Search/Find

Owned by Kat and Steve. (We need to look at requirements for other systems – FRBRized systems, faceted browse systems – to make sure we're not missing anything)

Keyword searching

yes

P1

level 1; just indexes whatever is there

User enters a term into an input box or boxes; system responds with Aquifer search results that match that term, regardless of where that term appears in the record.

Fielded searching

yes

P2

level 1: creator, title, rights; level 2: subject, language, date, typeOfResource; level 3: genre, form, non-topical subjects; level 4: abstract

User enters a term into an input box and specifies which field to search that term (e.g., "title", "author"); system responds with Aquifer search results that match that term only in the field specified. (Need to itemize all fields we want to be searchable.)

Search limits

yes

P2

level 1: rights; level 2: language, date, typeOfResource; level 3; genre, form

User enters a term and specifies to limit that search to a particular field value (e.g., just books, just records from Emory); the system responds as specified. (Need to itemize all fields we want to use as search limits.)

Search by clustered values

no, only on DLF portal

P4

level 2 (implies some kind of controlled vocabulary as output from automated clustering process)

Users choose field values from a controlled vocabulary (usually hierarchical) to limit their searches.

Search by selecting hyperlinked field values

yes

P2

n/a

User views search results and sees that some fields are hyperlinked. User selects a hyperlinked field value and system responds with Aquifer search results constrained to just that field value (initial search is not included in the constraint).

Expose facet values for result sets

no

P2

various; see fielded searching/search limits

For any arbitrary result set (including a query with no constraints), the system displays one or more facets (e.g., language), facet values (e.g., English, Spanish, French) that apply to the query. Facet values should also display the number of records in which that facet value appears in the result set, recognizing that many fields in a record may have multiple values. For example, if a query for "toast" in the keyword index results in two records and the first record has values of "English" and "Spanish" in the language elements and the second record has a value of "English" in the language element, the language facet should have two facet values: English (2 records) and French (1 record). Null facet values (Spanish) should not display. (If browse is supported, need to itemize browsable facets and their facet values.)

http://flamenco.berkeley.edu/demos.html http://www.lib.ncsu.edu/catalog/ http://worldcat.org/ http://blacklight.betech.virginia.edu/

Apply constraints to a query by selecting facet values.

no

P3

various; see fielded searching/search limits

When user selects a facet value from the facet display on the results screen, the system appends this query to the existing query and displays results that match the constraint. Using the example immediately above, when the user selects the "French" facet value, the query becomes: (keyword=toast) AND (language=French).

Indicate active constraints

no

P2

n/a

For any result page, the system should display to the user any constraints that have been applied to the query (e.g., facet value limit, search limit, query term, Boolean operators used). After selecting a facet value, that facet value should not be hyperlinked and should indicated "you are here".

Remove facet constraints

no

P3

n/a

From the display of the active constraints, user should be able to remove any arbitrary constraint and generate new results. For (keyword=toast) AND (language=French), if a user removes the (keyword=toast) constraint, the system should display results for the (language=French) query.

"More..." facet values

no

P2

n/a

There should be a system-defined, configurable threshold for how many facets values to display within a facet as a default. User should be able to choose to display all facet values for a given facet.

Hierarchical facets

no

P4

TBD: date? geography? subject?

Some facet values may be hierarchical (e.g., Michigan --> Ann Arbor). Selecting "Michigan" should expose a page that lists all cities in Michigan, for example, and then the user can select "Ann Arbor". (Spec further if hierarchical facets are a go).

Sort order facet values

no

P4

n/a

The facet values should be sorted by number or records (descending) as a default. The system should also be able to display by number of records (ascending) and display alphabetically (ascending and descending)

Metadata analysis interface

no

P4

n/a

Aquifer metadata librarians (e.g., Kat, Jenn) should be able to see the distribution of facet values, find outlier facet values, null facet values, etc..

See attachment metadata_analysis_interface_requirements.doc

Roy's tool Rob Casson at MU-Ohio video of UNT's tool

Refine query

yes

P2

n/a

User is able to view the original search query and change it by adding or removing query values.

For a particular record, find "more records like this"

no

P3

n/a

User identifies a record as input for a subsequent search; system responds with Aquifer search results that are similar ("query-by-example").

For a particular work, find "more works like this"

no

P4

n/a

While viewing search results that are grouped by FRBR-ized work sets, user identifies a work and indicates s/he wants to find similar works; the system responds with search results grouped by work sets, of works that are most similar to the one the user indicated.

For a particular work, find all versions of that work

no

P4

n/a

While viewing search results that are grouped by FRBR-ized work sets, users identifies a work and indicates s/he wants to find all expressions/manifestations of that work; the system responds with the master work record and all records in that work set.

View results

yes

P1

n/a

System shows user brief and/or expanded metadata (including thumbnails) for Aquifer records that match a query. Brief display can be thought of as the "Google model".

View faceted results

no, only on DLF portal

P4

various; see fielded search/search limits

System shows user one or more facets and facet values (with counts) for all the records (works?) that match the query.

Sort search results

yes

P2

n/a

User is able to sort record results by relevance (which algorithm(s)?), author/title, date, provider, etc.

Group all versions of a work in a single search result

no

P4

n/a

System groups all versions of work on a results page (assumes FRBRish work sets). A specific task that takes advantage of this functionality is to "Find related works".

Results per page

yes

P2

n/a

System returns a certain number of results per page, and the option to decrease, increase or show all results.

Pagination

yes

P1

n/a

System provides the ability to proceed to the first, last, next, previous, and any arbitrary results page.

View search history

no

P3

n/a

System allows users to return to a set of results from previous search(es), to view and/or to refine.

Saved searches

no

P4

n/a

User can choose which queries in their search history they want to save across sessions (assumes user accounts). (Also see Obtain for the ability to download saved searches/records and/or subscribe to saved searches via email/RSS.)

Search in multiple languages

yes, can enter CJK, etc.

P4

level 2

(What is this: entering non-English text as a search term? providing a UI that is non-English? translating records from English to non-English in the display?) (Additionally: searching using language codes or plaintext? mapping from codes to text?)

Identify

Owned by Katherine (help is welcome)

Determine relevance

yes

P2

n/a

User is able to clearly see relevance score. User may also click on that score to see a description of how "relevance" is ranked by the system (a verbal description of the ranking algorithm's components).

Aquifer/Institution branding

yes

P2

n/a

System includes prominent branding (icons?) on every record. That branding identifies the record's institutional origin.

Display foreign language (from metadata)

yes

P3

n/a

System can handle and display properly all basic foreign languages.

Translate metadata into foreign languages

no

P4

n/a

System can translate metadata into other languages, especially translating English to Spanish and vice-versa

Obtain

Started by JV - any help here welcome!

Implementation of Asset Actions would solve many of the requirements in this section.

Link to content

yes

P1

level 1

Important to have a persistent link directly the item (MD + digital object), not to collection level.

Select object

not sure

P1

n/a

SET: What does this mean? ability to cognitively select it in your mind or is this a UI widget to add it to your basket?

Metadata

yes, descriptive metadata

P1

level 1? various?

Need both descriptive MD about intellectual content and tech MD about the digital object.

SET: need to itemize fields at a future date.

Obtain full or partial citation

yes, obtain full citation

level 1? various?

SET: need to itemize fields at a future date.

Export metadata

yes, but only fielded BibClass

P1

n/a

Export in standard bibliographic formats: Zotero RDF, BibTeX, MODS, Endnote, RIS, Unqualified Dublic Core RDF, Wikipedia Citation Templates (above, taken from Zotero list).

SET: is each of these formats P1? My guesses: are that RIS might be P1 but the rest are P3

Make metadata openURL friendly

sorta, there is a hack

TBD; see question

SET: what does this mean: ability to generate a valid OpenURL (0.1 or 1.0?) or are there certain fields that are required?

Show viewers supported

no

P2

Not sure which MODS field this is!

Inform user what viewer needed and, ideally, point user to viewer download site.

Show rights/permissions

yes, if is in metadata

P1

level 1

Self explanatory?

SET: is rightsMD required on ingest; if not required, what do we do for records that lack: do we insert anything?

Show size/extent

yes, if is in metadata

P1

level 3?

So user can determine how long download of digital object will take. Ideally would provide info about download speed based on bandwidth.

SET: P3?

Manage/Use

Owned by Jennifer and Deb

Bookmark/bookbag

yes

P1

n/a

Mark items for later use such as exporting metadata or downloading files. Group files into 'sets' for either personal or group use, similar to collection feature in ArtStor

SET: ability to relabel sets, annotate sets, or export set-level metadata?

Manipulate object

no, because only pointing to the object

P2

n/a

User can change the object in some way. Manipulation may be dependent on file types for example, zooming in on a detail and cropping the picture to save that detail would be available for image files.

To what extent is manipulation dependent on external software? What capability comes with the objects themselves and what is stored on a server. --Is the ability to store objects post manipulation important?. If so, where?

Remix/Reuse

no

P4

n/a

Take content of an object and reorder so different meaning may be inferred/derived. For example, take a set of images, select a subset and change the order of that set and submit the new set in the new order or take the scenes in a video and reorganize them and resubmit.

SET: if this one stays in, need to break it down further for each format.

Annotate object

no

P2

n/a

Add text to item. This could be tags or lengthy text. The annotation could be kept in personal space on server linked to an account and accessible to that account, or available to a group or be in public space. Some tools annotate content within a system and other tools allow users to annotate content from anywhere.

need to think about whether we want to provide annotation tool or annotate through external systems like flikr or del.icio.us. If annotation could be placed in a wrapper that can be downloaded and transported with the file keeping everything as an object, how would it be edited?

Examples of annotation tools: Delicious, Library Thing, ArtStor, Aluka.

Translate object

no

P4

n/a

For text, providing it in different languages. The assumption is that this happens on the fly through translation software. Translation could also be changing file types for example changing file from MSWord into PDF or tiff into jpeg

Would we want to store translations to facilitate multilingual search?

Assessment

Portal reporting usage statistics, sources, partner institutions from search partners

Transaction logs

"

Queries

"

Clickthroughs

"

Search Terms

Reporting from search partners (Google, Sakai, etc.) usage logs

Aquifer architecture

Clustering by Subject

Aquifer – Subscribe to updates

Aquifer – Find more in Aquifer

Export normalized controlled vocabularies to search engines

Aquifer – Translate into another language

Assumptions

Support SRU Level one

UTF 8 compliant

xml 1.1 compliant

must use persistent URL using OAI id field

support real time query of aquifer server