Open access books through open data sources: assessing prevalence, providers, and preservation

E-books and their preservation

The preservation challenges related to e-books have existed roughly as long as the medium has had any significant volumes of content published. It is around two decades since Frank Romano authored an article titled “E-books and the Challenge of Preservation” that identifies three related challenges to the preservation of e-books: “1. The location of the stored information, 2. The organization storing the information and its long-term viability and commitment to preservation, 3. Technical issues involving coded and recorded format, interfacing, and rights management.” (Romano, 2003). When it comes to the context of OA books, we can today argue that the first two are still largely unresolved and a motivation for initiating this study, while the third could be argued to be partially resolved through the mature standardization of the most commonly used formats for representing static print digitally (PDF, EPUB) and the availability of reuse rights and the lack of digital rights management considerations for OA content. A central theme in Romano's paper is the uncertainty of the different responsibilities between publishers and libraries when it comes to content in the purely digital domain. Therefore, preservation organizations such as Portico and CLOCKSS, owned and governed by publishers and libraries together, have been set up and contribute to bridging this gap for scholarly journals and books. An earlier version of Romano's text was included as part of a report commissioned by the Library of Congress and the Council on Library and Information Resources in the USA titled “Building a national strategy for digital preservation: Issues in digital media archiving” that included similar chapters that related to other mediums where digital preservation needs were emerging (e.g. periodicals, websites, sound and video) (CLIR and Library of Congress, 2002). A report titled “Preserving eBooks” and published by the Digital Preservation Coalition efficiently summarizes the key challenges when it comes to the preservation of non-OA e-books (Kirchhof and Morrissey, 2014). A central finding of the report was that e-books outside of the OA context are substantially different in comparison to traditional books in that readers and libraries usually only gain access to a limited license to access the contents of the book rather than owning the book itself. This foundational difference combined with the challenges of still evolving technical format standardization, competing systems for digital-rights management and emerging business models introduced a substantially new environment for preservation service providers, Hurley (2019) provides a comprehensive history, description and comparison between the largest preservation service providers in the e-book space: Portico, CLOCKSS and the Global LOCKSS Network.

In order for something digital to be uniquely identified, there needs to be a widely adopted standardized system for how to assign identifiers to objects. In the paper-based world, ISBNs functioned reasonably well for this purpose since the standard was created in the late 1960s, but as Scott and Orlikowski's (2021) study on the digital transformation of the book industry reveals, the ISBN system has started to show some serious limitations with not all books getting indexed in the system, and differing practices for its use among publishers for e-books. The study points out that the ISBN standard commonly requires that a separate ISBN be issued for every edition and format for a book, which means that PDF files, EPUB files and other formats all need their own ISBNs to be compliant. Digital files under different licensing or digital rights management restrictions also need their own individual ISBNs. The exception to all this necessary ISBN assignment is if an edition of a book, or part of a book, is available through one single source only and does not need to be processed into the broader book supply chain. The authors point out the fact that Amazon does not require books to have been assigned an ISBN to be sold on the Kindle book store to be a notable shift in the practice of ISBNs use in the e-book context. The next section will delve more particularly into how digital object identifiers (DOIs) work in this space, as they have grown into use in parallel with the emergence of digital publishing, further challenging the perception of ISBNs when it comes to unique identifiers for books.

The context of OA book publishing

During the last decade, there have been several national and European projects that have supported the building of infrastructures for OA books. For someone new to this space, the number of projects and acronyms can seem confusing. Stern (2021) provides a helpful narrative for the origins of OPERAS (open scholarly communication in the European research area for social sciences and humanities) and how many of the related initiatives like OAPEN (Open Access Publishing in European Networks), DOAB (Directory of Open Access Books) and OpenEdition are connected in this context. A recent project with a significant focus on preservation is the COPIM project (Community-led Open Publication Infrastructures for Monographs) where a work package is dedicated to archiving and digital preservation. At the time of writing the project has produced a scoping report which covers a brief overview of existing technical methods for digital preservation together with findings from interviews and workshops, where a main initial conclusion is that diverse solutions are needed and that it is unlikely that any single model will provide a solution for everyone when it comes to preservation (Barnes et al., 2022).

A key study in the context of scholarly OA books is Neylon et al. (2018), which presents an overview of their visibility and integration in the digital landscape of scholarly works. The authors conducted both a web survey and analysis of the OA book content and associated metadata for seven publishers (including indexing inclusion in various web services), all partners of the European OPERAS network. While preservation is not directly dealt with in the report, there are many identified aspects upstream and in the overall technical landscape that influence whether and how preservation can later take place. The study highlighted several key challenges that OA books have in comparison to OA journal articles. Books are often distributed through multiple online platforms and the publisher's website might or might not be one, and there is no persistent identifier for the overarching work which strains the use of persistent identifiers such as DOIs and ISBNs for the various manifestations of that work. There is currently no comprehensive collection of usage data as there is with journals via COUNTER. The open systems used for cataloguing, indexing and discovering OA books are much younger than they are for OA journals, which shows in their lack of consistency and reliability. While journals have strongly shifted to online only, there is still a larger demand and practice for books to be printed, suggesting that the processes for digital and print will remain parallel at least for some time. The seven studied publishers were small organizations with limited resources and capacity, calling out for coordination, shared services, infrastructures and standards in their survey responses. The publishers delivered their metadata in various formats and levels of quality, from various file types to APIs, demonstrating the diversity in managing and making data available of published works. A particular challenge in the metadata was the inconsistent use of persistent identifiers, where multiple ISBNs could be reported for different manifestations (e.g. editions and formats) of a book, in addition to a potential DOI that could also be inconsistently reported in cases where individual chapters also had their own DOIs. Only around 10% of all OA books from the publishers were associated with a DOI. The authors argued the variable quality of book metadata created challenges for reliably studying their presence and indexation in various web services as the study compared publisher-provided records and that of various external services and indices (WorldCat, BASE, Google Books, DOAB and OpenAIRE). Most of these resources were at least 80% comprehensive; however, BASE had a low inclusion rate of around 40%. The study found no major difference in the degree to which content in different languages was included in the various services.

Echoing many of the same observations with the challenges of OA book metadata as Neylon et al. (2018), the 2019 report “The State of Open Monographs“ by Grimme et al. (2019) provides a helpful review of the historical and technical origins for why the data landscape for OA books is so fragmented. The report highlights the basic difficulty of even exactly knowing how many OA books are published, using a multi-method approach to come up with an approximation of around 80,000 for the year 2013.

Based on the earlier findings of a lack of standard practices for metadata for OA books, a substantial part of the COPIM project was dedicated to developing a minimum metadata requirement for OA books based on the needs of key stakeholders. This work was part of a more encompassing Open Dissemination System, which also included guidelines for the standardized use of persistent identifiers (Stone et al., 2020). In the COPIM project, Barnes et al. (2022) found, through their interviews with stakeholders, that there are some publishers that upload published content to local repositories. The role that repositories could potentially fill in the context of preservation is still largely unexplored and undefined. One challenge is the heterogenous landscape of repositories, operating with varying organizational backing and technical expertise for ensuring long-term access to content. The way through which repositories perceive themselves as responsible for the long-term preservation of the outputs of an institution varies and is not in any way an underlying requirement for running a repository (Francke et al., 2017). A recent literature review of 21 studies dealing with long-term preservation in institutional repositories showed concerning findings where the review “… has not found clear evidence about how institutional repositories are implementing digital preservation” suggesting that more clarity into the roles and responsibilities of repositories is needed (Barrueco and Termens, 2022, p. 161). Another recent study found that about a quarter of repositories registered in widely used repository indexing services gave an erroneous response and were inaccessible when an attempt was made to visit the registered URL to the repository (Mannocci et al., 2022), a finding which is very concerning. There is an active discussion ongoing about what kind of information would be required to evaluate which repositories can be trusted for long-term preservation (Lin et al., 2020), but so far there is no wider adoption of any practical standards outside of CoreTrust Seal which currently has certified only around 190 repositories and data archives (coretrustseal.org, n.d.).

From the reports and studies done on the landscape of organizations involved in OA book publishing, it is known that many actors are small. In such contexts, it is important that IT systems automate and guide as much of routine workflow steps as possible. One practical example of this is given as part of a case study of a university press where Taylor (2019) notes the following related to the selection of a particular publication management system: “Features which particularly appealed to us included the automatic registration of digital DOIs, the ability to send content to a preservation service at the click of a button …” (p. 6). Comprehensive and standardized metadata for facilitating aggregation into external services is also important and should be supported and automated by the publishing platform. In a study of how users access books from the OAPEN Library website, 73% directly accessed the full-text file without opening the actual website, suggesting access by other means such as aggregators, search engines and libraries (Snijder, 2019).

Momentum is building for libraries to get more seriously involved in the structural funding of scholarly OA books (see e.g. doabooks.org 2022) which brings the content closer to libraries and their potential preservation processes. Recently, UKRI (UK Research and Innovation) commissioned a gap analysis of the infrastructures for OA scholarly books, where preservation was included in the comprehensive scope of analysis (Ferwerda et al., 2021). Regarding preservation, the authors identified a gap as “Technical challenge of preservation and ambiguity concerning who is responsible for the preservation of OA books.” (p. 7) with a recommended action to collaborate with UK legal deposit libraries and international partners.

The strategies and processes for how national libraries are approaching e-book preservation have appeared in the academic literature, with some prominent examples being China (Wei et al., 2014), France (Derrot and Clément, 2014; Derrot et al., 2014) and UK (Gooding et al., 2021), but none of these mention or deal with OA materials specifically. One way that many national libraries have for centuries tried to preserve works published in the country is through legal deposit, where publishers are required by law to submit copies (be it digital or printed) to the national library upon publication. Muir (2001) is one of the early seminal works studying how libraries approach the deposit of digital publications. More recently, Roudik et al. (2018) provided a review and comparison of how digital legal deposit is implemented in 15 countries. The International Internet Preservation Consortium (IIPC) also maintains a list of countries with Legal Deposit laws covering web archiving IIPC (n.d). Unfortunately, most access to these archives is restricted to either on-premise access or based on the evaluation of individual applications. This approach to long-term preservation, therefore, does not facilitate sustained and uninterrupted access to OA or other published works. An exception to this is an initiative focusing on OA books specifically that is run by the Library of Congress, where they are ingesting titles for which they already have print holdings directly from DOAB, both for preservation purposes and for making them available through their own website (Cassidy-Amstutz et al., 2022). Implemented at a larger scale, involving more libraries and titles, these kinds of actions could help contribute a robust layer of resilience to the preservation of open materials.

From this overview of previous research, it can be concluded that broad and systematic studies on the three main focus areas of this study (prevalence, providers and preservation) for OA books have not really been attempted before. Overall, the literature related to these topics is dominantly populated with project reports and conference proceedings, suggesting that there is potential for making substantial contributions to the academic literature through empirical studies.

OA book data

Not all books on the web are of key interest to this study, where focus is on non-fiction academic books. Most bibliometric databases provide filtering to either “Book” and/or “Monograph” with very few offering further ways to reliably narrow the scope down from there. There is no widely used tag for “peer-reviewed” or similar that would make it possible to filter the large quantity of entries down, leaving it up to the inclusion criteria/data harvesting methods of each service provider to what is included and what is not. Further, as categories are so wide there can be theses, reports and individual book chapters sprinkled in among the search results which are hard to identify and separate in any automated way. This is not only a factor that concerns metadata but also overall transparency and knowledge about what kind of editorial processes are behind published works. For the sake of replicability, it is not viable to manually edit the data without clear criteria. Ambiguity is also introduced by the concept of OA, as some sources allow filtering to content available in full text for free (potentially without reuse rights in perpetuity and therefore not an OA type), some do not have OA filtering at all and some have very granular metadata concerning OA metadata.

As described earlier, the information environment concerning OA books is heterogenous and an appropriate study design should take this into account to avoid drawing conclusions based on the circumstances of books included in just one of many information sources. As there is no single data source that would comprehensively list all currently available OA books or their metadata, sampling book records from multiple sources is an efficient way to get a good grasp of the characteristics of preservation coverage for materials listed or stored across different services.

In Table 1, an overview of the bibliometric sources containing records of OA books is provided, with columns describing the URL to the service, the search criteria used and the number of results it generated, the prevalence of ISBNs and DOIs among the results, and the point in time when the service was accessed.

Data were aggregated from a variety of sources, scoped both narrow and wide, and some having exclusively OA book content, while others are general-purpose bibliometric databases. OpenAlex (Priem et al., 2022), The Lens and OpenAire (Baglioni et al., 2022) can all be considered to be broad scholarly bibliometric databases that require a lot of filtering in order to narrow down to the specific information relating to the target group of OA books. The Lens and OpenAire offer no way to select specific OA mechanisms to be included/excluded, which means that they include all types of OA under one single content metadata tag which is not optimal for research purposes like the one currently at hand. This study is primarily interested in OA books made available directly through the publisher or similar primary distributor, as that has arguably a connection to having a potential preservation coverage compared to, e.g. self-archived green OA copies available on the web. Not being able to filter out green OA is also a drawback in attempting to minimize the occurrence of thesis ‘and dissertations ending up in the dataset since they are often hosted in repositories which are a location type almost exclusive to green OA. Nevertheless, the methodological choice was made to include the OA book content from The Lens and OpenAire since they are such substantial data sources, relying rather on post-hoc filtering out content based on content metadata tags identifying them as thesis and dissertation works. OpenAlex has the option to exclude Green OA/repository copies which were utilized to primarily obtain content that had been published OA directly by the publisher. WorldCat is also a broad database but offered quick ways to filter down to the relatively small number of records that related to this study directly from the webpage. The data from DOAB and Scielo Books was imported wholesale since they contain only data relevant to this study, with the exception of excluding individual chapters for a small part of the DOAB data.

Though the volume and quality of openly available metadata concerning OA books are better than it has ever been and is constantly improving, there are some obstacles to straightforward duplication-checking when data is aggregated from several complementary data sources. There is varying use of unique identifiers for books, where DOIs are mostly the key identifier used among bibliometric data providers included here. Matching only by title is not optimal due to even small differences in spelling, format and punctuation leading to incorrect matches, hence why that should only be considered as a last resort after primarily using unique identifier data. For performing the data cleaning and analysis Microsoft Excel version 16 was used. The exception was the OpenAire and OpenAlex datasets that first had to be imported into OpenRefine version 3.4 in order to parse the JSON data into spreadsheet format to become comparable to the rest of the data sources.

After the extraction of all data from the data sources, there were 750,136 items in total. Of which, 99% had a DOI/URL in the metadata; 6% had an ISBN number in the metadata; 0.2% had neither ISBN or DOI/URL. As the first step, all items with an identical DOI/URL or ISBN were deduplicated, which resulted in 430,745 entries remaining. The next step was deduplicating based on the exact title of the book, considering the first 250 characters of each title as per the restrictions in Microsoft Excel. This process resulted in a total unique record count of 396,995. As such the title matching was not a very impactful deduplication method relative to the unique identifier approach used in the previous step, but still offering some additional matches that would have otherwise remained in the dataset.

Preservation data

The challenges mentioned so far have concerned creating a comprehensive dataset of OA books, but none of the data so far can provide an indication of which titles are reported to be preserved through a trusted preservation service. CLOCKSS (2022), Portico (2022), Global LOCKSS Network (2022) and Cariniana Network (Márdero Arellano and Abbud Grácio 2021) all provide open datasets that describe which books they have included in their holdings. None of these four provide DOIs for their records, only ISBNs which is not optimal as most of the major bibliometric service providers focused on OA book content rely on DOIs. Table 2 provides an overview retrieved preservation data.

National libraries have good data within them but programmatic access from outside is still limited. Barnes et al., (2022) found that some OA monograph publishers deposit copies into national library holdings, something which would be very interesting to obtain more information about on a larger scale. However, the holdings of national libraries around the world are not easy to query programmatically from the outside.

Determining web domains for OA books

This study explored what domains host the OA book content, by checking which URLs their DOIs direct to when queried. This is not a way of verifying the preservation status of the books, but such an exploration can shed light on the nature and capacity of the long tail for OA book providers. An automated process of querying was set up in the Octoparse software application, simply recording the URL that was received as a response (if any response was received) when a web browser queried the DOI web address. This process was performed individually for the DOIs of the records in four of the six largest OA book data sources. In the case of Scielo Books, all books are hosted on the platform itself, and in the case of WorldCat, the number of DOIs found among records was low. In total, 745,525 DOIs were queried during June, July and August of 2022. The reason for the high number of DOIs in comparison to the overall number of unique records in the final OA books dataset (396,995) was that queries were performed for all records per data source and in parallel with the deduplication and analysis process. The results for this part of the study are also presented by the data source to better describe the content distribution for individual databases. In some individual cases, likely due to blocking frequent queries from the same IP address, the DOIs would not automatically be resolved. In such cases, the same web domain was recorded as for other records with the same DOI registrar prefix.

Matching OA book data to preservation data

In order to establish which records from the OA book dataset were also represented in the preservation dataset the common denominators were ISBN and book title, both of which were used to find matches in the datasets. Due to the lack of DOI data in the preservation datasets, matching had to be performed on only these two data points, where a match on either would be considered to indicate that the title was included in the holdings of a trusted preservation service. Some books were recorded with multiple ISBNs in both the book data and the preservation data, and matching was performed on all these ISBNs in all possible combinations.

Limitations and considerations

There is a need for further study into the correctness and data quality of the bibliometric data offered by the sources utilized in this study. For particularly the broader datasets, it is apparent that not all records classified as “book” or “monograph” are actually that in reality. This would also extend to the verification of the OA status and classification of content. Due to the size of the dataset and emphasis on the replicability, this study relies on the classifications given by the data providers with all the uncertainty that entails.

Books, similar to journals, can also be available in physical print format. For OA books, the specific practice of print-on-demand is also characteristic, meaning that there can be small quantities of printed OA books in circulation. What this study cannot establish is the print preservation status of OA books that are, or have at some point been, available to purchase in print form. It would be useful to gain more insight into the current adoption of print media for books that are available OA; however, establishing this knowledge reliably is best suited for a dedicated study. Another useful study would be on the inter-relationship of print and digital preservation practices in the library world. Digital preservation is needed in addition to print preservation if functionality and reader experience are to be preserved as well as the content itself.

One of the challenges that e-books introduce in addition to their storage and distribution media is the potential to integrate dynamic multimedia content rather than just static information. While it would be interesting to look more closely at the unique preservation challenges and risks associated with the content of such books, a wide-scale study like this is not suited to go deeply into these issues.

The Venn diagrams produced to visualize the data distributions were produced using DeepVenn (Hulsen et al., 2008). They are calculated to be area-proportional, however, due to the complexity of several overlapping datasets and the limitations of circular shapes that are not always completely possible. As such we advise the use of these visualizations to be approximations of the distributions. Please consult the exact percentages and numbers provided in Appendix for any exact calculations.

What is the current prevalence of OA books?

To answer this question, descriptive statistics for the query results from the six bibliometric data sources are presented (i.e. The Lens, OpenAIRE, OpenAlex, DOAB, WorldCat and Scielo Books), both individually and together as a deduplicated dataset. The breakdown of content identified through the individual databases was presented in the methods section, with the result of deduplication being 396,995 records.

Since all data sources provided a metadata field for the year of publication, an analysis of what age the content provided was from was performed. The results can be seen in Figure 1, where it is clear that a considerable share of content provided through The Lens (130,413 records), and to a lesser degree OpenAlex (30,979 records), had been published before 1975. There is only a relatively minimal amount of content published between 1975 and 1999 with a consistent growth trend starting from the year 2000 onwards. Since the amount of older digitized materials (published earlier than 1975) is so large in proportion to the latter individual years, Figure 2 presents a zoomed-in view of publication years from the year 2000 onwards. It warrants a mention that the data is not consistently capable of telling when a specific piece of content was made OA, only when the original work was published.

The next step for understanding OA book prevalence, and for informing future studies in the area, is to see how content is distributed across the different bibliometric datasets. Table 1 in the methods section already presented how many records each individual data source provided but that presentation did not analyze for overlap. Figure 3 presents the content distribution across the six bibliometric data sources with the overlap record counts across all data sources being available in Appendix.

From the results of this analysis, it is clear that there is a lot of overlap in content between the sources, which explains why almost half of all records were merged in the deduplication process (from 750,136 to 396,995). As Figure 3 illustrates, most unique records were contributed by The Lens (27% of the deduplicated dataset) followed by OpenAire (10%). The Lens shares substantial overlap with both OpenAlex and OpenAire, and 41% of records were found in all three. OpenAlex and The Lens are almost completely overlapping with under 1% (3,280 records) of the final dataset being unique to OpenAlex. What is perhaps a bit surprising is that despite the DOAB data being considerably smaller than the larger data sources included in the study, it still contributed 5% of the unique records for the final dataset. Based on this distribution, it would be hard to argue for selecting just one or two data sources as the basis of future studies in this space if maximum coverage of unique records is the target.

What web domains offer full-text access to OA books?

The next step in the inquiry was to find out what web domains the DOIs of the records point to when queried. The methods section describes the approach used (web scraping the 745,525 DOIs found in the four largest data sources used). The results are visualized as treemaps in Figure 4 with a summarizing breakdown of the top domains provided in Table 3.

While not providing exact numbers, Figure 4 conveys a general content distribution overview for the four data sources. Despite the difference in volume of records, they share the general trait that around half of the content, or somewhat over in the case of OpenAlex and DOAB, is hosted by three individual web domains. As Table 3 indicates, link.springer.com and biodiversitylibrary.org can be found in the top three web domains for three of the four data sources. biodiversitylibrary.org contains a large volume of digitized materials, with a lot of it being rich in images as it focuses on natural history literature and archival material. The prominence of link.springer.com in particularly The Lens is both due to Springer being one of the larger book and OA book publishers in the world, but also because some of the materials are separated as individual chapters since in some cases every chapter has been assigned a separate DOI, but at still picked up as entire books by the search function in The Lens. In the bottom row of Table 3 is a summary of the remaining web domains that did not fit the table, giving an indication of the “long tail” of content provided by these domains. DOAB is exceptionally clustered with only 7% of the content from 188 domains that did not have domains listed in the top frequency list. Content from the other three data sources was found to be much more widely distributed with 23–32% of content being held on 1,453–1,816 web domains that did not fit into the frequency top list.

Due to the sheer volume of web traffic already created by querying the almost 750,000 DOIs content analysis of page content or download of full-text copies was not performed, so this study is not capable of providing information about actual download capability/availability on the pages that are directed to. Though it would warrant a more detailed dedicated investigation, the long tail of web domains contains some clearly volatile services (e.g. Dropbox, Google Drive, organizational subpages, etc.) as well as some DOI addresses giving HTTP 404 errors indicating that the web page is no longer accessible.

To what degree is this content able to be verified to be included in the coverage of content preservation services?

This step of the study cross-matched the OA book records found from the six bibliometric databases with the data retrieved from the four preservation service providers. Table 4 presents a breakdown of how the content records retrieved from each OA book data source were represented in the various preservation services based on ISBN and/or book title matching.

Table 4 illustrates that OA book content listed in DOAB is covered to the highest degree by at least one of the services (46% of all DOAB records) with WorldCat (33%), OpenAire (25%), OpenAlex (13%), The Lens (10%) and Scielo Books (9%) following in descending order. Among the preservation service providers, Portico provides the overall highest numbers with 31% of coverage for both DOAB and WorldCat and also the highest numbers for any service for the rest. Matches were only minimally found in records in the Global LOCKSS Network and the Cariniana Network preservation data, ranging between 0% and 1% depending on the OA book data source. The high share of preservation coverage in DOAB is largely influenced by the collaboration between OAPEN and Portico, since the web domain analysis revealed that OAPEN (library.oapen.org) hosted 38% of all OA books in DOAB.

As a secondary perspective on the preservation coverage, Figure 5 presents a visualization of the uniqueness and overlap of preservation coverage for the deduplicated dataset of 396,995 OA book records. Portico provides 14% of preservation coverage uniquely, with CLOCKSS having 2% and sharing a 3% total coverage overlap with Portico. Overall, this analysis also provides the total coverage for the preservation of the deduplicated dataset based on the data sources utilized and compared to each other, which is 19%.

Recommendations

How should collaboration evolve among major stakeholders (e.g. publishers, libraries and preservation services providers) develop in order to establish higher coverage and flexible workflows?

It could be argued that OA content would benefit from OA status information for preservation, i.e. that there would be practices and data in place that would make it easy to both deposit and verify where specific pieces of openly available content are properly preserved. Concerning preservation data national libraries could on their own or through collaboration make available open machine-readable data concerning which books are preserved in their digital holdings. A service similar to The Keepers Registry that the ISSN International Centre maintains for journals would be very helpful for books as well, so preservation service providers could automatically report which titles they include in their holdings.

For future research, the open data produced by this research should help facilitate extended and deepened data-driven inquiries into the landscape of OA books. The study also functions as a detailed snapshot of the current situation on the entire spectrum, opening up for comparative studies in the future. The study lays an empirical foundation to develop theoretical connections between preservation and the concepts of time and temporality within library and information science (Haider et al., 2021). Preservation is through its inherent actions preparing for a future state in time, that in the best case scenario will not have to be utilized, and scholarly explorations in this domain would likely prove fruitful.