week 5 notes: DL demos

Greenstone: after copying files to a collection, are they saved somewhere else, or just point to the original file? Will my DL be messed up if I move a file or delete from the desktop after copying it to the DL?

I might at some point try to install DSpace just for "fun" but I've got too much to do right now just doing required stuff.

week 3 muddy point (due 9/25)

I need some clarification on the term project. In lecture it was mentioned that to digitize things that you do not own the copyright for, you must obtain the copyright. This is probably not possible for us on this project. Therefore, should we only only be looking for Open Access items? If not, are we expected to scan whole books??

I have a very elementary grasp of encoding and use of "7 bit" or "8 bit" or "31 bits" for Unicode. when choosing which bit to use, is the point to find the smallest bit that will include as many characters as you need?

week 4 notes: Metadata (due 9/25)

Introduction to Metadata
• metadata may be specifically crafted to meet the users' needs of a particular user group
Although so much needs to be standardized (content, value, standards) inevitably there is going to be metadata assigned to items that is so specific that a casual user will have no use for it. In assigning metadata, should you also consider the casual user as well as a specific? (pg. 4 mentions "baffling hierarchical metadata").

•the multiple types and functions of metadata is useful (pg. 7)


witten5.4-5.7.pdf
• describing MARC records in detail (again recap of Organizing Info class), good information
• Dublin Core's purpose is to describe electronic materials
• BibTeX: scientific and technical (TeX)
• Refer: computer scientists
• TIFF used for today's digital libraries
• TIFF tags include information about the image, size, color, optional creator
•"most digital library projects...use the TIFF format to store and archive the original captured items, even though they may convert them to other formats for display." So do digital libraries have 2 formats of the same image, one to hang on to and the other to access via computer? How does that affect $ and storage for things like ArtStor?
• the speculated uses for MPEG-7 sound nuts

witten2.2.pdf
• gives some basic "whys" of metadata
• "entity" encompasses all sorts of items, not just books
• "helpful digital libraries need to present users with an image of stability and continuity, as though electronic documents were indentifiable, discrete objects like physical ones"
I wonder why Witten thinks that electronic documents are so fluid? I don't think this is necessarily so - an mp3 may be removed from a digital library, but so may a book from a traditional library. Electronic documents are not *that* unstable, maybe just able to change or replace more quickly than a traditional book. If an update is replacing an older document, the metadata can easily show that and even include the older document alongside the newer. Witten addresses this idea a few paragraphs later, seems to contradict self.


Border Crossings_ Reflections on a Decade of Metadata Consensus Building.pdf
A recap of the development of Dublin Core...more in depth than the introduction from Organizing Information class, but not a whole lot of new information.

annoying.

it looks like Greenstone only runs on Mac 10.5 Intel, which I don't have.

DSpace. No idea how to interpret the installation instructions. I doubt I have SQL or
Oracle and the 'configuring' stuff is beyond me.

Fedora says it has a version for PPCs. It's an ISO image of a DVD and is going to take a couple hours to download, and try to burn a CD from the image...it says on the webpage I can burn it onto a DVD *or* a CD, but a CD doesn't have enough space.

Now what?

should I just drop the class now?

week 3 notes: Representation (due 9/18)

Arms chapter 9:
• interesting points about computers vs. books. For my online classes I save all PDFs or print webpages to PDF so I may mark them up with Skim as if they were paper copies. I can't take the time, paper, or ink to print all that stuff out.
• Oxford English Dictionary SGML project very interesting
• I was interested in the description on OCR and scanning resolution since I have been using this technology for a number of years.
Manually retyping stuff may be cheaper but Arms clearly states it is often outsourced
• the discussion of binary numbers is beyond me
• example of SGML? I got HTML. XML? and I didn't really get CSS to do what I needed to do for 2600 which was frustrating.

Lesk chapter 3:
• ironic that this starts out talking about text documents as images, because that's exactly what this PDF is - which means I can't use Skim to highlight and take notes :(
• a 2nd annoyance with scanned books is although 2 pages can open and be scanned together, it takes more effort to navigate around the page when it becomes a PDF. I must go left-right as well as page up and down.
• "many people of course still prefer to read from paper"
• compares storage costs for real books and digital depositories

lesk-cha2.1-2.2-2.7.pdf
• Evolution of typesetting to computer document prep and printing
•"SGML is a syntax and a philosophy" The Huck Finn example basically looks identical to HTML, but HTML can do links. Does it have more complicated tags? or can you make up what you want and what tags apply to?
• searching and abiguity

Paskin
• Digital Object Identifiers: abstract basically describes a DOI as a handle
• not content, but location. URLs do not fulfill handle requirements
• DOIs are different from other handles "by metadata and policy." Changes are made to a DOI directory if the URL changes.
• "Since such persistence requires a social infrastructure, policies as well as technical infrastructure need to be defined and implemented."
• The DOI system was the result of a publishing industry initiative in the late 1990s, which recognized the need to uniquely and unambiguously identify content entities, rather than refer to them by locations

Lynch
• "Identifiers take on a new significance in the networked environment. To the extent that a computational process can allow a user to move from the occurrence of an identifier to accessing the object being identified, identifiers become actionable."
• URLs that tell web browsers what to do - HTTP, FTP
• URN doesn't specify an access service like a URL
• Example of a URN??
• When developing new identifiers a significant number of questions need to be addressed

week 2: muddy point (due 9/18)

Are the 3 dimension / 4 dimension / 5S models something to actually follow when building a DL, or are they used for analyzing or verifying a DL?

Hands on Point 1:
Stream: images, text
Structural: hypertext, not really metadata but descriptions
Spatial: not sure what this means
Scenarios: "most beautiful" artifacts chosen by the national libraries of Europe, administered by the European Library? Users browse, no searching.
Societies: not a comprehensive collection, so use limited to those interested in browsing but probably not enough stuff for a lot of research.

Hands on Point 2:
• what is the handle for current Web pages?
URLs

• Any problems with it?
filenames may not be unique, many hosts can have duplicate filenames
Some URLs are long and complicated, or buried in many directories, may not be easy to obtain or resolve.

• Can URL act as the identifier of online digital objects?
Not a good one! a URL can change or disappear without notice

week 2 notes: Architecture (Suleman, Arms, Payette, ARMS ch2)

It seems that many agree that there is a problem with "interoperability" in DLs. My question is, if there are systems available that someone has developed, why not use them? It is expense? I'm looking at the OMI-PMH page and the description seems reasonable. Why would anyone with a difficult-to-use DL not want to use this protocol?
Maybe I am totally missing the point here.

Suleman states that "while technically sound, [Dienst] required an investment in software, methodology, and support that some prospective users were not willing to make" (2). I imagine this is the perspective for many on why one thing or another is adopted, it is probably cheaper to try and develop something on your own although it causes problems later.

I am still kind of unclear how all this really relates to real life DLs - the theory is good for background but some real-life examples on how DLs do not "interoperate" would provide some clarification for me.

The Payette article was too much "computer guts" in many places that was beyond my understanding. My notes on the PDF again say a very basic example would help me a lot.

week 1 notes: Introduction (Candela, Borgman, Paepcke, Lupovici)

the DELOS manifesto makes an attempt to include all the facets of a digital library, in spite of numerous opinions on what a digital library is and should consist of. The Paepcke article describes the interruption of the WWW in the DLI - seems that the insertion of rampant capitalism, lack of clarity on ownership and copyright issues intervened with the vision of the project. Apparently computer scientists and librarians have a relationship that is similar to engineers and architects. I like Borgman's definition of a digital library - it could be used to determine if something IS a DL, or a guide for creating a DL.
Lupovici states "The most important sector requiring librarians in the future is preserving digital documents in the long-term." I disagree, feeling that the primary purpose of a librarian is to help people find the most relevant material, not simply preserving it.

week 1: muddy point (comment, really)

After looking at the graph showing the drop in storage costs for digital materials, I'm wondering how the "life" of digital materials compares to that of paper and film. In the past microfilming craze people discovered that microfilm is not a permanent medium and deteriorated badly, and of course poorly archived paper crumbles too. Are there any statistics comparing the life of a digital item and the life of a paper/film saved item?