Draft Phase 2 Priorities Table
Draft Phase Two Priorities
- Front-end changes – changes that affect the user experience
- Back-end changes – changes in implementation and design
- External changes – changes outside the scope of the Sakaibrary project
Priorities are 1 - 5, 1 being the highest and 5 being the lowest.
Difficulties are 1 - 5, 1 being the easiest and 5 being the hardest.
IU-P is IU's Priority Assignment; UM-P is U-Michigan's priority assignment.
Front-end Priorities
Item |
UI |
Svc. |
OSID |
Diff |
IU-P |
UM-P |
Notes |
Related JIRA Issues* |
---|---|---|---|---|---|---|---|---|
Full Text Only searching capabilities |
|
|
|
|
1 |
|
As per usability testing: have the capability to search for or retrieve only full text article citations from databases |
|
Style Guide options |
|
|
|
|
4 |
|
As per usability testing: Allow users to manipulate the citation information in such a way as to conform to established style guides such as MLA and APA. |
|
Terminology changes |
|
|
1 |
1 |
1 |
As per usability testing: Change field values to be more clear; explore other terms for Citation List ("Bibliography" or "Reading list") |
|
|
Add next/prev page arrows to bottom of search results |
|
|
1 |
1 |
1 |
As per usability testing and standard web practice |
||
Re-order citation editing/display windows |
|
|
1 |
1 |
1 |
Reorder fields so most used/salient field (e.g., volume, issue) are near the top with lesser-used fields (i.e. Rights) near the bottom |
|
|
Split citation editor into standard and advanced fields |
|
|
|
1 |
2 |
2 |
Move lesser-used fields into an initially-collapsed "Advanced properties" section to avoid confusion. |
|
Improve citation editing/display (in general) |
|
|
|
1 |
1 |
1 |
Fixes for non-editable fields, window-sizing, "Add Another" window-jumping, and required field checking is required. |
|
Make window navigation less confusing |
|
|
1 |
1 |
1 |
Make it more clear that citations are being added to the citation list when 'Add' is clicked; Clicking on the title of a citation should open a new window with its full details; Maintain scroll position on page during a refresh. Much of this is related to "Refactor UI", below. |
|
|
Add more explanatory text |
|
|
1 |
1 |
1 |
Add help text explaining the steps needed to create a citation list; Let users know how results are sorted; add descriptions to search targets so informed choices can be made about what to search; number search results so people know where they are; add sample searches so people can see the syntax available. |
|
|
Stabilize citation order |
|
|
|
3 |
1 |
1 |
Fix bug that reorders citation list whenever a citation is edited |
|
Allow user to cancel ongoing search |
3 |
2 |
2 |
Searches that take a long time or are started in error should be abortable without closing the window |
||||
Give user more information up front |
|
2 |
1 |
3 |
Provide an easier way to get at the abstract and other metadata before adding a citation to your list. |
|
||
Re-think Export UI |
|
|
3 |
3 |
3 |
Current export procedure is confusing to users; provide more information about the process and the resulting file format; remove checkboxes from standard display unless/until there are other multi-citation operations. |
||
Create Research Guide document type |
|
5 |
1 |
1 |
Prototype of research guides. Need ways to store the components, ways to embed citations/searches/targets, a serialization format, and a UI |
|
||
Add more citation types |
|
|
1 |
2 |
2 |
Add a few more citation types, along with some "good guesses" for determining type based on available metadata |
|
|
Import from RIS |
|
|
2 |
2 |
2 |
Import RIS files to an existing (or new?) citation list |
|
|
Google Scholar integration |
|
|
3 |
1 |
1 |
Need to figure out how to target a specific (authenticated) user's specific citation list |
|
|
Advanced (fielded) search |
2 |
1 |
1 |
Add title/author/whatever search |
||||
Choose specific search targets |
4 |
1 |
1 |
Need a UI, an internal way to store lists of targets, a way to pass possible targets out of the OSID, and a way to pass targets into the OSID as a search property. (How) Do we support pre-set categories/subcategories of databases by subject? |
|
|||
Refactor Search UI |
|
|
|
4 |
2 |
2 |
Replace popup windows with floating divs; may want to refactor the whole searching UI while we're at it (e.g., to include a running list of saved citations where the user can see it) |
|
Refactor Citation List Editor UI |
|
|
|
|
|
|
Take advantage of the coming changes in Sakai 2.4 (we have control over the entire frame) and change to a more modal Citation List Editing process. |
|
Error Reporting |
|
|
|
|
|
|
Define user-friendly messages for all possible error conditions. |
|
Export to Refworks |
|
|
|
3 |
3 |
3 |
Similar to creating an OpenURL (or, perhaps, an inline RIS); need to investigate |
|
Canned searches |
|
|
3 |
1 |
1 |
Need a UI and a datatype to hold search targets (and optional terms) |
|
|
Sakai as SFX Target |
|
|
5 |
5 |
Need to learn more about SFX; similar problem to having Sakai be a Google Scholar target |
|
||
Sort citation list |
|
|
4 |
1 |
2 |
Working within a citation list |
|
|
Search/list/sort citations across citation lists |
|
4 |
4 |
4 |
Depends on where we decide citations live |
|
||
Explore a browser plugin interface to the citation system |
|
|
4 |
4 |
4 |
Greasemonkey/plugin approach to searching and saving via Sakai, rewriting links to use the URL format designed for the Google Scholar project to target Sakai from native interfaces |
|
|
Create a Citation List icon |
|
|
|
|
|
|
Create an icon that will represent the Citation List Resource type in Sakai. |
Back-end Priorities
Priorities are 1 - 5, 1 being the highest and 5 being the lowest.
Difficulties are 1 - 5, 1 being the easiest and 5 being the hardest.
Category |
Goal |
Priority |
Difficulty |
Notes |
---|---|---|---|---|
Advanced Search |
Determine which advanced search facilities the metasearch engine can support |
1 |
3 |
see Web2Bridge Advanced Search, /wiki/spaces/SLIB/pages/2921169066; this involves a fair amount of experimentation to determine what is actually supported by the metasearch engines due to relatively weak documentation. |
|
Define a "context set" (namespace) for using CQL |
|
3 |
CQL is a flexible query language and we can choose/create a "context set" for it - for example, we could use the Dublin Core CQL context set. CQL context sets could be defined for individual repository implementations - for example, the Web2Bridge CQL context set would completely describe the search context and functionality for Web2Bridge. The X-Server would have a slightly different CQL context set. |
|
Define a way to get search-type information from the OSID |
|
2 |
A certain metasearch engine may offer title, author, year whereas another one may offer title, author, year and subject - the consumer should know which search types are supported. Once a CQL context set is defined, it could be passed to the consumer. |
|
Define a way to get the logical search relations (and, or, not, etc.) from the OSID |
|
2 |
Same reasoning as above. |
|
|
|
|
|
Getting Search Target Hierarchies from the OSID |
Determine what can be provided by the metasearch engines |
|
1 |
see Web2Bridge Advanced Search, /wiki/spaces/SLIB/pages/2921169066 |
|
Define how categories/subcategories and hierarchies in general can be defined/configured |
|
2 |
This can be done in a number of ways, controlled by the Consumer. The Consumer would configure where the OSID should look for search targets - it could be the metasearch engine, a URL or a local file. |
|
Define how categories/subcategories and hierarchies in general can be represented in the OSID and passed to the Consumer |
|
2 |
A new Asset Type needs to be created to represent a search source. |
|
|
|
|
|
Speed vs. Quality |
Improve the quality of search results |
|
5 |
Doing this will adversely affect the time it takes to get those results. |
|
Improve the speed in which we get search results |
|
5 |
Doing this will adversely affect the quality of the results we get. |
|
|
|
|
In Phase 1, we're trading quality for speed - we have a very fast search, but unpredictable search results. Tweaking the balance between speed and quality combined with novel features such as advanced search and hierarchical search target selection may require a novel way of fetching records within the OSIDs. This is a discussion that would involve everyone (usability, design, technical, etc.). |
|
|
|
|
|
Search Result Numbers |
Determine a way to effectively deal with search status information (i.e. total number of results found) |
|
3 |
This may be more of a UI issue than a technical one. |
|
Develop a more robust way to determine duplicate records OR how to deal with "phantom records" at the end of all search results (i.e. Viewing 1630 - 1641 of 1647) |
|
5 |
Researching options for duplicate checking would be a technical task whereas dealing with messages to the user about Search Result Number conflicts is more a UI discussion. Currently, all searches of multiple search sources with the X-Server (i.e. General Interest Set) end in the phantom records issue if they yield results over a page or two. |
|
|
|
|
|
Configuration |
Move all configuration outside of the OSID jar |
|
2 |
Need to define what "configuration" means (what things need to be configured) and need to define where "outside of the OSID jar" is and how that configuration gets into the OSID implementation. |
|
|
|
|
|
Authorization/Authentication |
Determine how to handle authorization and authentication of metasearch engine users through the OSID |
|
5 |
We need to be able to pass user authentication/authorization tokens from Sakai out to the underlying metasearch engine. Stopping this at the article-download level is insufficient because for many databases, the metadata is the product (e.g., ISI). At the same time, we don't want to deal with collecting and protecting passwords if we can help it. |
|
|
|
|
|
Embedded citations |
Allow any document type to call our code with a unique citation ID and get back data sufficient to render that citation |
|
|
Do we go our own way? Wait for RSF to be ready? Do a proof of concept implementation in a single document type (e.g., perhaps try to work with the Syllabus people? |
|
Provide a well-defined API for getting citation data |
|
|
Follow RSF conventions, but also explore exposing this data as a web service |
|
|
|
|
|
Citation scope and permissions |
Provide clear ownership/scoping rules for citation objects |
|
|
To what do citations "belong?" The user? The course? The user in a course, or across the whole Sakai instance? This encompasses both permissions for individual citations (Who can read it? Who can edit it? Who can delete it?) and issues surrounding both search and targeting (e.g., how do we define the target of an RSF rendering request and determine if it's allowed?) |
|
|
|
|
|
More use of SFX |
Use the SFX X-Server API to (try to) check on fulltext availability |
|
3 |
Can be expensive (in time and resources) and very error prone to try to determine fulltext availability; could fall back on a less-certain-sounding "I feel lucky" button ala Google |
|
Abstract SFX API |
|
3 |
Need to discuss how tightly to tie ourselves to SFX. All SFX calls would have to be optional and probably should be abstracted into a library that could (in theory) target other OpenURL resolvers. |
|
|
|
|
|
Library Server |
Implement a remote metasearching server under library control |
|
4 |
Need to determine if we should proceed with this for Phase 2 for sure. If so, we need to determine communication protocols as well as "new" OSID connections. |
|
|
|
|
|
External changes proposed
These changes are outside the scope of the Sakaibrary project, requiring changes to the Resources Tool, integration with other Sakai services, or perhaps decisions made on a Sakai-wide scale.
- Changes to "Title" and "Description" field names. These are controlled by the resources tool; suggested changes are "Title of the Citation List" and "Description (text will appear above citations)", obviously controlled on a per-resource-type basis. The confusion prompting this change may disappear as the Resources tool moves to a more wizard-like approach.
- Templating syntax to embed citations The entire Sakai framework is moving toward using RSF as the mechanism to allow documents to give up rendering control to another service, e.g., to embed a citation. Any short-term solution would require working with the creators of a specific tool or tools while we wait for a Sakai-wide solution to evolve.
- Provide hinting/help for the document types Help or descriptions of the different document types that can be created in Resources would allow people to make better decisions about what to create. A Resources Tool issue.
- Change the Resources 'Add' button to reference specific doctype ('Add this Citation List') Another Resource-tool level issue.
- Collapse/explain the properties better. The properties section of the resource add screen confuses users. Collapse these into an "advanced" section.
Notes
*issues are as of 18 Dec 2006