re n. 4, Collation:
1.  proposed to structure the actual schema as
<collation><formula></formula></collation>
--re n. 5, Catchwords:
1.  attributes with fixed values
(horizontal/vertical/up-down/down-up/unknown with the last as default?)
for catchwords
2.  decoration on a catchword, if significant enough to be mention, should
be discussed in the            context of the Decoration, rather than
under Collation
--re n. 7, Layout:
        1.  handle number of columns/lines as attributes
        2.  <lineCount type=written> and/or <lineCount type=ruled>

LAYOUT from Eva's list:  pricking, position, technique, number of columns,
ruling, medium, ruled space, number of lines.
LB: Oxford DTD proposes <dim> with type=leaf, ruled, pricked, written.
Within this, elements height, width, depth(!?!), with attributes unit,
scope (all, must, range).
AP: grouping support with leaf dimension, etc.  <dim> perhaps should be
part of the broth.  This was decided to be a fine idea.
Attributes: columns, lines.  How to deal with lines?
LB:  <dim type="written"><height scope="most" units="lines">42</dim>
        MD:  <lines type="ruled/written">42</lines>

--re n. 8, Binding:
        1.  note that dimensions were proposed for thickness (as well as
height and width)
        2.  need indication of binder added to present list
        3.  add style/type of binding, tagged as <term>, e.g. legatura
alla greca, Grolier, a la fanfare
        4.  attributes of date/contemporary for the binding


Digression:  back to agenda in following through the order of the
description:

The Parts are changed to <addMsDesc>.  CP wants <msId> available here but
the option was rejected (again).

Use <form> rather than <format>.

As per later discussion, added <dim> as a separate element.

<marginalia> AP: an interesting access point, for example all the texts of
Augustine with marginalia.  Casual, not a glossed text (which is a
recognized text).  Can indicate provenance, can be illustration, can be
text.  MP: maybe this should be part of our broth.  After much discussion,
it winds up as broth or a brothy thing.


6 March 1999

<script>content=script type, prose description
        <scribe> content

Discussion of difference between <scribe> <artist> etc. and <persName
role="artist">.  We will return to this discussion.

LB:  <term> for flagging important terms,  such as <autograph>; discussion
of <autograph> as a concept.

LB: what is the difference between script, hand, and scribe.  We must
provide waterproof definitions for each of these.  AP proposes <script>
                                <hand>
                                        <scribe>

LB: reminds us of <handList>, <hand> used in the header (which can be
moved into text as part of physDesc, presumably?).  We can suggest changes
to existing definition, and use <hand> within <script>.  For example,  an
attribute named "style" might be better (and less ambiguously) named
"script."  This would allow us to combine <hand> and <scribe> in a
sensible way.

<music> contains prose.

[Given in summary above, but actually discussed on 6 March at this point]
<binding> [back to Eva's list] <dim> of course.  Add <width> to elements
within <dim>.  <binder>.  For types of bindings, LB suggests <term> again.
At this point the group seemed to catch on to what this meant and what the
possibilities really were.  For example, pulling all of the <term>s out of
a given description, collection, catalog to see what type of terminology
was being used.  Grolier example, given by PK.  Suggestion for <materials>
tag.  CWD: we aren't sure what binding experts really want, but date would
be one of the key access points.  LB: then perhaps date should be an
attribute with an absolute date value.  AP: contemporary also needs to be
an attribute.  LB: sure, but come up with definition for it.

<condition> perhaps this belongs in <adminInfo>.


Decoration

Should decoration be part of the physical description, or external to
this?  Description, according to PK, doesn't belong in physical
description because when there are images they are in a way texts.  EN:
how would this affect scholars who are searching and who are doing markup?
LB: it shouldn't affect them one way or another, but would be disastrous
to put, for example, technique in one place and content in another.
Decided in the end to leave it where it is for now <physDesc> but perhaps
this should be revisited later.


History (origins and provenance)

<origins> how to deal with dated, as opposed to datable.  LB suggests an
attribute (dated:  yes/no/unknown with default as "unknown").  Another
attribute on datable (also as yes/no/unknown, with default to "unknown").
Discussion of <date> and <dateRange> already available in TEI.  This group
will supply lots of examples for this.  LB suggests also <dateFrom> and
<dateTo>.

<place> as in TEI, we need to decide on hierarchy within this.  LB: unless
we can come up with suitable definitions for subparts, perhaps we
shouldn't divide this up.  CWD: but it's so important for people to be
able to find all Italian manuscripts of a certain period that we must have
this ability.  <country> as in TEI.  <city> is the same as <settlement>
which some people in this group think has a very bad name.  <institution>
isn't in there, but <org> is.  LB proposes a <placeName> element with
various type attributes.

Structured content within this?  So <place> first of all?  And then <p>?
CWD: troubled by treating these two things, date and place, in perceived
"different" ways, one as an attribute, and one as an element.  CWD also
troubled by "date" coming "first," because she would always localize
something before dating it.  AP does this in the reverse order.

Give up dates as attributes, structure instead as:

<history>
        <origin>
                <origPlace localized=y/n/u localizable=y/n/u>
                        <placeName type="city">Berkeley</placeName>
                </origPlace>
                <origDate dated= y/n/u datable= y/n/u dateFrom=xxx
dateTo=xxx>1999
</origDate>
        </origin>

Place and date above are both required, with the idea that searching for
what is unknown is interesting and not possible unless the tag is there
and the information is missing!

AP prefers term "unavailable" rather than "unknown"

<provenance> contains <p> which can contain <date> <dateRange> <placeName>
LB: put "dateFrom" and "dateTo" as attributes on <provenance> .  Shot
down.

Names of religious orders should be treated as <name>.

Suddenly, the conversation turns to msId and how to handle place names in
provenance vis a vis what's already in msId.  PK: useful to be able to
pull out previous or obsolete shelfmarks out of <provenance>.

New structure for <identification>

<summary>
        <country>
        <region>
        <msIdentifier>
                <city>
                <repository>
                <collection>
                <idNo>
        </msIdentifier>
        <nickName>
        <shortDesc> etc.
</summary>

and then <msIdentifier> can be pulled into <provenance> (note:  this is
the reason for removing country and region from <msIdentifier>).

Then, back to provenance.

<acquisition> PK wants to search in this for which lots transferred, at
what time, from where.  Can't count on this being the same as the last
provenance in all cases.


XZY

Audience attribute?  LB says something about security codes (?).

<additional>
        <p>
        <listBibl>
        <adminInfo>
<recordHistory> could include:   source/ name of cataloguer/  name of
institution/ reviser's name/ date. LB suggests a <change> element
(including respStmt from the header) here.
                        <change>
                        <source> with <p> which would allow for bibl.
<custodialHist> <p> or <date>, <event> with attributes, can include loans
and exhibitions, conservation, photography
<availability> from the header
        <surrogates> contains <bibl> <p>

Outstanding question: do printed facsimiles go in <surrogates> or in
<listBibl> under <additional>

We need examples of how to cite web pages in <bibl>

7 March 1999

MD had breakfast with LB; LB wasn't able to attend the day's session,
hence this "breakfast" report courtesy of MD:

Proposal to call surrogates refs (perhaps he means <xref>), and to call
spans <ref>s as Lou suggested in Paris.  The group felt that this was
ill-conceived and needed to be re-thought.

MD: proposes:  internal to <surrogate>:
<surrogateID type="microfilm"> <city> <institution> <repository>
<collection> <idNo>

We like this as above.

What to do about web pages, or electronic versions?  MP suggests:
<surrogate><p> blah blah blah <xref>http://www.mydoc.com/</xref></p>

On the issue of span:  If they are to point to something, then by all
means, mark them as <ref> or <xref>.  Otherwise they should just be text.
The folios or folio ranges are not the only places from which one might
wish to build links; one might wish to link from the title of a text or
description of a miniature to a transcription or detail in separate files.

Secundo Folio

Discussion of problem; this is not always literally a "secundo folio," but
this is a good enough name for it.  Those who follow other traditions will
be able to use this regardless of what this is called.  So an element
called <secFol> should have prose content.  AP suggests that it go within
<physDesc>; EN suggests that it goes in <history>; PK: suggests that it go
into <textDesc>.  After some discussion, there was agreemnt to put it at
the end of <physDesc>, before <binding>.


Accompanying material

MD:  Recounting of AM slips.  AP agrees that this is something that needs
to be catered for; he also has items that are loose in the manuscript and
he mentions them in his catalogues.  PK would put the AM slips as a
separate part, or perhaps as separate texts.  This seems weird since some
slips are bound in, and by PK's standards should be counted as parts, and
others are just sitting in the manuscript, and should be counted as
accompanying material.  We would like a tag called <accompMat> with
content of <p>.  Would contain description of the materials and perhaps
quotations.  How it is used is up to institutions.  MD proposes that it be
part of the physical description.  Distinctions between in the manuscript,
in the box.  Skull caps and handkerchief.  It was decided to put this
under physical description after binding.


addMsDesc (Note:  this term later changed to <msPart>)

The group was in accord that <addMsDesc>, as a formal distinguishing name,
should be an element, in order to eliminate any confusion in the case of a
composite manuscript (note:  this is in contradiction to an earlier
proposal whereby a strict sequencing of text-to-physical would necessarily
imply another part when one moved to a text after some physical
description).  <addMsDesc> contains: <textDesc> <physDesc> <history>
<additional> (excluding <summary>).  Discussion of whether <additional>
should be allowed within <addMsDesc> or excluded.  It was decided to keep
it.

<textDesc>
        <msText> was <msBibl> on Lou's list/Paris meeting

<author> issues here are: as in ms, supplied by others, supplied by the
cataloger.  Cataloguers present here are pretty clear that these are the
important issues.  Author is already a tag in the TEI; it has two
attributes, ID and lang, so we are going to propose the element as if it
doesn't exist.  If the attributes can be added to the existing element,
fine.  Can use <author><name reg=""> to deal with regularization, but what
about the rest?  <author type="as in ms, supplied in ms, externally
supplied, supplied by cataloger, regularized, other">  The reason for
adding "reg" to the proceding list is to deal with all of this in a
consistant manner.  This is apparently not consistant with use of other
regularized <name>s in the ms description, but whatever.

<title> we looked at TEI <title> and the "type" attribute.  Type=parallel
is interesting, says CWD.  We should use type attribute as above: "as in
ms, supplied in ms, externally supplied, supplied title (i.e.
regularized), uniform title".  This discussion was halted with the idea
that someone should run off and check AACR2 (probably for author also).

<persName> with role to identify: editor, translator, compiler, glossator,
corrector, annotator, etc.
The following passed without discussion:
<rubric>
<incipit>
<explicit>
<finalRubric>
<colophon>
<langUsage>
        <language>
<status> with type= "incomplete, begins defectively, ends defectively,
internally defective"
<listBibl>/<bibl>
</msText>
<textDesc>
Can a text have more than one incipit, for example incipit of the
prologue(s), incipit of main text?  One suggestion (AP) is to maybe have
each of these "subtexts" each as its own text with its own elements.
Another, more satisfatory solution is to allow these elements to repeat.

MD: wants a way to flag defective manuscripts.  The concept of a
<condition> element came up again.  Resolved by using <status> element.

Use <persName role="scribe, artist, binder, owner, etc."> for corpororate
entities <orgName> (<orgName> will need a role attribute).


Return of marginalia

Discussing once again, after reviewing everything in the broth, (EN) this
sticks out like a sore thumb.  MD would put <marginalia> in <physDesc>.
PK think this reflects mushy thinking on our part, as with <secFol>.
Perhaps this only stands out because there are not similar items in the
broth, such as heraldry, which we haven't dealt with yet, or <secFol>
which might need to be revisited.  We were getting pretty tired by this
point.


Documentation responsibilities

Decoration, Summary: Peter
TextDesc: Ambrogio
Script, SecFol, Binding, AccMat, History: Consuelo
Additional: Merrilee
PhysDesc as far as Music included (except Script): Eva

Deadline: 30 April.


Next meeting

mid-late October in Berkeley (note:  Manuscripta conference in St. Louis
on 8-9 October 1999; Medieval Guild at Columbia on 23 October).


History of the discussions
1.  Studley Priory, November 1996
2.  EAMMS, Collegeville MN, December 1996
3.  EAMMS + DS + SagaNet, New York, November 1997
4.  TEI WG, New York, July 1998
5.  TEI WG at Claremont Conversations + e-mail discussion, New York,
January 1999
6.  MASTER, Paris, February 1999
7.  TEI WG, Rome, March 1999

Appendix 1:  EN list for codicology discussion

Appendix 2:  form of dtd at the end of the Rome meeting
