|
Metadata Home > Metadata Steering
Group > Minutes
January 12, 2005
Present: Sherry Lake, Erin Stalberg, Ann Whiteside, Thorny Staples, Bob Thomas, Janis Kessler, Judy Thomas
Tibet Image Database
- In continuing the MSG's review of the Library's various image databases and their content creation (cataloging) tools, Judy Thomas demonstrated the Tibet Image Database. The Tibetan images are cataloged using FileMaker and its FileMaker Entry Forms. There are different entry forms (layout screens) for varying types of metadata entry detail. There is a form for entering very basic information (creator information and file number) for batches of slides. There is also a very detailed form divided into 4 sections: "Content Analytical Metadata," "Geographical Metadata,' "Administrative Metadata," and "Technical Metadata." The Judy handed out documentation for filling out the detailed form.
- Although there is no controlled vocabulary in the THDL, there is a 3-tiered thematic hierarchy managed through pull down menus. IRIS allows subject term hierarchy, but does not have pull-down menus, based on hierarchy, with choices.
- The MSG discussed how GDMS could represent this information and talked about the many "views" GDMS could provide. The images could be grouped, using GDMS, to represent views based on time, trip, place, buildings, etc.
- As for the assessment of using IRIS for cataloging the Tibet images, it seemed like there is a baseline of metadata for these images that would need to be identified for minimum requirements. IRIS could handle the minimum requirements, but David Germano's needs require that there still be advanced metadata tools on top of the baseline to provide additional information for his world-wide scholarly collaboration. Because of this collaborative environment, the Tibet database provides field level administrative metadata (who created the metadata field and date). IRIS only provides record level administration metadata.
- Further more, the geographic information as captured in the Tibet database would be hard to accommodate in IRIS. IRIS cannot not handle multiple location information such as discovery location, former, or present location. \
January 20, 2005
Present: Sherry Lake, Erin Stalberg, Beth Picknally Camden, Ann Whiteside, Judy Thomas, Thorny Staples, Bob Thomas
Geostat Aerial Photos
- As part of its image assessment, this week the MSG looked at the scanned images of the Aerial Photographs from Geostat. Sherry showed the Aerial images, described their metadata and its creation and the images' search and delivery method. Geostat recently scanned its collection of 1974 and 1966 aerial photographs of Albemarle County. This collection is comprised of nine-by-nine inch black-and-white images taken at a scale of 1:40,000 and 1:20,000 respectively. Both TIFF and JPEG format images of the photos were created within the library by Digital Library Production Services (DLPS). The JPEG images are accessible through the Geostat website (using a MapServer interactive interface) while the full resolution TIFF images are available on DVD-Rom in the Geostat lab.
- These and other map images in Geostat were not cataloged using a database (Filemaker) or a database tool (such as IRIS). For the maps, Sherry started with "series" templates created by the USGS. Using a Perl script, she then created FGDC xml files (metadata files) for each image. For the aerial images, Sherry created a "series" FGDC xml template using the cataloging information from the aerial photographs' MARC record. Then using a Perl script, and the data from the IVIEW dump (created by DLPS when scanned), she created individual FGDC metadata files for each image (and each format, JPEG and TIFF).
- With the current map and photo images already cataloged programmatically (and pretty much stable), the MSG decided that it did not make sense to add these images to a central image cataloging tool.
Miscellaneous Comments
- Judy questioned the MSG about DescMeta and the minimum requirements. The minimum requirements (from the UVa Community Metadata Guidelines Quick Chart - http://www.lib.virginia.edu/digital/metadata/communityquickchart.html ) are Title, Agent, Date (of creation of the content and/or of the creation of the digital object), Content Type, Identifier, Access Rights, and Resource Type. Of these, Content Type, Access Rights, and Resource Type could be created on batch. The MSG then discussed who decides what can go into the Repository. The MSG agreed that there should be different selection criteria for Faculty projects and for purchased material. Erin pointed out that there was a Selection Form being developed that will help in the assessment of DL "readiness" and inclusion in the Repository.
January 27, 2005
Present: Sherry Lake, Erin Stalberg, Beth Picknally Camden, Ann Whiteside, Thorny Staples, Bob Thomas, Janis Kessler
Follow-up Relationship Discussion
- From the 12/17 meeting's discussion on conceptual relationships, the MSG decided to create the element <corpgroup> so David Germano could use it to describe religious sects. This term would be used (until a better nomenclature was decided). At that meeting, Bob Thomas was asked to research possible relationship terms. David Germano is using the <corpgroup>.
General Discussion on Metadata Content versus Search Interface
- The MSG discussed how do users get stuff out of the DL, if it's not explicitly in the content. Problems with cross-collection searching and different thesauri/terms were discussed. If certain terms were used in one collection, a search interface would need some behind the scenes manipulation to normalized the search term across collections, or present the user with some "you may also like to search ?.. this term". The search interface can either do a subject normalization and present its results, or let the user have a more interactive interface to make decisions on their own.
- Many ideas were discussed from creating a GDMS object for people or objects and their associations; having different search interfaces for the whole collection, and different domains; building a UVa DL thesaurus, designing a search interface that could deal with misspellings; creating a fun and user driven interface; having the interface articulate the scope of the DL and its parts; capturing user keywords and including them with the objects for others to use to search; and using the controlled lists from all collections together (via data mining the subjects and schemes), normalizing them and putting back this new term.
Preservation Metadata
- Erin handed out the Appendix from "A Metadata Framework to Support the Preservation of Digital Objects" from the OAIS Information Model. She wanted to know if the DL's AdminMeta (with administration metadata and technical metadata) fit this framework. Thorny pointed out that this document was only a guideline. Thorny said we should probably discuss (for future discussion after completion of the image assessment) putting this information in each object's AdminMeta too and setting standards for metadata for the disseminators. The Preservation Metadata describes the structure of the object (which could possibly be automated) and the semantics of how a human is suppose to use the object. We discussed whether the Fedora architecture would keep track of relationships and versioning.
February 3, 2005
Present: Beth Picknally Camden, Ann Whiteside, Thorny Staples, Erin Stalberg, Bob Thomas, Judy Thomas, Sherry Lake
IRIS Demo
- Ann led us through IRIS entry pages and general discussion on how images are currently being cataloged in the Fine Arts Library.
- Image Resource Information System (IRIS) is a collection and catalog management system originally developed at Brown University (in 1999) for images. It uses FileMaker 6 as its underlying database. On initialization, Authority files are activated. The initial discussion focused on the maintenance of the authority files. The Geography authority files (3 separate ones) were received with the original software from the IRIS developers. We discussed whether the TGN (Getty Thesaurus of Geographic Names) was available via a web server. There is a place for links for web authority files in IRIS; this opens a new browser. Ann pointed out that the subject authority file has duplicate entries from the creation and geography files such that their contents can be searched as subjects. Many terms are duplicated this way.
- Entering data in IRIS starts with creating a Source Record that contains bibliographic info on where the image(s) came from. for the image(s). Problem here is that you must tell it how many images there are from this "source." If you later get additional images, from this one source, you have to create a new (duplicate) source record for them. Next a Work Record is created for each image. There are 5 "pages" of potential information, some of the fields use the values from the authority files. Finally, there is the Image Record; some of the fields are automatically filled in from the Work Record. The Image Record contains mostly general descriptive content for the specific image.
- Lastly, we discussed whether we want a central image cataloging tool and if so, do we want to use IRIS. These are 2 separate questions.
February 10, 2005
Present: Beth Picknally Camden, Ann Whiteside, Thorny Staples, Janis Kessler, Erin Stalberg, Bob Thomas, Judy Thomas, Jack Kelly
SnapDragon
- Ann led a demo of SnapDragon, the image cataloging system developed and used by Smith College. Smith is interested in exploring the possibility of a development partnership with us if we choose to leave IRIS. The group was impressed with the software. It seems cleaner than IRIS and more directive in terms of guiding the cataloger through the records. SnapDragon also seems to address a number of the issues that the art & architecture community has with IRIS.
- The MSG had the following questions for Smith:
- When you change an authority record, does it update the corresponding work/image records? If so, then do the fields need to match? i.e. if you use a Culture authority heading in the Work/image Subject field, will it update?, or only if you use it in the Culture field of the Work/image record?
- Display date: is that texty field repeating or is it just one big text box? What is happening when you hit "enter" in that box to go to the next line?
- To pull together collections, do you need to create a Work record for the collection?
- Can we change the workflow steps on the order record to reflect our workflow?
- Things we think we'd like:
- Date certainty ... (or uncertainty) ...
- Repeating texty dates (see #2 above)
- longitude/latitude for location
- loading in or using web services to connect to authority files.
- To change the workflow steps on the order record to reflect our workflow? (see #4 above)
- A change in the display of the hierarchies in the authority records, so that the term you are interested in is always in the same place on the screen. Your eyes have to get reoriented each time you change the hierarchy.
- Ann will take those questions back to Smith and also engage in a conversation of what Smith means when they say they are "are seeking a co-developer." Are they looking for programmer support? Or money? How much money, etc.?
- Jack will take a look at the Filemaker database structure to compare it with IRIS and consider whether SnapDragon will be easier or more difficult to load/extract records.
February 15, 2005
Present: Sherry Lake, Erin Stalberg, Ann Whiteside, Thorny Staples, Bob Thomas, Judy Thomas, Beth Picknally Camden
Update on Snap Dragon
- Jack Kelly is looking over Snap Dragon. Its platform is a Mac Server, thus Snap Dragon is cleaner on importing and exporting, but necessarily easier than IRIS.
Plant Images in IRIS
- Landscape Architect images of plants have been cataloged in IRIS. There is a Work Record called "Plants". Each Image record has more detailed description of the image and plant.
- Ann talked with Nancy Takahashi to see what the Landscape Architecture Department wanted in cataloging their images. This information was applied to the IRIS database. Nancy Takahashi set up the hierarchy for subjects. To create a GDMS "collection" record, the images with "Plant" Work Records could be pulled out. It was pointed out that the plant images in IRIS had a horticultural point of view not a biological view point. For example the locality information in IRIS would be more like "shaddy, eastern Virginia" than an exact location where the plant was found. The locality is more general for the Landscape Architecture plants and most of the times come from a plant reference book.
BIOTA Briefing
- Eric Nagy, faculty in Biology and Associate Director of the Mountain Lake Biological Station, has proposed a project with the SEL Lab to catalog and create digital images of a selection of the biological specimens. They plan to bring the specimens to RMDS in groups to be imaged. They will undertake item-level cataloging for each specimen, using a system that they already have in place called Biota, designed specifically for biological specimen cataloging that is commonly used in the Biology community.
- As Biota is a licensed product and installed only in SEL and at Mountain Lake, the MSG could not see a demo of the product. The MSG looked at web pages that were output from Biota < http://viceroy.eeb.uconn.edu/biota >. It was decided that IRIS could probably accommodate this collection, but would need authority files for Division, Class, Order, Family, Genus, and Species. Individual specimens could be handled like the proposal for Holsinger, as individual one offs. There could also be a work record in IRIS that serves as an authority record for the plant species. Those records could then be related to the particular specimens. Like in IRIS, there would be a great benefit in creating relationships between records for the species and records for the particular specimens. One possible problem in using IRIS, you lose the specific Biology Classification and Biota's ability to automatically create relationships. Those entering in data would most likely be Biologist who would be more comfortable with a cataloging tool that uses their jargon (i.e., Biota). Also the plants in IRIS are now cataloged with a "horticulture" view point rather than a biological one with biological classifications.
- Concluding remarks were that Snap Dragon and IRIS are currently biased toward Art and Architecture images. Should we use one of these for base level metadata creation?
February 22, 2005
Present: Sherry Lake, Erin Stalberg, Ann Whiteside, Thorny Staples, Beth Picknally Camden, Leslie Johnston
OAI-DC Mapping Review
- For the American West Project, selected UVa metadata records (DescMeta) were mapped to Dublin Core for OAI metadata harvesting. Last October the California Digital Library reviewed our DC metadata as well as all partner institutions' metadata. Leslie presented a synopsis of the results of that review. She also presented notes from a new document that the American West team distributed 2 weeks ago: "Baseline Metadata Requirements." This document contains the standard requirements and guidelines on what the OAI search interface expects. They do say that compliance is NOT required for project participants. That said, their document lists the REQUIRED elements and the categories of data that MUST be present and which elements they map to. Leslie has asked for an interpretation of the highest prioritized requirements.
- One problem reported was duplicate subjects. Duplication of subjects was discovered with GetChild Descmeta and corrected. De-dupping needs to be done for the DC mapping also. Even though the OAI mapping is unqualified DC, another comment on our UVa data was that there were too many duplicate elements that were not qualified. One suggestion (from the OAI review) was to join all instances of an element into one element, or to put just select the single most important mapping.
- For <dc:date>, unqualified Dublin Core is the only option for delivering OAI data for harvest, it is preferable just to include the date of creation, and not the date of digitization of the object, or other dates not essential for the end user, thus <surrogate><date> was dropped from DescMeta-DC mapping. If there are two <date>s inside a single <time>, concatenate them together with a hyphen in between (i.e. 1912-1930).
- The mapping was picking up all <title> elements, so it was decided to only use <title type="main"> concatenated with <title type="sub"> (if it exists). If there is no "type" on <title>, use the first instance of <title>.
- For <dc:type>, the OAI guidelines suggest the medium of the original item be mapped to <dc:type> instead of <dc:format>. The MSG decided to ignore this suggestion keeping our mapping to <dc:type> and <dc:format> as we have documented it.
- The last comment from the review was regarding no spaces in the <dc:contributor> field. This is a TEI-to-DescMeta mapping problem and not a DescMeta-to-DC mapping problem.
- To allow OAI to harvest chosen sets, OAI Set descriptions must be used. The element <setSpec> will be mapped to the <set codes> element. More than one <setSpec> is allowed. The element <setName> must also be included. This needs to be the "human readable" version of our set codes codes.
- Leslie questioned whether the statement "Copyright UVA Library" in <dc:rights> was accurate. It was verified "yes," this is the copyright statement for the metadata file (surrogate). With the OAI data, we are talking about the content and not the metadata file, so we are not mapping <rights type="copyright"> for the surrogate. As for the other surrogate <rights> elements, type="access" and type="use" will be mapped. For top level <rights>, all are mapped. An additional <rights> statement will be added pointing to a URL with the UVa Digital Library general rights pages.
- Two other OAI suggestions were nixed by the MSG. In one, it was suggested that AAT subjects be mapped to <dc:type> instead of <dc:subject>. This met a most definite "no" and resounding "stupidly bad idea." The second suggestion on mapping the medium of the original item to <dc:type> instead of <dc:format> was also rejected.
- One last comment from the OAI harvesting, Leslie noticed that there were no <dc:coverage> elements. There are <geogname> elements in DescMeta and mappings set up, but the stylesheet translation must have missed this.
Credit Line/Sponsors
March 15, 2005
Present: Sherry Lake, Erin Stalberg, Ann Whiteside, Thorny Staples, Beth Picknally Camden, Judy Thomas, Bob Thomas
Work Records and Images
- We confirmed what was said in e-mails per Leslie's question about Work Records and Images. In cataloging plants in IRIS, each plant (Acer Maple, for instance) is a "work", with images. Plants could also be a work of works and one image could belong to more than one work. For Holsinger images, the Work described is the photograph, and that an image record is still needed to record metadata about the digital image of the photograph. In the image record, one record's administrative metadata, i.e., that it is a digital image. A work record for each "genre" (WW I, UVa, Charlottesville, African American), a work of works, could be created from the individual works.
Image Data Tool
- Lots of discussion about what type of cataloging tool do we need: a staff tool (like Workflows) that staff use, or a tool that staff and faculty could input into.
- A unanimous opinion is that one cataloging system cannot solve all needs. Different collections need differing amounts and types of metadata. The MSG also discussed the feasibility of having a central tool, where Faculty could contribute their records. The central tool would only capture standard fields. Problems here would be when the Faculty member updated their database record, the central one would not be updated.
- Another problem with using a central catalog system is field names (tags). Having an input interface that would vary based on community specific terms would make it more useful. There could be more than one database within one cataloging system.
- Finally we agreed that a central image cataloging system is desired, for use by all library staff unless there is a good reason not to use it.
April 19, 2005
Present: Sherry Lake, Erin Stalberg, Ann Whiteside, Beth Picknally Camden, Judy Thomas, Bob Thomas, Janis Kessler
Review of Central Image Cataloging Tool Proposal
- In November 2004, Martha's Managers charged the MSG with making a recommendation for the utility of a centralized image cataloging tool. MSG starting reviewing various UVa Image databases and collections in January 2005 and in March 2005 had finalized their assessment in the "Image Collection Assessment" document.
- The MSG discussed a draft proposal for a central cataloging tool for all library staff use. Slight modifications were made to the draft. The document, Central Image Cataloging Tool (CCT) for All Library Staff Use, will be available and sent to lib-metadata listserv once approved by Martha and her Managers.
- Erin brought color charts of the proposed metadata paths for images. She took the group through metadata creation, mapping, dumping into a Central Cataloging Tool, and finally, ingestion into the Central Digital Repository. Faculty will be using 2 different tools to catalog their collections, Jack's User Collection Tool and iView. Mapping other sources of images and metadata from vendors and library-created legacy sources to the Central Image Cataloging Tool were also discussed.
- Further discussions lead to a list of unresolved issues: can there be just one mapping of the output from Jack's User Collection Tool; do we care about getting the data-enriched (enriched in the Central Cataloging Tool) metadata back into the faculty tool; how do we refresh already ingested faculty data, with new data or changed metadata; what will we do with our legacy collections that will still need to be published and delivered through a library server (outside of the Central Digital Respository) for some time?
- Erin is invited to the next meeting of Martha's Managers to present the proposal. Once approved, the "Central Image Cataloging Tool Proposal" and the "Image Collection Assessment" document will be made available on the metadata website and through the lib-metadata listserv.
April 28, 2005
Present: Sherry Lake, Erin Stalberg, Ann Whiteside, Beth Picknally Camden, Janis Kessler, Thorny Staples
TEI Multi-Volume Numbering and Sorting Issues
- Some issues of the Cavalier Daily have Misprinted Volume/Numbers. In a browse display by Volume/Number, a patron would be confused by seeing 2 (or more) Volume 78, no. 19. The MSG discussed the best way to correct this in the TEI, but still keep what exactly appears on the paper. It was decided that the TEI headers would be changed to:
<biblScope type="volume"><num value="78">Volume 78</num></biblScope>
<biblScope type="issue"><num value="20"><corr sic="19" resp="mgp2c">No. 20</corr></num></biblScope>
The display style sheet will render the original form with the correction (as seen by the <corr> tag) on display; something like this:
No 19 [i.e. No. 20]
- Another problem encountered with muti-volume sets is when the numbering for a series starts over, such as the UVa Record that had volumes 1 - 7 published and then started over numbering the volumes with "New Series, volume1." This could cause potential problems with sorting. It was decided to keep the browser sorting program the same; this is not a wide spread problem.
Presentation of CCT Proposal
- Erin presented the Central Image Cataloging Tool (CCT) Proposal at Martha's Managers meeting. It was approved. The criteria for UVa Library's CCT (point 1., bullets a - h) will be discussed with Smith College, the developers of SnapDragon, an image cataloging system. Ann Whiteside has set up a video conference call between Smith and the MSG Team to discuss these points as well as others gathered during MSG demos of SnapDragon and IRIS.
- The "Proposed Metadata Paths for Images" was discussed at Martha's Managers meeting. It was suggested that a new workflow be created such that there was one input to IRIS and then this metadata would go in to a Central Image Cataloging Tool. The MSG needs to be thinking about different kinds of metadata workflow for "non-big" batches of just metadata.
Brief Overview of Simile
- Thorny gave the MSG a brief overview of SIMILE. SIMILE (Semantic Interoperability of Metadata and Information in unLike Environments) is a joint project conducted by the W3C, HP, MIT Libraries, and MIT CSAIL. The SIMILE Project is working to make it easier to wander from collection to collection and to find your way around in the "Semantic" Web. SIMILE uses RDF technology on metadata to break it down into triplets (subject, Predicate, object) which are related to each other. Using these relationships, crosswalks to-from different metadata schemes can be developed as well as linking Digital Library objects to each other.
May 11, 2005
Present: Erin Stalberg, Ann Whiteside, Janis Kessler, Thorny Staples, Judy Thomas
Snapdragon Conference Call
- The MSG engaged in a conference call with Smith College. Smith is interested in finding a "development partner" for their image cataloging tool, Snapdragon. The discussion was very general to gage whether or not we have similar image cataloging needs & vision. The conversation was very positive and it was evident that both institutions are very much on the same page. Discussion points included the following:
- Background on Snapdragon
- Smith's vision for it's future
- Beyond Art & Architecture
- Technical environment
- Platform: FileMaker 7.0?, XML?
- Vocabulary service: Webchoir tools & other authority control needs
- Other tools involved: Insight, Metalib (Smith), UVa's tools
- Funding
- Intellectual Property
- Next Steps
- The MSG has a next meeting May 12, 2005 to debrief.
May 12, 2005
Present: Sherry Lake, Erin Stalberg, Ann Whiteside, Beth Picknally Camden, Janis Kessler, Thorny Staples
Debriefing on Conference Call with Smith
- On May 11th, members of MSG (Ann Whiteside, Thorny Staples, Erin Stalberg and Janis Kesslier) and interested parties from Smith College discussed the image cataloging tool, Snapdragon. See the notes sent to the Metadata-Lib mailing list from Erin Stalberg. During the debriefing, the MSG group talked specifically on their impressions of the conference call and on the interest Smith College has in bringing in a development partner.
- The MSG talked in general about upgrades, improvements and the time all of this would take for IRIS and for SnapDragon. In deciding on either program as a Central Cataloging Tool, there would be costs (as it would be for any "new" image cataloging tool). The MSG outlined the following 4 options for choosing a Central Cataloging Tool:
- Take IRIS as is continuing to work with the consortium (and with Brown upgrading it to FileMaker 7 with "some" enhancements.
- Take IRIS and move IRIS where we want (UVa customizations done by UVa programmers).
- Move to another existing cataloging tool (primarily SnapDragon).
- Do in-house development (likely as an xml tool).
- The MSG also layed out a simple cost-benefit analysis (lines specific to SnapDragon are preceded by "S", lines specific to IRIS are preceded by "I"):
Cost of Moving to SnapDragon OR Staying with IRIS (and upgrading to FileMaker7)
S Re-developing SnapDragon ?
S Upgrading to FileMaker7 Plus enhancements
I Upgrading IRIS to FileMaker7 (need FileMaker7 licenses)
I OR "not" upgrading and “finding” FileMaker6 licenses
S WebChoir Fee
S Software
S Data from Getty
S Development for use with other UVa DL tools
S/I Local Development for use in DL
(Cost with IRIS depends on version 7 table structures)
Importing/Exporting data
S/I Design Time
S/I Migration
(Cost with IRIS depends on version 7 table structures)
Stop work time
Mapping/Programming
Data Prep and Clean up
- Using this Cost-Benefit Analysis with the 4 options above:
- Option #1 is deemed not to be attractive, as the IRIS consortium is not interested in expanding its scope much beyond Art & Architecture and is not currently meeting our needs.
- Option #2 is deemed not to be terribly attractive either, as it would require a significant Filemaker programming investment & cost on our part and probably still leave us with many limitations (caused by the underlying architecture of IRIS).
- Option #4 is deemed not to be a short-term option. Development would likely be many years off, require significant funding (such as a grant), and would compete with other DL development priorities.
- Option #3 , therefore, is most attractive to the MSG. Particularly in relation to Snapdragon, where Smith and UVa seem very much in line regarding functional requirements. At the conference call, Smith said they would send Ann their functional-requirements/wish-list document for the next version of SnapDragon. When we get that, we will compare our needs.
May 27, 2005
Present: Erin Stalberg, Beth Picknally Camden, Ann Whiteside, Sherry Lake , Janis Kessler, Judy Thomas
Feedback from the Imaging Group
- Discussed feedback from the Imaging Group minutes.
Smith Conference Call
- Further discussion was held on points resulting from the Smith Conference call concerning Snapdragon.
June 2, 2005
Present: Erin Stalberg, Beth Picknally Camden, Ann Whiteside, Sherry Lake , Thorny Staples, Janis Kessler, Judy Thomas
CCT Proposal
- Revised CCT Proposal unresolved issues section to include staffing issues.
SnapDragon discussion
- Video collections
- Video clips—i.e. video news, video interviews, video movies
- Maintain flexibility for hierarchy and relationships so moving images can be accommodated.
- Requirements
- General—types need to be available on any field and the system needs to be configurable to require fields.
- Authorities—Include MESH
- Longitude/latitude data—allow multiple schemes; type=point; type=bounding box; allow repeatable values.
- Works, Date needs
- Change “year” to date; need qualifiers and typing on the pair of dates; texty dates—change to: dates expressed as phrases.
- Additional UVa needs/questions
- Permissions—add record editing/creation
June 6, 2005
Present: Erin Stalberg, Beth Picknally Camden, Thorny Staples, Janis Kessler
getCitation and getTEIheader
- Worked on disseminator mappings for getCitation and getTEIheader. We agreed to map getCitation from the DescMeta rather than TEI/EAD/GDMS, so that DL citations will be consistant across collections. We chose the MLA citation format. Mappings will be distributed when they are finalized.
- getCitation from DescMeta (not TEI, EAD, GDMS)
- Objects that speak for themselves, fields for the disseminator mapping:
- <agent type=”creator”><name>
- <title type=”main”>
- <identifier type=”UVA collection number”>
- <place type=”publication”><geogname>
or
<place type=”current site”><geogname>
or
<place> with no typing, first instance only
or
<place> if typed something else besides “publication” or “current site”, but only first instance
- <time><date type=”publication”>
or
<time><date type=”creation”>
or
<time><date> with no typing
- <agent type=”provider”>
- <surrogate><time><date type=”creation”>
(year only, sometime it's in format: 20040607)
- today's date (in format: 07 June 2005)
- URL (from default GetFull View disseminator) with time/date stamp to retrieve version as of citation date.
- (note: <identifier type=”UVA collection number”> will only exist in the DescMeta generated from EAD files)
- Stylesheet:
<agent type=”creator”><name>. <title type=”main”>. <identifier type=”UVA collection number”>, <place><geogname>, <time><date>. <agent type=”provider”>. University of Virginia Library Digital Collections . <surrogate><time><date type=”creation”>. University of Virginia . Version from today's date: <URL from default GetFull View disseminator>.
- The text in italics is not coming from the DescMeta. It should be supplied.
- The text in italics “ University of Virginia .” (before [today's date]) should only be supplied if the following does not exist: in the DescMeta:
<agent type=”provider”> Special Collections, University of Virginia Library .
- If an element is repeating, use semicolons to separate values.
- If creator is “Unknown” start with title.
- Punctuation goes with the element that precedes it. So, if the element does not exist, the punctuation that follows should also be left out.
- Objects that do not speak for themselves, as we have been doing with other child objects (in getChildDescMeta):
- for GDMS children:
Concatenate <title type=”main”> : <res><title>
- for TEI children:
Concatenate <title type=”main”> : page [value of n= (page number)].
June 9, 2005
Present: Erin Stalberg, Beth Picknally Camden, Thorny Staples, Ann Whiteside, Sherry Lake, Janis Kessler
CCT Functional Requirements
- Continued work on the CCT Functional Requirements document
- General functionality—changed work/image to hierarchical that will support multiple hierarchies/relationships
- General Element Requirements—Add “Configure labels for community individualized fields”
- Authority files—Add “Authority records need to have unique identifier which the system would provide”
- Locations and dates—Add “type” attributes
- Notes—Add description
- Relationships—change the order of the list
- Scalability—Add size of fields and individual records
- System infrastructure—Add server requirements, open software, etc.
- Erin will take this document to the Imaging Group on Tuesday for their additions and feedback. This document will be made available once it is finalized.
getCitation and getTEIheader
June 17, 2005
Present: Erin Stalberg, Beth Picknally Camden, Ann Whiteside, Judy Thomas, Janis Kessler
Discussion of ContentDM
- Database structure is a concern.
- Speed of getting things available was discussed.
- Issue of authorities was discussed.
- Does not seem to have capability of relationships
- Erin will send inquiry about this product to Autocat; Ann will send to museum list and Judy will send to video list.
- Erin and Ann will check on ContentDM at ALA.
SnapDragon
- SnapDragon comments have been sent to Smith and they have agreed with those comments.
Annual Report
June 23, 2005
Present: Erin Stalberg, Ann Whiteside, Thorny Staples, Judy Thomas, Janis Kessler
CCT funcional requirements
- Review of CCT functional requirements continued. Ann will send to Smith to review and see what they might be considering for the next Snapdragon version.
NINES
- Permissions/policy discussion
- Policy–unrestricted for 10 years
<policy>
<access>restricted</access>
<time>
<date type=”begin”/>
<date type=”end”/>
</time>
</policy>
<policy>
<access>public</access>
<time>
<date type=”begin”/>
<date type=”end”/>
</time>
</policy>
- Within Access need to be able to accomodate more complexity (i.e. restriction to classroom or particular class).
- Brainstorming session on policies is needed.
June 30, 2005
Present: Erin Stalberg, Beth Picknally Camden, Ann Whiteside, Thorny Staples, Judy Thomas, Janis Kessler
ContentDM
- Review of CCT funcional requirements for ContentDM continued. Erin and Ann reported on the meeting with Greg Zick, president for ContentDM.
- New release is coming out soon.
- Primarily used for coherent collections.
- Hierarchical relationships available at delivery end, but no hierarchy in the records.
- Something can't be child of 2 parents which would be problematic.
- Licensing model is problematic–pay by the number of images.
- Retrospective conversion of the world rather than production.
- Elaborate authoring tool.
- Configurable (i.e. adding fields, renaming fields)
- Workflow tools available to set up template.
- Controlled vocabulary list is a text file with no references and without the capability to change records when the list is updated.
- OCLC is doing marketing and training.
- Has spell check.
- OCR and JPEG 2000–integrated.
- Can merge production and test server.
getCitation
- Thorny is working on the getCitation stylesheets.
July 14, 2005
Present: Erin Stalberg, Ann Whiteside, Thorny Staples, Janis Kessler
Review of CCT functional requirements for PICtor .
- Access database.
- Cleaner looking than IRIS.
- Moving from Access to MySQL.
- Mac use questionable.
getCitation
- Put versioning in. Add current version?
- Citation needs to be place, publisher, date.
- Need to check MLA and versioning.
- Continue discussion next time.
July 28, 2005
Present: Erin Stalberg, Beth Picknally Camden, Ann Whiteside, Bess Sadler, Thorny Staples, Janis Kessler, Judy Thomas
Introductions for Bess
- Bess was introduced to the group.
- Ann gave an overview of the Imaging Group.
- Erin gave an overview of MSG.
CCT functional requirements
- Review of CCT functional requirements for Snapdragon continued.
- Further discussion points for Smith
- General functionality, authority files, locations, dates, relationships
- Recommendation of MSG for CCT
- Options
- Stay with IRIS (with consortium)
- Take IRIS and make what we want
- Look for other tools—ContentDM, PiCtor, Snapdragon
- Build our own system from scratch
- Recommendation is to continue conversations with Smith regarding Snapdragon
- Demo Snapdragon for Imaging Group and ask for feedback on our recommendation.
August 4 , 2005
Present: Erin Stalberg, Beth Picknally Camden, Bess Sadler, Thorny Staples, Janis Kessler
CCT
- Erin will prepare one page summaries for Snapdragon, IRIS, PICtor, CONTENTdm and a summary of the MSG recommendation.
- The cost of moving to Snapdragon or staying with IRIS was also discussed.
getCitation
- Note explaining citation vs. bookmark
- This object may change from time to time
- Bookmark will always give you the current version. Version from [date] will give you the version of the object as it appeared on that date.
- Examples:
- Version from July 5, 2005 : http://
- Bookmark: http://
August 11, 2005
Present: Erin Stalberg, Bess Sadler, Thorny Staples, Janis Kessler, Beth Picknally Camden
CCT
- Review of recommendation document for the joint MSG/Imaging meeting
August 16, 2005
Present: Metadata Steering Group & the Imaging Group
CCT
- Demo of Snapdragon for Imaging Group
- The tool was largely well received
- Issues:
- FileMaker support
- Local development costs
- Staffing
August 25, 2005
Present: Erin Stalberg, Bess Sadler, Thorny Staples, Janis Kessler, Judy Thomas, Ann Whiteside
CCT
- Review of recommendation document for Erin & Ann to present to Martha's Managers on August 31st.
August 29, 2005
Present: Metadata Steering Group & the Imaging Group & interested other parties
CCT
- 2nd demo of Snapdragon for the Imaging Group & others who were interested in seeing the tool.
- Clemons 407 was full
September, 2005-December, 2005
The MSG got neglectful in these months in recording minutes. Many apologies. Many of these discussions surrounded the CCT. Ann & Erin met with Martha's Managers on August 31st and, from there, continued conversations with Smith regarding SnapDragon. Both institutions compiled their "top 10" lists of functional requirements --"low-hanging functional fruit" -- and proceeded with conversations about developing a next generation tool.
|
