Category Archives: digital collection

alexander architectural archives, digital collection

Linking Architectural GeoData

Over the past several years, a team at UT Libraries has been developing the UTLARCH GeoData project, a prototype for a new kind of geospatial database that can map and link together data derived from various collections at the Alexander Architectural Archives. In this post, I will summarize our work to date.

View the UTLARCH GeoData project on ArcGIS Online

The core project team has consisted of myself, Josh Conrad, a graduate research assistant, along with Katie Pierce Meyer, Michael Shensky, and Jessica Trelogan. We have also relied heavily on the amazing data work from UT Libraries staff and students including Grace Hanson, Irene Lule, Alyssa Anderson, Abigail Norris, Katie Jakovich, Stephanie Tiedeken, Beth Dodd and Nancy Sparrow.

Buildings of Texas

The project began in 2018 after the Alexander Architectural Archives received a new collection donation from architect and historian Gerald Moorhead containing the research and editing material for the two volume book series Buildings of Texas, published in 2013 and 2019. This donation consisted of over 20 boxes of paperwork, 12,000 photographs, and, importantly for our project, a series of eight Excel spreadsheets cataloging over 5,000 places represented in the book series. These spreadsheets included addresses, place names, dates of construction and renovation events, as well as architects and other contributors to these building events.

View the Buildings of Texas Collection finding aid on TARO

View the original Buildings of Texas Excel spreadsheets in the Texas Data Repository

The Buildings of Texas spreadsheets provide a wealth of data about a wide range of architectural projects throughout the state, including many projects represented in other AAA collections. The AAA projects database and finding aids also store a lot of information about the various architectural projects represented in the collections, but many gaps exist in AAA data, especially geographic information such as addresses and coordinates which were not included on donated items such as drawing sets. With the acquisition of the Buildings of Texas datasets, the intriguing possibility arose of filling in these gaps by linking Buildings of Texas data to existing AAA projects data.

From this seed of an idea, the project concept expanded: why not connect geodata from other AAA collections as well, such as the recently digitized David Williams photography collection and Atlantic Terra Cotta Company photography collection, both now hosted on the UT Libraries Collections portal. Numerous other un-digitized architectural datasets also exist in AAA collections which in the future may be able to be included in this growing linked database, such as the Texas Architectural Survey collection and the historical surveys conducted by architects and historians such as Eugene George and Wayne Bell. Could we include data from outside of the AAA as well?

Linked data model

These early brainstorms coalesced around a plan for new kind of geospatial database that could store and display place data from various linked datasets together on a map. We saw this concept as akin to pre-digital geographic indices such as atlases and gazetteers. We also envisioned this database as a kind of spatial finding aid that researchers could use to find all the items about a place such as photographs and drawings that may be held in multiple collections. Further, if this potential geodatabase will be digital, we could use it to develop map-based digital exhibitions and other kinds of library discovery tools.

We explored a variety of database models for representing this diversity of data. A single data table would not be sufficient. Places represented in AAA collections are complex and involve multitudes of contributing individuals. Our data model would need to be flexible and allow one-to-many relationships throughout. In architectural data, places are often associated with a number of specific events in time such as dates of construction and renovation. Further, each event is often associated with multiple different people and companies such as clients, architects, builders, and other contributors. And, to add further complexity, events are documented by multiple different artifacts, such as photographs and drawings. We wanted to develop a spatial database that could link people and artifacts to places while also being able to store temporal data that could contextualize them within the history of that place.

Below is the Entity Relationship Diagram (ERD) of our prototype data model. Centrally, the core table is the Places table. Places represented in this table are labeled with the name of the collection from which the data originates (place_source_term). In order to map the data with a geographic information system (GIS), we connect each place to its geographic representation, a point, line or polygon located somewhere on earth. The three tables at the top (in red) are the spatial tables, or “feature classes” in GIS lingo, that are required by the GIS. One feature class represents points, one represents lines and one can represent polygons. Place data is often represented with a point, essentially a map pin located on or near the place’s location on earth, but places can also be represented with lines, in the case of roads or trails, or polygons in the case of building outlines, land parcel boundaries, and neighborhood boundaries.

The core places table is then connected in a one-to-many relationship to various attribute tables (in blue) that can store multiple values per place. In the Place Names table, places can have many names, such as current and historic names as well as addresses. In the Place Type table, places can have many different types or uses, such as residences, commercial buildings, public buildings, and landscapes such as parks and neighborhoods. Further, places can be sites of multiple events over time, including but not limited to construction and renovation events, documentation events such as the photographing and drawing of that place, and cultural events such as celebrations and commemorations. Thus, the Events table, as our model’s temporal table, stores the dates and event data associated with each place. From the Events table are connected two more attribute tables: the Contributors table, which stores names of people and companies associated with each event; and the Artifacts table, which stores the data about collection items associated with each event.

Within this spatial and temporal data model, multiple data sources can be combined. However, if two or more sources represent the same place, that place will be represented by multiple unconnected records in the places table, one record for each source. Ultimately we want to be able to match these records so that researchers can view all the data available about each place. Preliminary ideas from our team included the idea of creating a “Unique Places” table where matching places could link to a single shared place record, essentially an authority record that could store a standardized geographic representation for that place. We would also want to create an authoritative “Unique Contributors” table.

The prospect of creating our own authority files brought up the question of how we might work with, or contribute to, existing authority file databases such as Getty Vocabularies, who manages two authority ontologies relevant to us, the Art and Architecture Thesaurus (AAT) and the Union List of Artist Names (ULAN). Further, as a means to broaden representation and allow for multivocality and multiplicity, we felt that we should also contribute our authority records to the linked open data community via Wikidata.

Wikidata

I usually think of Wikidata as the database version of Wikipedia. Similarly to Wikipedia, Wikidata is an open collaboratively-edited platform developed by the Wikimedia Foundation using the CC0 public domain license. But whereas Wikipedia is composed of records of various entities and concepts written in an encyclopediatic narrative prose, Wikidata represents unique entities as structured database records with a data model that they call a “knowledge graph”. A graph data model, which I have also heard called a linked data model, a network model, an ontology model, or an object-oriented model, is an alternative data model used by encyclopediatic knowledge databases and newer authority file systems such as Getty Vocabularies, the Library of Congress Linked Data Service, the Virtual International Authority File, the CIDOC Conceptual Reference Model, and Wikidata. In a graph model, data is stored, not in fixed table rows like in traditional relational databases, but rather as a series of “triples” composed of two objects (or “nodes”) and their relationship (or “edge”), and stored in a specialized databases called triplestores. In this more-flexible structure, each unique object can have many different relationships with many other objects. In Wikidata, item records are able to have many different attributes which are each linked to other unique items. The result is a flexible hyperlinked database. The image below shows an example of a Wikidata record containing an attribute that is in turn linked to another Wikidata record, forming a network.

Because Getty Vocabularies and Wikidata both use a linked data model, we are able to match and link records from our different data sources by connecting, or “reconciling”, each record to its authoritative Getty or Wikidata record. However, because Wikidata offers an open-source collaboratively-edited platform, if a Wikidata record for a place or contributor does not exist, we can simply create it on the fly using the Wikidata reconciliation API and Open Refine. This important feature distinguishes it from Getty Vocabularies which is not a publicly editable database (though they do accept new data through partnerships with libraries). Because it allows us to directly manage records, Wikidata can essentially become integrated into our data infrastructure and workflow. In other words, Wikidata can simply become our authority file. In this configuration, our data model is greatly simplified. As shown the ERD above, instead of creating new “unique” tables, the Places and Contributors tables both contain attributes labelled wikidata_qid in which we simply store the Wikidata identifiers (QIDs) that represent each authority record.

Database infrastructure

Because our ultimate goal is to create a geospatial database, we required a GIS database infrastructure. A critical component of this project, one which really made everything possible, was our ability to use the ArcGIS geodatabase and server that UT Libraries ITS has recently installed to support the creation of the Texas GeoData Portal. UTL ITS has been providing the libraries with an incredible data infrastructure from which to build new innovative kinds of discovery, access and digital preservation tools. We are also forever grateful to UT Libraries GIS experts Michael Shensky and Jessica Trelogan for being a part of our core team and teaching us how to navigate the complex matrix of settings and workflows required to host data within this infrastructure.

The GIS infrastructure is essentially composed of three components: a ArcGIS enterprise geodatabase built with Microsoft SQL Server, the ArcGIS Server, and ArcGIS Online. Data tables created on the SQL server are linked to a GIS feature service which can be then added to an ArcGIS Online web map as a set of GIS feature class and tables. Any updates made over time to the core data tables are then automatically updated in the web map.

In order to populate and update the geodatabase over time, our team established a documented data processing workflow that extracts the original data source, transforms it into our linked data model (including reconciling with Wikidata), and loads it into our database infrastructure. This Extract-Transform-Load workflow involves the combined use of the data processing tool Open Refine as well as ArcGIS Pro and can be applied to new datasets in the future.

Results

To date, we have uploaded three datasets into the new geodatabase: the Buildings of Texas Collection, as well as two datasets exported from the AAA projects database representing the Karl Kamrath Collection and the Volz & Associates, Inc. Collection. As part of this process, we have created nearly 7,000 new Wikidata records representing both people and places.

Below is a screenshot of our ArcGIS Online web map and interactive data dashboard. The image shows an example of a house represented in two separate datasets, linked spatially by having the same point coordinates as stored in Wikidata. This house, located at 8 Tiel Way in Houston, was designed by Karl Kamrath. The Kamrath dataset provides data about the drawings he produced for the project which we represent in the geodatabase as “drawing” events using the date at which the house was drawn (as recorded on the drawings). The Buildings of Texas dataset includes data about the home’s construction, which we represent as a separate “construction” event. Together the two events begin to provide a more temporal depiction of this place than each dataset does alone.

The two datasets from the AAA projects database provided us with a workflow template for adding other AAA projects datasets in the future. Digitized datasets extracted from UT Libraries Collections portal metadata are next. Over time we hope to continue refining and filling in the gaps in each of these data sources by developing new editing workflows using the interactive tools in ArcGIS Online. As both the Wikidata and ArcGIS platforms improve over time, we hope to find new ways to link the two more intimately. Currently, for example, ArcGIS Online is not able to asynchronously query external APIs such as Wikidata, so we must continue to store authority data locally. ArcGIS developers has also started integrating graph data models into their products which may allow us to migrate from our relational data model and more-fully embrace linked data practices.

We are excited to be a part of these developing technologies and hope to continue experimenting with new ways of sharing and linking architectural data. Be sure to visit our project dashboard as well as our project StoryMap which explores how linked architectural data can be useful for new research.

alexander architectural archives, digital collection

Mission Possible

What does an archivist do when the archives are closed?

Mission Espíritu Santo de Zuniga restoration drawing, 1933

Along with institutions across the globe, the Alexander Architectural Archives and Architecture and Planning Library closed our doors in March for the safety of students, staff, and visitors. Usually we spend our days surrounded by the Alexander’s collection of papers and books at Battle Hall, physically tied to the materials we work with. While we miss the stacks (and each other!), the transition to remote-work presented the perfect opportunity to focus our attention on digital projects at the archives. 

Digital work gave us something to keep our hands full, but, more importantly, it gave us a way to continue to make archival resources available to researchers, also facing the challenges of remote work. Contributing to the University of Texas Libraries’ Collections Portal offered the best of both worlds. This recently launched Digital Asset Management System (DAMS) provides a platform for UT repositories to store and describe their digital materials, and share them with the public.

Screenshot of the collections portal

Luckily, this isn’t our first rodeo when it comes to digital projects at the Alexander. For as long as digital technologies have existed, the archival profession has been brainstorming ways to both preserve these technologies and take advantage of them to make archival materials more accessible. In 1999, the Alexander launched an online digital exhibit titled, Spanish Colonial Architecture as Represented in the Alexander Architectural Archive, described as “a collection of digitized drawings and photographs of eighteenth-century Spanish missions and [the] San Antonio’s Governor’s Palace. These records, dating mainly from the 1920s through the 1950s, reflect how the structures looked before various efforts of restoration and reconstruction. Selected from collections within the Alexander Architectural Archive, they include works by Harvey P. Smith, Stewart King, Ayres and Ayres, Robert Leon White, and measured drawings by UT Austin School of Architecture students in the Texas Architecture Archive”.

Twenty years later, this collection of digital scans is gaining new life on the Collections Portal. In my role as a Graduate Research Assistant at the Alexander, I’ve been immersed in a world of arches and colonnades, uploading assets from the Missions exhibit into the DAMS and adding descriptive metadata.

Mission San Francisco de la Espada, drawing of patio

Following UT’s shutdown, the staff at the Alexander shifted our focus to remote projects, including working in the DAMS to ingest an array of items, from floorplans to historic photographs. With over 2,000 combined items from the Alexander Archives and Library, our collection is only growing. As researchers continue to limit travel and conduct their work remotely, access to online resources has become more important than ever. Even as UT begins to implement new safety strategies on campus and we consider plans for reopening the archive, digital portals will remain a vital access point and a way for both institutions and researchers to see archival materials in a new light. When it comes to the digital, we as archivists are in it for the long haul, rain or shine, just like our commitment to preserving traditional paper artifacts. 

We look forward to welcoming you back to Battle Hall in the future, and in the meantime, we hope you’ll enjoy browsing our materials virtually via the Collections Portal.

Sources:
digital collection, special collections

UPDATE: Southern Architect and Building News Digitization Project

2 Comments

Hello, People of the Blogosphere! We figured it was time to provide an update on the efforts to digitize our copies of the journal Southern Architect and Building News. The bottom line: all our issues of Southern Architect are now available on the new University of Texas at Austin Collections website! There are still some kinks being worked out and changes being made to the site, but there are now hundreds of items available from institutions across campus for you to peruse.

One thing worth highlighting about the site is the sheer quality of the images. Particularly noteworthy is that the images lose none of their quality when users zoom in. Not only that, but now you can invert images to be negative, adjust saturation and brightness, rotate the image, and more.

Screenshot of the University of Texas at Austin Collections website, showing an unedited image
This is the same page in the journal as pictured above, only zoomed in on the images and with the colors inverted so that the images of the plans are clearer.

Additionally, you can overlay the metadata for the issue over the image itself, which is a very useful feature when you are in the midst of reading the journal or looking at the images. This means you can see all the information you might need without having to abandon the image you are on or losing any edits you might have made to the image.

Here you can see the overlay of metadata over the image.

There are also several different ways of viewing the images, including scroll view and book view. This allows you to choose the format that best fits your needs, whether you want to read the article as if the journal was in your hands (book view), or see numerous images at a time and essentially flip through the whole journal several pages at a time (scroll view).

Book view
Scroll view

There’s tons to explore on this website. There’s also still kinks to be worked out. But please go and look at these amazing online collections from us, the Alexander Architectural Archives, and our friends around UT Austin!

archive, digital collection

Disk ≠ Image: Creative Solutions to Issues in Born-Digital Processing

Hello. It’s Abbie again, back with another installment on processing and publishing the Volz & Associates, Inc. collection. I recently had the opportunity to present on this collection at The University of Texas’ 2019 Digital Preservation Symposium, where I discussed some of my solutions to born-digital access and preservation issues.

In the course of processing this collection, some of the most significant roadblocks were related to problematic disk images and metadata generation. These revolved around two distinct media types: zip disks and CDs. The collection contains 85 inventoried zip disks. When these were imaged and analyzed in BitCurator, the report came back riddled with error messages. Even though the disk images had been generated successfully, something was keeping fiwalk from generating the correct metadata. This made it impossible to determine necessary information like disk extent, content, and the presence of possible PII or malware.

Blog1
Results for the zip disk analysis. Not what you want to see!

When processing CDs, I noticed a group of about a dozen disks each had the exact same metadata. While some similarities are expected, I thought it was strange that that many disks from different projects could each have identical specs. I analyzed the disk images and physical disks and found that they were all CDs of photos processed by the same company. The original content of the disks had been overwritten by the company’s built-in structural and technical metadata, making the disk images virtually useless for research purposes. However, when these disks were mounted in FTK Imager, a write-blocked environment, it became clear that the original images were still extant – they just weren’t being picked up in the disk images.

Blog2
Optical disks with identical metadata.

In both of these cases, I knew that the original content could be reached but didn’t know how to access it while adhering to the best practices set up by both the digital archiving community and the resources at my disposal. The disks with these issues amounted to nearly 1/5 of the imaged collection, and many hours of research and testing were put into how I could reprocess this information. Ultimately, I found myself faced with a choice – I could either preserve inaccurate representations of the disk by retaining the existing disk images or preserve inaccurate representations of the metadata by extracting extant files instead of preserving a disk image. The issue with preserving the metadata as it stood was that it was wrong to begin with, and the resources at hand didn’t provide a solution.

When thinking about how best to reprocess this information, I was drawn to this quote from a study undertaken by Julia Kim, the Digital Assets Manager at the Library of Congress:

“Most of the researchers emphasized that if it came to partially processed files or emulations and a significant time delay in processing, they would take unprocessed and relatively inauthentic files. Access by any means, and ease of access were stressed by the majority.”

“Researcher Interactions with Born-Digital: Out of the Frying Pan and into the Reading Room,” saaers.wordpress.com

While we certainly don’t want to prioritize patron opinion over archival standards, this quote made me question whether sacrificing access and preservation to retain imperfect metadata could truly be considered “best practice.”

Working closely with the UT Libraries Digital Stewardship department, I created a workflow whereby I extracted files using FTK Imager and generated SIPS and a bulk_extractor report using the Canadian Center for Architecture’s Folder Processing Tool in the BitCurator environment. The only impact on the metadata was a change in the “Date Modified” field, which now showed the date the files were extracted and could be amended to reflect the most recent date in the file tree. While this is an unconventional approach to digital preservation, the resulting AIPs and DIPs are better representations of the original disks and will allow us to provide more comprehensive data for future researchers.

Blog3
The amended workflow.

Blog4
Our new results.

After my presentation, I had the opportunity to talk to a digital archivist about this workflow. We discussed how best practice is ultimately not about perfection, but about preserving and providing access to our materials and documenting the process. While somewhat contrasted from the theoretical approaches I’d both learned in class and adapted from online resources, this approach seemed more natural to me. It affirmed both my processing decisions and the opinions I’ve developed about what best practice in the archival community is and should be. This process has shown me how vital the user is to the archive, especially when developing workflows for digital materials.

As I mentioned in my previous post, I love having opportunities to exercise creativity and problem solving in this position. It’s even better when those opportunities lead to breakthroughs that help me grow professionally and enhance the data we provide patrons. I’m excited to see new developments in our born-digital workflow as we get closer to making this collection available to patrons. Check back soon for a (final?) update from me – I hope you’ve enjoyed learning more about the behind-the-scenes work of born-digital archiving.

archive, digital collection

Consider the Floppy: Exploring Access Issues with the Volz & Associates, Inc. Collection

A black 3.5 inch floppy disk with a label reading "RJA: Geomsn.zip"
A floppy disk from the Volz & Associates, Inc. collection.

Hello! This is Abbie Norris, back with an update on the born-digital Volz & Associates, Inc. collection. For those who haven’t read my previous blog post, I am the digital archives Graduate Research Assistant at the Alexander Architectural Archives. I’m currently working on the Volz & Associates, Inc. collection, which documents the work of a historic preservation firm based in Austin, Texas.

When I published my previous post, the collection was in the midst of being processed. I’m happy to report that processing for this collection is complete – all 813 floppy disks, CDs, zip disks, and flash drives of it. Processing is one of the first major stages of getting a collection from the donor to the public. It’s when the bulk of archival preservation happens. In this case, as in many born-digital collections, “processing” involved imaging (essentially, copying) the disks, capturing metadata like disk size and file types, and recording everything for documentation in the finding aid. We’re now able to determine the size of the collection, the types of files, and what we need to provide access to them.

One of the things I love about born-digital archiving is the problems that arise that require creative solutions. This is especially true for an archive’s pilot born-digital collection, as is the case for Volz. Items like CDs and floppy disks degrade at a faster rate than paper materials, meaning that sometimes you try to open a disk that physically appears fine, and it won’t show any of your files. One major question people have is, “If all of the information is stored on a CD, why do you have to copy the contents in a disk image? Why can’t you just continue opening the CD to access the contents?” Luckily, the answer is simple.

Imagine you have a 13th century codex and a 1990s floppy disk. Which one is easier for you to read? With the codex, all you have to do is open it. It will be fragile and you might not know the language in which it’s written, but much of the information held in the book will be visible to you. Now, consider the floppy disk. When was the last time you used one? Does your computer still have a floppy disk drive? More than likely, the answer is, “No.” Even if it does, think about the files on that disk. Can you still open a WordPerfect document from 1992?

Given that the Volz collection dates between 1980 and 2009, the types of files present on the disk vary widely. Some, like .txt and .tif files, are still widely accessible and are projected to remain active filetypes in the future. Others, like .jpeg, are still accessible but are not recommended for preservation because of their lower quality and the potential for their use to cease. Finally, there are the files that you try to open with modern software…and nothing works. These files can be either old versions of proprietary software and discontinued software.

This is where creative solutions come in. There are a variety of tools, many borrowed from the criminal forensics world, that allows us to look at files from twenty years ago. Because the digital archive field is still developing and many of the projects use open access tools, the software an archive uses to read and provide access to old files can resemble a patchwork quilt. Now that I know exactly what types of files are in the collection, I love exploring access solutions and finding answers to questions that have persisted since I began working here.

In many ways, finishing processing feels like finishing the first segment of a relay race. I feel accomplished for finishing a major task, but there is still a long way to go. Now that processing is complete, we have to finish writing the finding aid and establish methods for researchers to access the collection. It’s going to be an exciting few months, so check back here to learn about what providing access to a born-digital collection looks like at the Alexander Architectural Archives.

digital collection, library, special collections

Southern Architect and Building News Project Update

Hello, People of the Blogosphere!  We’ve been very busy working on a number of projects this summer, one of which has been the Southern Architect and Building News digitization project.

Part of what’s so neat about this project is that we get to collaborate with other wonderful folks around the UT Libraries.  The downside of this is that it means we have to coordinate and make decisions together about the workflow, standards, and overall goals.  This does make it take a bit longer than if all the work were happening in-house here at the Library.

IMG_5243
Picking up digitized volumes of SABN!

So far, we have sent several batches of about 10 volumes over to our friends in Digitization Services to be scanned and ingested into the Data Asset Management System (DAMS).  They have input some metadata for us, but part of what we have to decide here at the Library is what our metadata schema will look like.  There is a lot of information we could include in our metadata, so we have to make some choices about what will be most meaningful to our end users and how much information it is realistic to enter for each item.

IMG_5257
Packed to go to PCL!

The last batch of Southern Architect has been sent to PCL to be digitized, which is pretty exciting!  Once this is finished we will pick them up and they will join the rest of Southern Architect to await metadata entry!  Currently all the already digitized copies of Southern Architect are back at the Library, so once the paged content issues with the DAMS are fixed, we should be ready to go all-in on metadata!   

We’re slowly working our way through some of the issues with the DAMS.  This project is for the long haul, so it is going to take some time to get Southern Architect and Building News digitized and available online.  But, we also known what a special publication it is and how fascinating it is and its value to architectural history!  We can’t wait to keep working to make sure every one can see for themselves what makes Southern Architect so special and important!

archive, digital collection

The Volz & Associates, Inc. Collection: Born-Digital Initiatives at the AAA

One of the many structures VOH Architects worked on: the Littlefield House on the University of Texas at Austin campus.
One of the many structures VOH Architects worked on: the Littlefield House on the University of Texas at Austin campus.

Hello! My name is Abbie Norris, and I am the current digital archives Graduate Research Assistant at the Alexander Architectural Archives. My primary job is processing the born-digital content received in the Volz & Associates, Inc. collection. This collection contains the records of the Volz & Associates, Inc. architecture firm, which is focused primarily on preserving and restoring historic buildings and interiors. The collection showcases notable buildings from Texas and United States history and is an excellent resource to discover how much is needed to keep historic buildings authentic and alive.

A gray CD reading "Images for Volz: Elisabet Ney Museum, April 2007"
A sample CD from the collection. Born-digital archiving requires preservation two ways: retention of the original media and capture of the data for long-term storage.

The Volz Collection is significant to the Alexander for several reasons, but most importantly, it is the archive’s first large-scale born-digital accession. In addition to analog records and building materials, the collection includes roughly 450 floppy disks, 250 CDs, 90 zip disks, and one lone flash drive. These materials document the life of the firm from the early 1980s to the mid 2010s. So far, we have imaged over 100 filetypes representing everything from office files to construction reports to historic photographs. It’s a diverse array, and as the project moves forward, we’re faced with many questions about how best to provide access to researchers.

As diverse as the filetypes are the kinds of buildings included in the collection – though many are tied by one important identity. Volz  worked on buildings of many functions, styles, and preservation needs. While these buildings span the United States, the majority of them are located in Texas. Included are the Governor’s Mansion, the Alamo, the Lyndon B. Johnson Ranch, and the Alexander’s own Battle Hall. I love working with this visual representation of Texas history. Whether it’s by noticing design similarities between county courthouses or the way historic landmarks are used and maintained, the collection is an in-depth look into how architecture shapes our state and its identity.

Scaffolding covers a green dome atop a white tower.
Restoration underway on the Colorado County Courthouse dome. Photo credit: Volz O’Connell Hutson Architects, (http://voharchitects.com/projects/colorado-county-courthouse/).

In my four months of working with this collection, I’ve learned an incredible amount about both the intricacies of born-digital archiving and the breadth of work architects do. Through the frustration of software bugs and the triumph of imaging previously unreadable disks, this is a fascinating collection that provides many learning opportunities.

The next steps of the project are to finalize the creation of a finding aid for these born-digital materials and to determine methods of access once the collection is published. Check back here soon for collection updates and an in-depth look at the world of born-digital archiving at the Alexander Architectural Archives!

digital collection, special collections

Special Projects: Southern Architect and Building News

Southern ArchitectHere at the Architecture and Planning Library, we have long been obsessed with fascinated by our impressive collection of Southern Architect and Building News journals.  Printed from approximately the late 1880s to the early 1930s, Southern Architect was published to highlight the architecture and architectural news of the Southern United States, with similar content to American Architect.  Our Special Collections contains the largest known collection of Southern Architect ranging in date from 1892 to 1931.

The journal itself contains articles and advertisements relating to all things architecture in the South.  The articles include pieces on homes and buildings around the South, complete with drawings and/or photographs.  Taken together, Southern Architect documents trends in architecture, society, technology, and advertising that make it an important part of architectural history.

It has been a longtime goal of our Librarian, Dr. Katie Pierce Meyer, to digitize Southern Architect.  We’re so pleased that we have finally received funding to do so!  We are currently working with the Head of Digitization Services for UT Libraries, Anna Lamphear, and her team to digitize and make available on the UT work-in-progress Data Asset Management System.  Digitization Services is producing high quality, searchable scans of the journal and then uploading them to the DAMS.

The digitization is going much faster than we anticipated, so we are now in the process of adding further metadata on the digitized issues.  Part of that process is developing a workflow and a standard for our metadata.  Our Digital Initiatives GRA, Zach, Dr. Katie, and I are currently working on identifying the most important metadata attributes and creating a standard for us to follow as we all work on entering metadata.

This project is still in its early stages, so that’s all I can say for now about it.  But, we’re really excited about finally getting Southern Architect digitized and making it accessible, and we hope you’re excited too!  Stay tuned for more news on this awesome project!

 

digital collection, exhibition, library

On Omeka: The Good, The Bad, The Ugly

Hello y’all!  Katie (2.2) here.  I admit that Battle Hall Highlights has not been quite as active on the blog front from the Library compared to other years – apologies for that, but it was for a very good reason!  The University of Texas at Austin Libraries have been working on implementing Omeka as a platform for publishing digital collections and exhibits from UT institutions.  Our own Katie Pierce Meyer (the Librarian here at the Architecture and Planning Library) was instrumental in bringing Omeka to UT, and APL is proud to now have three exhibits available via Omeka.  As GRA for the Library, I spent the majority of my semester migrating content from our website, which was housed and designed via Drupal 6, to UT’s Omeka.  The three exhibits on Omeka now are “Our Landmark Library: Battle Hall at 100,” “Their Maya Story: George and Gerrie Andrews,” and “Eugene George: Architect, Scholar, Educator, Photographer.”  Each exhibit posed unique challenges in migrating the content, and have provided invaluable experience in creating exhibits via Omeka.

So, what is Omeka?  It is a free, open-source platform for publishing digital collections and exhibits.  Developed and updated by the Roy Rosenzweig Center for History and New Media, Omeka has several iterations (Omeka S, Omeka.net, and Omeka Classic), but originally started out as a platform designed for small institutions with limited resources.

The Administrator view of Omeka.org.
The Administrator view of Omeka.org.

Omeka is meant to be easy to use and simple to upkeep, especially for those with limited technological know-how (e.g., me).  So, with some persistent encouragement from Katie Pierce Meyer, UT eventually decided to install Omeka as a platform for the UT Libraries to upload digital collections and exhibits.  So only our three exhibits, and one about South Asian Popular and Pulp Fiction Books, are available via Omeka.  There are various reasons Omeka has not gained traction: it is necessary to first have a digital collection about which to build an exhibit which takes a great deal of time, it has been time consuming to try to coordinate with the UT Data Asset Management System (DAM) to develop a metadata standard, and, simply, Omeka is not perfect.

Omeka's Exhibit Builder from the Administrator view.
Omeka’s Exhibit Builder from the Administrator view.

Where other exhibit builders (such as Scalar) are more image-focused, Omeka is quite metadata-heavy.  This is evident in the limited theme options (basically, themes are the design layout, or style, of the exhibit) currently available on UT’s Omeka, which tend to be text-based with smaller images.  Hopefully, as more institutions digitize items and see the existing exhibits on Omeka, the site will grow in popularity as a platform for UT’s Libraries to share more of their unique collections.

The former UT Omeka homepage on Omeka.org.
The former UT Omeka homepage on Omeka.org.

In late December of 2017 to early January of 2018, all Omeka content was migrated from Omeka.org to Omeka.net.  What the Libraries and the IT team decided to do was host content via Omeka servers rather than host it themselves.  This created the opportunity for each institution to have their own independent Omeka site that they control.  When we worked in Omeka.org (the homepage of which is pictured below), everything was in the same bucket, so to speak.  At that point, two of our exhibits were completed, and the Architecture Library and the South Asian Pop Culture Collection were the primary users of Omeka.  Now, the South Asian Popular and Pulp Fiction Books has its own site, and we have our own website now (which you can see here: http://utlibrariesarchitecture.omeka.net/) that we can customize to fit our needs as an institution.  The site is still a work in progress, as we are working on adding more content, but we now have three exhibits (all discussed thoroughly below) publicly available for everyone to enjoy and explore!

Battle Hall at 100: Our Landmark Library

Pros: Learned how to import items one at a time

Cons: Took three to four months to complete, took a while to standardize metadata (had to go back and change early entries), and there was a lot of trial and error (in both uploading items and in building the exhibit and making it look well-done)

Theme: Thanks, Roy (the fullsize image display is rather small, as are the thumbnails, menus are confusing because Omeka menu is on left and exhibit menu is on right of the text, and the metadata page is text-heavy with a thumbnail image at the very bottom, nice galleries of images, though)

Our Landmark Library Exhibit as built in Omeka.
Our Landmark Library Exhibit built in Omeka in 2017, migrated from Drupal 6.

The first of the three exhibits I migrated was Our Landmark Library: Battle Hall at 100.  This one was migrated “by hand,” if you will, meaning I moved each item individually, entering in the metadata and adding the image files myself.  The main takeaway?  This method of importing items is slow.  Containing approximately 140 items, there was a lot of trial and error (mainly human error) in migrating Our Landmark Library.  Namely, it took time to find the best standard  for the metadata, because I wanted to be sure to include all the information included in the old exhibit on Drupal.  Eventually, I found a standard, wrote it down, and followed it for the remainder of the items, before going back and changing the ones I had already uploaded.  So uploading the items was the first phase of this project.  The second phase was reconstructing the exhibit in Omeka.  I

The Battle Hall at 100 Exhibit as built in Drupal 6 on our legacy site.
The Battle Hall at 100 Exhibit as built in Drupal 6 on our legacy site.

wanted to maintain the original order of the exhibit, which was simple enough.  The hardest part of building the exhibit was putting together the galleries, which were pretty big in this exhibit.  Formatting the exhibit to have a nice flow to it (no big gaps or spaces between what are called “blocks” on Omeka, a means of separating and formatting parts of a page), keeping the images in the correct order, and writing captions for each image was not challenging so much as time-consuming.  Even though this exhibit took a long time to complete because of the method of uploading each image myself, I am glad I had the experience.  I became intimately familiar with how Omeka works, and the rhythm you can get into when uploading every item and formatting an exhibit.  I definitely have a great appreciation for how long building a digital collection can take.  I did not have to digitize any of the items or create original content for the exhibit, but it still took a long time to move everything.  If it took three, almost four, months for me to migrate an exhibit of only 140 items, I can only imagine how long it would take for large collections.

The new exhibit on Omeka!
An example of a gallery built in Omeka!

 

 

 

 

 

Their Maya Story: George and Gerrie Andrews

Pros: Quick with CSV import, less trial and error, building the exhibit only took two days; Easy metadata standardization when done in Excel ahead of time

Cons: Figuring out the CSV import (must be done very quickly or it has to be started all over again); Mapping elements (sometimes difficult to match column headings with element names for DublinCore metadata)

Theme: Neatscape (nice display of the images,  and large, easy-to-read text, the menu at the bottom is the only downside, only available in Omeka.org); Big Picture (current display in Omeka.net, large image display, allows easy navigation between items in galleries and between pages in the exhibit)

Their Maya Story Exhibit on Omeka.
Their Maya Story Exhibit on Omeka.

After (finally) finishing the Battle Hall at 100 exhibit, I moved on to the George and Gerrie Andrews exhibit, Their Maya Story.  This exhibit Katie Pierce Meyer and I imported into Omeka via a CSV (Comma Separated Value) file that had been generously created for us.  When importing a CSV file, you choose “elements” of metadata that align with the columns in the file (e.g. the “Title” column matches to the “Title” element in Dublin Core, or the “Geographical Location” column might match to the “Spatial Location” element); it is important to note that not all column headings have a corresponding element to match to, so it helps think about this somewhat beforehand.  After a mishap during which Katie and I took too long matching columns to elements, we successfully uploaded all 116 items to Omeka in less than two minutes, complete with metadata and attached image files.  I then began to build the exhibit in Omeka.  I again wanted to be faithful to the original exhibit.  I wanted to try a different theme this time, so I chose the theme called “Neatscape,” which does not permit any customization or changes to the display the way some other themes do.   The only

The original Their Maya Story Exhibit in Drupal 6.
The original Their Maya Story Exhibit in Drupal 6.

problem I have with Neatscape as a theme is that the menu for the exhibit is at the very bottom of the page, so you have to scroll through the whole page in order to reach the menu.  One of the perks of Omeka is that it allows users to change themes without changing the content or layout of the

Metadata display in Drupal 6.
Metadata display in Drupal 6.

existing content of the exhibit.  We decided to go with Neatscape so that we have an example of what it looks like as an exhibit, though we may change it in the future if we find a theme that works well for our content.  A major asset of Neatscape is that the metadata pages for individual items does include a small thumbnail immediately to the right at the top of the page, meaning that (unlike in the Thanks, Roy theme in the Battle Hall at 100 exhibit) users do not have to scroll all the way down the page to see only a small thumbnail image that, when clicked on, then leads to a larger image.  Building the exhibit was much the same as before, only this took merely two days, between using the CSV import and having had the experience of building an exhibit.  It is amazing how much time the CSV import cut out, so I definitely learned that it is necessary to have all the metadata standardized in a CSV file in order to make the process much easier and more enjoyable.

The image and metadata display in the Big Picture Theme used for the Andrews Exhibit.
The image and metadata display in the Big Picture Theme used for the Andrews Exhibit.

After the migration from Omeka.org to Omeka.net, the Neatscape theme was no longer an option, so I chose to use the Big Picture theme, which emphasizes images.  The thumbnails for the galleries are large, and the images displayed when a user clicks on an item is larger than any of the other themes.  We still have not figured out how to add zoom functionality to our exhibits.  That is a task that the Library’s new Digital Initiatives GRA, Zach, is going to be working on in the coming months.  So the Andrews exhibit required a little bit of tweaking after the migration from Omeka.org to Omeka.net, but we wound up with a better outcome because of it.  The Big Picture theme is currently my favorite among the three I have tried so far because of the large image display, which is preferable for some of the items we might include in future digitization efforts.

Eugene George

Pros: More theme options because of the move to Omeka.net, quick because of CSV import, nice to have the exhibit ready to go whenever the CSV file was done so that all that remained was to add images and galleries; took only one day to complete the exhibit

Cons: Harder to build an exhibit without the images ready to go in, quite a bit of wait time due to the migration of all Omeka content from UT’s servers to Omeka.net

Theme: The Daily (there is a nice menu that scrolls down alongside the content to navigate between pages, large text, large image display)

Eugene George Exhibit built in Omeka in 2018.
Eugene George Exhibit built in Omeka in 2018.

This exhibit about Eugene George was done in a slightly different order from the others: instead of importing the items and then building the exhibit, I built the exhibit first since our CSV file was not ready to import.  So I copied over all the text from the exhibit, and created blocks of text and galleries to mimic the original order of the exhibit.  We had to wait on the CSV import until after the migration from Omeka.org to Omeka.net, which took longer than expected, but in early February, we received the go-ahead to import items into the new site.  Katie, Zach, and I sat together and worked on standardizing the metadata in the CSV file before importing it into Omeka, which mostly consists of renaming columns and copying

The original Eugene George Exhibit built in Drupal 6.
The original Eugene George Exhibit built in Drupal 6.

over the desired metadata content.  After several failed attempts at importing the CSV file, we realized that because our website is now a legacy site (http://legacy.lib.utexas.edu/apl) that the links included to connect the JPEG images with their respective metadata would not work.  So, we changed those links to include “legacy” before the link, and

Metadata display in the Eugene George Exhibit, which used the theme The Daily.
Metadata display in the Eugene George Exhibit, which used the theme The Daily.

the CSV import worked perfectly!  With those items now in Omeka.net, I was able to add them into the pre-built exhibit.  I then went back and made sure that the galleries and text looked nice, made the necessary edits, and the exhibit was ready to go!  For this one, I chose The Daily as the theme.  Overall, the theme looks nice, with large text, large image display within the exhibit.  The menu that scrolls along beside the content allows for quick and easy navigation among the pages of the exhibit.  The pages where item metadata is displayed are less text-heavy than in the Thanks, Roy theme used for the Battle Hall exhibit (and feature the image at the top of the page instead of a thumbnail at the bottom), but the image displayed is not as large as that in the Andrews exhibit.

By the time I built the Eugene George exhibit, I had a far greater understanding of how Omeka works and how long it might take.  With the CSV import and the exhibit pre-built, this exhibit took only one day to complete.  This means that just one day of intensive work is needed to build an exhibit when the CSV import is used to add the items, versus nearly four months doing an item-by-item import for the Battle Hall exhibit.  It is easier to build the exhibit as you go, as I did with the Andrews exhibit, rather than having it already built.  I found it far easier to make changes as I went along building the Andrews exhibit than to have to go back afterwards as I did with Eugene George.

Conclusion

In spite of my limited technological expertise and lack of knowledge about Omeka, I found the platform very easy to use.  Exhibits and importing items (whether individually or via CSV file) takes time and patience.  For collections less than 50 items, importing items one-by-one is fine, but for larger collections, the CSV import saves a great deal of time.  It is intensive to create a CSV file with all the pertinent metadata, but it is preferable to having to individually import 200 items and type in metadata for each.  With a CSV file, the standardization is done before the items ever make it into Omeka.

In terms of Omeka.org versus Omeka.net and the Administrator side of things, not much has changed.  However, it is nice to have our own site to manage and customize.  Since the change to Omeka.net, we have been able to play with the website and are working toward making it exactly what we want.  It is still a work in progress, but it looks pretty sharp!  The ability to manage our own content, look, and navigation is something we never could have achieved through Omeka.org.  Once the Libraries adds links to the UT Libraries website to reach all of the Omeka sites (the South Asian Popular and Pulp Fiction Books site, the Benson’s site), it will allow users to easily access the incredible digital collections that are forming on Omeka.  Additionally, Omeka.net provides more options in terms of themes than Omeka.org.  Though some do not translate between the two, the increased options have allowed us to try different themes for each exhibit to see what is the best display and format for Architectural exhibits.  Hopefully Zach or someone with more technical coding expertise can find a way to add zoom functionality to our images, something that will increase the usability of our items.

The new Architecture Exhibits site on Omeka!
The new Architecture Exhibits site on Omeka!

Overall, working with Omeka was an enriching experience for me.  Coming from a library/archives background, getting to do a project like this was incredibly rewarding.  With a free platform like Omeka, anything is possible.  Even a relative luddite like myself can use it to build digital collections.  For the most part, I am handing off the reigns of digital projects to Zach, but I’m proud of the work I have done on these three collections.  They are not perfect, but I learned so much about building digital collections from the experience of migrating this content.  And currently the Architecture and Planning Library is paving the way for other UT institutions in using Omeka as an exhibit platform, about which we are extremely excited.  Please explore our Omeka site and enjoy the exhibits as much as I enjoyed building  them!  Signing off for now, your Friendly Neighborhood Omeka Semi-Guru.

digital collection, exhibition

We’re going (Way)Back in Time!

Hiya!  I’m Zacharia (Zach if you’re feeling friendly) and I’m the new Digital Initiatives GRA here and APL.  I’m going to be in charge of making sure you wonderful people will still be able to find our online exhibits once the sites have gone offline.

So I thought I’d give you a little how-to guide about how you can access our old exhibits while they’re in migration.  So without further ado let me introduce you to our new best friend: The Wayback Machine!

The Wayback Machine is a delightful little site hosted by the Internet Archive that is dedicated to archiving the entire internet forever (give or take a few hundred thousand versions).  Today I am going to walk you through how to access the old APL legacy site and all the associated content.

Step 1: Accessing the Site

The Wayback Machine works by automatically scraping webpages, creating exact replicas of the site on a given day which can be accessed from the main interface.  These are stored versions of a site at the time of scraping, and will not reflect any changes made to the site following the date of scraping.  Think of it as visiting a historical site.

The url for our old site is https://legacy.lib.utexas.edu/apl.  So we’re going to copy that and paste into The Wayback Machine, which should look something like this: (Please note that the images are best viewed in full screen so don’t fret if you can’t quite make out specific text, just click on the image.)

The Wayback Machine

Now I have already scraped the legacy site as of January 31 and February 6.  The next screen you see should contain the url you just entered and a series of calendars with the dates January 31 and February 6 highlighted.  These represent specific “images” or captures of the site as it looked on those days.  There shouldn’t be any difference between the two so pick your favorite date!

Step 2: Navigation

Now those of you familiar with the old site should recognize a lot of what you see, with the exception of the Wayback Machine interface, which will look like this:Capture

This is the Wayback Machine navigation bar.  From here you can navigate between different captures of the site at will and see a timeline of their development.  Now the navigation of the old site works the same as it once did, however when you select a link you may see this image:

Capture2The Wayback Machine archives each webpage individually, and must redirect and access different versions when you go to a new web page.  What that means is that when you click on a link you are redirected to the Wayback Machine’s most recent or closest temporal capture of the webpage that is being linked.

Step 3: Accessing the old Online Exhibits

My work at APL primarily revolves around making sure you fine folk can still access all of our old online content, so let’s see if we can’t access one of our old exhibits:  The Architectural Legacy of Herbert Greene.  As we are still on the main web page we will need to move over to the Works and Projects page.  Before moving on I would like to point out that the entire legacy UT Libraries site is not fully accessible on the Wayback Machine, just these exhibits which I have manually captured.  So keep that in mind as we go forward.

In any case, the Works and Project page should look like this:

Capture3

If your screen does not look like this please select the “Works + Projects” button. All of our exhibits are housed under Exhibits and Curated Resources, but all the other pages found here are fully accessible so feel free to look around.

Once you’re  looking at the Exhibits page, select the hyperlinked text Online Exhibits and Curated Resources to find the list of old online exhibits.  Fun fact, you can also access this delightful blog from here and see all the archived posts!

The Architectural Legacy of Herbert Miller Greene is the first item on the list.  If you click on it should take you to:

Capture4There is a chance that the images or other script may not load.  This is a result of a problem with the given date of capture.  If you are having this issue, access a different capture from the navigation bar (such as January 31) to see the full site.

Perhaps the most important function of The Wayback Machine  is the ability to emulate and preserve JavaScript and Flash programs, so applications and other non-still image media are accessible in their original forms.  This is especially important for an exhibit like The Architectural Legacy of Herbert Miller Greene, which is based heavily around an applet called Zoomify, which allows users to zoom in and examine blueprints and photos at a high resolution.  Select Firm Brochures from the navigation bar on the site to take a look.

Capture11Ta-da! It’s like the site was never taken down!  You can freely explore the exhibit to your heart’s content.

All of old exhibits will be available in new formats on the new libraries site, but in the meanwhile (or if you’re feeling nostalgic) you can use The Wayback Machine to find all of your favorite content just the way you liked it.

I’m looking forward to sharing our progress on the new site as time goes on, and I’ll be back real soon with more awesome stuff to show you!