You are on page 1of 11

Caitlin Corbett

May 8th, 2016


Haunted Camera: Spirit Photography Digital Collection

Part I. Collection description and its information objects:


1.1 Location of collection
This collection is wholly digital in nature. It is comprised of digitized versions of photographs,
and as such, the collection may be accessed anywhere there is a stable Internet connection. The
items that have been digitized originally resided in locations such as museums and personal
collections. The URL for visiting this collection in its present incarnation is as follows:
http://hauntedcamera.omeka.net/
1.2 Purpose of the Collection
This collection is designed for people with interests and research needs in the history of
photography, as well as the paranormal, and even simply the history of the Victorian era.
Researchers of photographic history would find such a collection useful and important for
examples of early camera use, and associated techniques.
It is also a helpful solution for those who do not have physical access to the photographs. This is
especially true as many of the items originate from varied sources and exist physically in
different locations, including within personal collections which are otherwise inaccessible to the
public.
1.3 Physical and intellectual aspects of objects
This collections physical format is digital photographs, and exist in file formats such as JPG and
PNG. In terms of intellectual aspects, the photographs possess textual metadata, including
descriptions and provenance information such as featured subjects and location.
1.4 Extent of the collection and plans for growth
This collection presently contains 1,100 such digital items, and this number has remained stable
in the months since it was established. Acquisitions made from as yet unknown private donors
may cause the collection to grow. But as museums and collectors across the country and the
world have been invited to participate in the digitizing and contributing of such items, it is not
expected that the collection will expand further in any considerable way.

Part II. Users of the Collection


2.1 User group(s)
Primary users of this collection would be researchers of the history of photography. The items
are grouped by photographer, and also tagged with metadata components such as the century of
origin. Such images are seen as quirky artifacts of modern history, with the advent of the camera

Corbett/p.2
and ensuing rise of both private and commercial photography. This collections items would
likewise serve as useful points of comparison with more modern or contemporary examples of
ghost photography, often popularized by television shows and films. Also among primary users
would be researchers who are simultaneously interested in the history of 19th century
Spiritualism and associated hoaxes.
Secondary users would include those with a more general interest in the 19th and early 20th
century historical periods. The presentation and metadata is structured to serve users both casual
and more intense or practiced. It is designed so that almost anyone who has interest in a
particular photo or set of photos could follow up on the items with the given information on
provenance.
2.2 Demographics
The users of this collection are representative of both genders, and tend to be varied along age
lines (between 18 and 70 years old). Site statistics indicate that most of the users are visiting
from United States IP addresses, with some originating in the United Kingdom as well. A large
segment of users also appears to be educated, indicated by their academic usage of the collection.
They are computer literate, which points to home or work use of related technology.
2.3 Types and levels of knowledge

Domain knowledge: All of the users of the collection are familiar with the existence and
basic aspects of historical spirit photography. They understand that the photographs were
created by multiple individuals, across different decades. Some users, such as
researchers, may visit the collection with a specific intent to focus on selected creators.
For this reason, the collection is organized and searchable by creator.

System knowledge: All users of this collection have a functional understanding of online
search procedures and mechanisms, including search boxes, links, and pagination. These
users demonstrate fundamental computer literacy.

Task knowledge: Most of the collections users will visit it with well-established goals
for information retrieval. The researchers of historical photography will likely be more
readily familiar with the items creators, as well as their creative methods. Their interest
may well be in finding other photographs with identical provenance, or in comparisons
with works of other creators. Users interested in 19th and early 20th century Spiritualism,
and the noted era at large, may have less defined search tasks in mind, perhaps being
more inclined to casually browse the collection for items that appeal to their curiosity.

World knowledge: Users of this collection tend to have at least some formal education,
with many being more highly educated (with college degrees). The primary users are
researchers of the history of photography and as such, they are routinely exposed to
nuances of culture and time period. Secondary users, as well, are aware of distinguishing
details of their historical era of focus. The fundamental realization of such photographys
existence indicates a broader awareness of human history and culture.

Corbett/p.3

2.4 Users problems and questions


While mostly unified in their seeking of the items for their very particular characteristics as
historical photographs, the primary and secondary user subgroups are delineated by the end uses
of their searches. The primary users are researchers of the history of photography, and as such,
they look for information to not only identify and link the items amid different creators, but they
seek commonalities of technique and creation context. A researcher might be very interested to
find the exhibit of all known works by William Mumler, who is considered by many to be the
father of spirit photography. Researchers might then seek to compare his techniques with those
of William Hope. Research-oriented users are also looking for points of study in the history of
19th century Spiritualism and associated hoaxes, as which the collections items most definitely
serve. These users are looking for dates of creation, location, specifically featured photographic
subjects, and other basic points for contextual reference. Beyond the more precise needs of
scholarly users, casual users may be interested merely in the sensational quality of the items, and
are often satisfied merely browsing the collection without looking for anything in particular.
They simply view the available items, and may note metadata of objects that especially appeal to
their interests.
User Question: How many spirit photographs did Mumler create in the late 1860s?
Suggested Attributes: Date, Creator, Subject
Desired Precision/Recall: High precision, high recall
User Question: Ive seen some old photographs featuring Lincolns ghost, but dont remember
who the photographer was. How can I find that out?
Suggested Attributes: Creator, Subject
Desired Precision/Recall: High precision, moderate recall

Part III. System design:


3.1 Entity level or unit of analysis and why appropriate
The entity level for this collection will be each digitized photograph. The items feature their own
body of metadata that links to other item records, and allows for users to sort more easily through
the collection (ie. by date, creator, and subject). This facilitates both research endeavors and
casual browsing, as well as renders easier any referral or later retrieval. The photographs
themselves are commonly organized at this same entity level among other collections.
3.2 Attributes chosen and their appropriateness to the collection

Name: Serves as the functioning title of the item, which enhances its findability and its
identification. In most cases, the title is pre-established, and carried over from the digital
items physical counterpart.

Corbett/p.4
Also indicates the smaller collection(s) to which the item belongs. This curation is helpful
to both researchers and casual browsers.

Subject: Notes the specific people, locations, etc. that are featured in the given image.
This allows for additionally targeted searching capability.
Also includes details about the contents of the photograph, such as what is apparently
occurring or being depicted. Provides basic contextual notes.

Creator: Indicates the photographer responsible for making and publishing the original
photograph from which the digital item is derived. This attribute is particularly essential
for researchers.

Date: Indicates the year of the photographs creation and/or publication. This is another
attribute vital to researchers and any others wishing to pursue study of the items.

Format: Indicates the items image file format (JPG, PNG, etc). This is primarily for
system administrators.

Classification: Provides the items unique numeric identifier. It also provides the actual
image file for the item.

3.3 Field names and indexing decisions


General Attribute

Field Name

Searchable? (y or n)

Name

Title

Yes

Name

Sub-Collection

Yes

Subject

Subject

Yes

Subject

Description

Yes

Creator

Photographer

Yes

Date

Date

Yes

Format

File Type

No

Classification

ID #

Yes

Classification

File

No

This system is designed to index as many fields as possible for robust and flexible searching on
the part of users with varying intents and familiarity with the items. As such, it will index all but
two of the fields. Researchers, in particular, require the ability to search titles, subjects,
photographers, and dates. These users, in addition to those with more casual purposes, will also

Corbett/p.5
be able to search by sub-collection, which, for instance, organizes and groups items by
photographers (ie. Mumler sub-collection, Hope sub-collection). Also of general purpose will be
searchability of item description fields. This process will utilize keywords, and help lead users to
items for which they have very little identifying information. File Type and File are the only two
fields that will not be searchable. The File Type is considered administrative data, and does not
bear any implications for researchers use of the collection. The File field contains the image
itself, and therefore will not be searchable.

Part IV: Semantics, Syntax, and Input Rules


4.1 Field Names and Semantics:
Field Name

Semantics

Title

The name of the item.

Sub-Collection

The separate sub-group into which the item falls, if


applicable.

Subject

Specifically identified people, time periods, or scenes in


the image.

Description

A detailed, but concise overview of the image; inclusive of


notes about physical appearance, such as handwritten text.

Photographer

The creator of the original image.

Date

Year, or year range of original photographs creation


and/or publication.

File Type

The image file format.

ID #

The items unique numerical identifier.

File

The items image file.

4.2 Record content and input rules


Field Name: Title
Semantics: The name of the item.
Chief Source of Information: Field information comes from the given title of the
original photographic item. In the event that there is no pre-existing title for the
physical item, the cataloger will assign one.
Input Rules: Capitalize the first words of the title, as well as all proper names;
& should not be used for and; do not include concluding punctuation; title
should not be a question; title should be no more than ten words in length. Do not

Corbett/p.6
wrap the title in quotation marks.
Example: P.T. Barnum and Lincoln
Lincoln's Ghost
Field Name: Sub-Collection
Semantics: The separate sub-group into which the item falls, if applicable.
Chief Source of Information: This field is assigned by the cataloger, as
applicable.
Input Rules: Enter the full name of the photographer or theme, capitalized
appropriately; follow with Collection, also capitalized. This field is repeatable.
Example: William Hope Collection
William H. Mumler Collection
Field Name: Subject
Semantics: Specifically identified people, time periods, or scenes in the image.
Chief Source of Information: Field reflects the people, objects, themes, or
scenes depicted by the image. In most cases, this information will be adopted
from the record of the original photograph. Consult the Library of Congress
Subject Headings and LC Name Authority File respectively for the terms and
names to be used in this field.
Input Rules: For identified people featured in the image, enter last name first,
comma, first name and middle name, comma. Follow this with a double-dash and
the years associated with the individual as given by the Library of Congress. If no
dates are known, then omit entry content after the first/middle name. Utilize up to
a maximum of five separate subject headings. Terms should each occupy a single
line. This field is repeatable, so multiple terms should be stacked, and reflect
alphabetization. If a desired term is not available within the LCSH offerings, or a
featured person lacks an LC Name Authority file, one may add it to the field, and
the record will be automatically shadowed and flagged as pending by the
system, and immediately subject to review. Subject analysis of each item should
be conducted at a basic level; people, places, and things identifiable and otherwise
viewable should be considered in the subject searching. The item should be
viewed holistically, and terms that summarize the main focus should be used.
Example: Barnum, Phineas Taylor "P.T.",--1810-1891
Lincoln, Abraham,--1809-1865
Field Name: Description
Semantics: A detailed, but concise overview of the image; inclusive of notes
about physical appearance, such as handwritten text.
Chief Source of Information: Field information is generated by the cataloger at
the time of record creation.
Input Rules: Capitalize the first word of each sentence, and each proper name;
field should contain no more than eight complete sentences; field should not

Corbett/p.7
include abbreviations.
Example: Photograph purports to show Barnum accompanied by the spirit of
Abraham Lincoln.
Photograph purportedly taken at the annual meeting of the "Society for
the Study of Supernormal Pictures" depicts a spectral mist; Sir Arthur Conan
Doyle and his wife are seated center-left.
Field Name: Photographer
Semantics: The creator of the original image.
Chief Source of Information: Field information comes from the holder of the
original photographic item.
Input Rules: Enter the full name; capitalize the first letter of each part of the
name, including middle initial if applicable; do not abbreviate any names (ie. do
not use Wm for William). Utilize the spelling provided by the LC Name
Authority File. If a company rather than a person is given as the photographer,
leave off any preceding articles such as The.
Example: William H. Mumler
William Hope
Field Name: Date
Semantics: Year of original photographs creation and/or publication.
Chief Source of Information: Field information comes from the original
photographs provenance data.
Input Rules: If a precise year is known, enter it in its full four-digit form. If only
an approximation is given, include ca. for circa, and follow with the year. If a
year range is given, then enter ca., space, followed by the four-digit form of the
starting year, hyphen, and then the closing year in four-digit form. Do not enter
data in this field more specific than the annual level, even if it is available. If this
information is considered critical, it should be entered in the Description field.
Example: ca. 1868-1872
1869
Field Name: File Type
Semantics: The image file format.
Chief Source of Information: Field information reflects the file type of a given
item. The field itself only reports this for the benefit of database administrators.
Input Rules: Enter image/ as the prefix, then follow with the three-letter
abbreviation for the given file type; do not spell out any file type, only utilize the
three-letter abbreviations.
Example: image/jpg
image/png

Corbett/p.8
Field Name: ID #
Semantics: The items unique numerical identifier.
Chief Source of Information: This field is generated by the cataloger at the time
of record creation.
Input Rules: This is a faceted scheme, and non-repeating. The system will assign
the item a starting number (prefixed with 0 for all numbers under 100) that
reflects its order of placement in the database; the cataloger needs to assign the
last four digits of the ID #, which should be the items year of creation/publication
found in the Date field; if only a year range is given, then enter the starting year
in that range. This, paired with the starting numbers, preserves the unique aspect
of the entire ID number even if the last four digits may be shared by multiple
items.
Example: 006-1869
010-1920
Field Name: File
Semantics: The items image file.
Chief Source of Information: This is uploaded and attached to the record by
database staff.
Input Rules: This is entered by database staff. This field is repeatable, to allow
for enhanced or alternate versions of the digitized photograph. However, most
records will only feature one image file.
Example: N/A

Part V: Records for your objects


Title

P.T. Barnum and Abraham Lincoln

Sub-Collection

William H. Mumler Collection

Subject

Barnum, Phineas Taylor "P.T.",--1810-1891


Lincoln, Abraham,--1809-1865
Ghosts

Description

Photograph purports to show Barnum accompanied by the


spirit of Abraham Lincoln.

Photographer

William H. Mumler

Date

1869

File Type

image/jpg

ID #

005-1869

Corbett/p.9
File

mumler_barnum_large.jpg

Title

Lincolns Ghost

Sub-Collection

William H. Mumler Collection

Subject

Lincoln, Abraham,--1809-1865
Lincoln, Mary Todd,--1818-1882
Ghosts

Description

Photograph depicts the ghost of Abraham Lincoln with his


widow, Mary Todd Lincoln.

Photographer

William H. Mumler

Date

ca. 1869

File Type

image/jpg

ID #

006-1869

File

Mumler_Lincoln.jpg

Title

Meeting Spirit

Sub-Collection

William Hope Collection

Subject

Doyle, Arthur Conan, 1859-1930


Ghosts

Description

Photograph purportedly taken at the annual meeting of the


"Society for the Study of Supernormal Pictures" depicts a
spectral mist. Sir Arthur Conan Doyle and his wife are
seated center-left.

Photographer

William Hope

Date

1922

File Type

image/jpg

ID #

024-1922

File

hope_001.jpg

Corbett/p.10

Title

Sance

Sub-Collection

William Hope Collection

Subject

Ghosts

Description

Photograph shows a sance in progress, with a spectral


hand reaching up through the floor.

Photographer

William Hope

Date

ca. 1920

File Type

image/jpg

ID #

011-1920

File

Hope_Seance.jpg

Title

Unidentified subject with two spirits

Sub-Collection

William H. Mumler Collection

Subject

Ghosts

Description

Photograph purports to show an unidentified man


accompanied by two spirits.

Photographer

William H. Mumler

Date

ca. 1870

File Type

image/jpg

ID #

020-1870

File

Mumler_(unidentified).jpg

Part VI: Project Summary


This project has been an interesting exercise, in a variety of ways. To begin with the
positive aspects, I chose to (re)organize this particular collection because of the inherent
challenge it brought. Spirit photography has always been an interest of mine, as well as
photography in a more general sense. So working with these items seemed like a fun

Corbett/p.11
undertaking. And it was both fun and challenging. It compelled me to think about the users of
this collection, what their needs would be, and how they might explore the collection depending
upon what they required from it. That naturally led to contemplation of metadata to include, and
the granularity of the data. While I did not include the ability for users to free tag in this design
iteration, it is something that I think would benefit a highly visual collection like this. The
crowd-sourcing possibilities are elements worthy of pursuit in future redesigns of this collection,
not least of all due to the sometimes mysterious quality of the images and the people depicted
within them. User-generated metadata would be useful for filling out those kinds of gaps in
subject and even provenance. For these reasons, I enjoyed the opening portions of the project,
which asked for descriptions of users, perceived needs, and ways that my system could meet
those needs. Building context in that way really lays the foundation for the fine-tuning tasks later
in the process.
While I certainly would not want to characterize any part of this process as especially
negative, I would surely say that some portions were more difficult than others to think
through and implement. For instance, defining the input rules really required me to think on a
much more critical level about everything I was doing. I had used Omeka to put the host
collection together initially, and that system was of course more flexible in terms of how I
wanted to represent the metadata. For this assignment, I had to identify and become somewhat
more familiar with all of the sources I needed to consult for field content entry. I had believed
going into this project that the Library of Congress would have all of my necessary subject and
name authority bases covered. But I eventually found that many of the items I was working with
featured people unknown to those resources, and according to my hypothetical systems rules,
many of the records would be flagged when the cataloger desired to enter new information for
these missing portions. Having to account for new subject heading requests was not something I
had thought enough about prior to this assignment. It showed me the limitations of such
resources, even though they are incredibly comprehensive. Our discussion of standards and the
often exhaustive review process they require were things I considered when making provisions
for new headings. My system uses LCSH, but when even that controlled vocabulary is missing
things, the next step is not quite so clear-cut. One thought I am having for future iterations is the
generation of a system-specific controlled vocabulary, independent of LCSH and others. The
arguably niche nature of the collection might justify that on some level.
Another area of some difficulty was keeping attributes and actual fields separate in my
mind. That was one portion that needed revision for the final product. Looking back, I
understand what makes them different, and how they are respectively used. Essentially, I learned
how easy it can be to become mixed up with the terminology and concepts, especially since
cataloging at almost any level is a multifaceted process. The complexities caused by specificity
can be daunting. But this is all the more reason to consider closely the collection items and how
they can be represented in clear and meaningful ways.

You might also like