Roger C. Schonfeld, the Ithaka S+R Manager of Research, asked us several questions in a comment to a post here last week (Who Do You Serve and What Do You Do?: Defining Your Role to Defining Your Role to Ensure the Future of Our FDLP). These were good questions and we welcome the opportunity to spell out some specific ideas we have here at FGI about the future of the digital FDLP.
Thanks for your comments and your questions. (And thanks for noticing that broken link! I fixed it.)
Here are few quick responses to your questions. (I’ll answer the question about libraries holding overlapping or even identical digital collections separately.)
Q. Is your proposal that, as a condition of participation, every Federal Depository Library be required to hold locally at least some kind of sub-set of the digital/digitized collection?
A. No: No more than we require every depository to have fiche. I do think, though, that it is not too idealistic to imagine a future when every FDLP library (or most of them, or hopefully at least hundreds of them) will have digital collections that include FDLP content.
I think that digital deposit could become a standard part of depository status in a number of different ways, probably starting out slowly and building up over time. I think that the strength of the FDLP has been and will be its flexibility and a digital FDLP should have new kinds of digital flexibility.
The design of exactly how to achieve that should be a process, not an either/or choice today. I would like to see us use a process different from the one we used to get microfiche into the FDLP because that process took years to implement. I would prefer to see a phasing in of digital deposit that would allow FDLP as a whole to move faster even as individual libraries moved at their own speed. I am not a lawyer, but I think that a simple change to SOD 301 to allow for digital materials to be deposited and for depositories to select digital format would be an easy first step.
I would think that phasing in digital deposit would allow for the community as a whole to develop an infrastructure and skills and experience and a variety of models to accomodate different kinds of libraries and collections. Over time, I would guess, more libraries would develop the technical capacity to participate and would be eager to do so.
Q. Or, do you see a complementary role for “virtual depository libraries” (see for example http://govdocs.library.arizona.edu/policy.html) and similar models that would rely on GPO and/or other libraries as collections providers?
A. I would like to see a digital depository program that would be flexible enough to include all kinds and sizes of libraries. That will be its strength.
I think the key to digital deposit is that a depository library would have physical control (in the OAIS sense) of the material. I would think that a digital depository system would be flexible enough to allow partnerships between libraries as well, so that “physical control” might include partnerships among libraries. That might mean, for example, that two library partners (libraries “A” and “B”) might share physical control though a shared data center or a data center at one of the partners. The “control” would consist of their shared administrative decisions over what was in the data center, how that material is organized and described and preserved, and the methods of access to that material. I can imagine a single library participating in several different partnerships as well. For example, a library might be part of a national digital law library and a regional water-shed library, both of which might include FDLP materials.
I do not think that any library can currently claim physical control over anything in FDsys or at any federal agency.
I do think that there is a place for libraries maintaining links to digital materials not in their physical control and that this can enrich their services. But, since a) anyone can do that without being a depository library and b) the library has no control over the materials they point to, I don’t think that providing such a service by itself is an accurate or useful way to characterize or define what it means to be an FDLP library. It might be possible to design a digital FDLP in which maintaining links to materials in FDsys (for example) would be one of many characteristics or even requirements of a being an FDLP library. I can also imagine, using the example above, that an FDLP library “C” might point to materials in the possession of libraries A and B as well as having its own digital collection.
I can imagine a couple of things that GPO could do to make the just-pointing-to-FDsys “virtual depository” model a more reliable, more technologically sophisticated part of a fully functional twenty-first century FDLP: 1) GPO could make FDsys an “interoperable” archive (in the OAIS sense) that worked with FDLP digital repositories. 2) Once there are at least a few large FDLP repositories, the FDLP library community could become part of the GPO “succession plan” (in the TRAC sense of the term) so that materials that GPO could no longer keep or preserve or provide access for could be easily migrated to FDLP libraries for preservation and access. Individual FDLP libraries could also rely on GPO and, as we get more digital FDLP libraries, other FDLP libraries for their own succession plans. This would at least solve a bit of the problem of pointing to things not in one’s control: you could be at least somewhat assured that they were in the possession of the FDLP library system and not at the mercy of a GPO budget cut or mandated mission change.
Finally, in an FDLP with digital deposit, we could have something similar to what is possible with DOIs that would allow maintenance of pointers to more than one copy of any particular item at more than one repository. (http://freegovinfo.info/node/2947)
Q. Now that GPO is making more of the digital content available for bulk download, I’m curious what success looks like from your perspective
A. Bulk download is a great thing to have, but it is merely a useful method of delivery, not a system of or replacement for digital deposit. Almost certainly it would be a useful tool for a digital depository system, but it is not synonymous with or, necessarily, a precursor of digital deposit.
The strength and importance and sustainability of a digital FDLP will be found in those things that FDLP libraries can do or guarantee that they will do that are not done or cannot be done by others. I would think that digital deposit would include more than the ability to download in bulk. For example, a digital deposit might involve the commitment by the FDLP library to maintain FDLP-specific metadata generated at the time of deposit. That might include identification of the materials deposited as officially depository items, when they were deposited, version information, original source url, fixity information, provenance, and so forth. It might involved participation in a shared DOI-pointer system as noted above. It might include shared, reliable union catalogs and inventories and full-text indexes that could be built and maintained across depositories because of the commitments made by the participants.
Success of a digital FDLP would have to include:
– Digital deposit of digital materials into the physical control of participating FDLP libraries.
– Agreements between GPO and FDLP libraries, and, probably, among FDLP libraries, for preservation and access.
– Accountability of FDLP libraries for preservation and access.
– When fully implemented, a digital FDLP would, in the aggregate (across all participating libraries), guarantee that every Title-44 digital object would be in the physical control of the FDLP community in addition to or instead of the control of GPO.
Success could also be measure by other benefits that would, I think, also accrue (to GPO, to the public, to participating libraries) once such a system was in place. For example, it would be easier to develop and enforce new publication and metadata standards that would make preservation and long-term usability easier. It would be easier to create publication processes to identify and distribute instantiations of content (an obvious example: Census summary files) rather than trying to scrape dynamic web sites (which is almost always incomplete and unsatisfactory). It would be easier for the public to find and rely on government information they needed. It would be easier for libraries to build and maintain complete collections of information for their designated communities. It would be easier to keep track of and preserve “fugitive” government information and non-Title-44 information.
I don’t want to minimize the tasks at hand to do what I am proposing. But I also urge you not to be overwhelmed by libraries and librarians and, particularly, library managers who think this is an impossible or never ending task. This isn’t a unique situation. I saw a similar situation in when the “impossible/never-ending” problems were overcome and libraries embraced what they once did not: A couple of decades ago, there were very few numeric data collections in libraries, but today almost every ICPSR member is a library and there are many data libraries in libraries.
The software for building digital libraries and trustworthy repositories exists and is being improved all the time. To mention just one specific example, the LOCKSS-USDOCS project is already working toward the model of a successful digital FDLP library that I describe here.
Other events outside of libraries and overlapping with libraries (e.g. the open data movement, citizen mashup and tagging, new forms of scholarly publishing and open access publishing, institutional repositories, print on demand, mandated deposit of publicly funded data collections, the so-called data deluge, etc.) are beginning to affect libraries and their roles. Libraries already are embracing and building digital collections. While digital deposit might have seemed far-out and impossible in 1993, it is now within our reach.
At FGI, we see digital government information as the canary in the library coal mine. If government information librarians work on and solve the digital ingest/preservation/access issues for government information, their libraries will be able to generalize those solutions to other digital library collections. That would be a HUGE benefit to libraries in general and a big incentive for libraries to continuing to participate in the FDLP. Everyone wins.