Project Listing and Clients Contact Information

• Client Listing and Contact Information:

* =   may require travel to Health Sciences Campus to meet with clients; there is an hourly shuttle between the University Park campus and Health Sciences.

The original station data for the voyages of research vessels sponsored by the University of Southern California are housed in the Hancock Library of Biology & Oceanography. The Velero IV data were collected between 1948 and 1983. Historically, researchers who used these data were working with specimens collected during the voyages, but the data are potentially much richer for a broader research base if they could be made searchable.  For the paper record, the only access is through the station number, which allows the related data to be retrieved.  This level of detail is not available from any other source than the station data sheets.  Subsets of the data have appeared in various publications of the Allan Hancock Foundation, but the information in the total data set, or simply a single specific station entry, is only available by direct request to the Hancock Library.  A table of all data elements is below.

      Cruise Number                                    Recorders
      Date of Cruise (MM/DD/YY)              Time Zone
      Station Number                                    General Locality

      Gear Used                                           Direction of Travel
      Starting Latitude, Longitude, Time and Depth in fathoms
      Ending Latitude, Longitude, Time and Depth in fathoms
      Wind Direction and Speed                    Remarks

Approximately 40% of the total data set (8mb) has been entered into a Microsoft Accessâ database and is housed on a zip disk. The original requesters want the data to be published in print form; however, the question of making these data available as a searchable file on the web has been suggested.  The searchability of the file would allow the historical information to be used to compare a like area with present conditions.  In some cases these are baseline data that have not previously been accessible in this way.

This project will investigate whether there is an appropriate way to transfer the data to a web searchable format.  These concerns must be considered in the project:

• What means exist to transfer the data from Accessâ to create a web database?
• How many searching functions can be combined?
• Geographic location must be searchable by either an exact location or a range of longitude/latitude parameters.
• Can geographic location be searched by named locality, e.g. Santa Catalina Island or within x miles north, south, etc., from

Santa Catalina Island?
• Other questions may also arise through discussion with the primary users, so the database access must be flexible enough to

allow a variety of combinations to be made.
• Results must be able to be downloaded and/or printed.
• The entire file must be safe from alteration, although editing with authority should be possible.

2.      Web Site Portal and Collaborative Filtering:

The Norris Medical Library’s Web site (http://www.usc.edu/hsc/nml/) includes links to a vast collection of electronic information resources, as well as information about the library and its services. As with other Web sites the proliferation of information on the Web makes it difficult for users to find the specific information needed. We are interested in enhancing our Web site to provide tools that solve this problem. We would like to provide various methods for helping users find information of interest including as many of the following as possible:

• Personalized portals (i.e. “My Library”)
• Collaborative filtering (i.e., amazon.com providing other titles of possible interest based on books purchased)
• Push technology based on Web utilization
• User profile

Linking to existing systems so that services requested through our Web page would automatically fill in the necessary user information so that the person wouldn’t have to re-input it each time a service was requested.  Also, if linked to existing University systems including financial system, fee services could be automatically debited to P.O.’s or deposit accounts etc.  The user profile could also be used to help generate the push technology applications.

• Live assistance for simultaneous Web site browsing and/or instruction (e.g., Hipbone software)

3.   Searchable Image Database for Pathology Slides:

As part of the Medical School curriculum, students study the pathological basis of disease and need access to a large variety of pathology slides and images.  Many of these have been digitized and made available through the Medical School’s Web page.  In addition, digitized pathology images are available through the Web sites from other medical schools.  We would like to have the capability of searching through a selected set of Web pages for a particular type of pathologic image (i.e. lung tissue infected with cytomegalovirus) and returning any related images.  Because some of the images on the USC site are copyrighted, it would be necessary to restrict access to USC students and faculty.  Also, when an image is located on another Web site, we would like the search engine to not only display the specific image and accompanying caption, but also reference the originating Web site so that that institution could be credited as the source of the image.

This project could be further enhanced by utilizing some of the tools developed and made available for free from the National Library of Medicine.  This suite of tools is referred to as the Unified Medical Language System.  Its purpose is to provide developers with the capability of retrieving information from a variety of sources which use different terminology to describe the same thing.  Further information on the Unified Medical Language System can be obtained from the National Library of Medicine Web site at: http://www.nlm.nih.gov/pubs/factsheets/umls.html
 

     Background

The Health Sciences Libraries (http://www.usc.edu/hsc/nml/) subscribe to a growing number of journals  (now about 750) in electronic form, available to USC users on the Web. These journals are housed at a variety of Internet sites, including both individual publisher sites and sites that provide access to journals from multiple publishers (e.g., Ovid Technologies).  These sites typically provide two levels of access, one for the general public that allows viewing of table-of-contents and sometimes abstracts, and one for subscribers only that allows for viewing of full-text articles (in HTML, PDF or both).  The library goes through a process of registering the USC IP domains with the publishers for those titles that it has rights to, either by virtue of having an existing print subscription, or because of a separate electronic subscription fee.  The publishers then allow or disallow access to the full-text displays based on the IP address of the viewing station.

The problem is that for a variety of reasons, access to the full text of articles at any given site can suddenly be discontinued, either because the publisher thinks the subscription hasn't been renewed, or a free trial period ends, or there's just something going wrong with the authentication software at the publisher's end.  The usual result is that when the user clicks on the full-text link, rather than seeing the article, he sees a screen asking him to enter his personal user ID and password -- a sign that the IP-based authentication has failed.  Unfortunately, unless the library is notified by the user, we are unaware that the IP access method has broken down.

It would therefore be immensely helpful if there were a way to regularly check each of the journal sites to be sure that links to the full text of the articles are actually working.  Obviously, a simple Web-based link checker won't work, because the user can still get into the journal site, he just can't get to the next step of viewing a full-text article.  Nor is there any simple script that could be written to automatically check full-text links for all journal titles, since each journal site requires different navigational strategies to get to the point where a full-text display becomes an option (e.g., you typically have to first choose an issue and then click on the full-text link that is part of each entry in the table-of-contents for that issue).  It might be that a different checking script would need to be developed for each journal -- although, in fact, many of the journals do have identical site architecture (because they are from the same publisher or aggregator), so that perhaps 15 or so different scripts might cover 95% of the titles.  Whatever solution is developed would need to be able to accommodate new titles as they are added, since we anticipate doubling the number of      electronic journals over the next three to five years.

Internal Revenue Service code and California law require employers to keep track of time worked for overtime-eligible employees. Information Services Division (ISD) would like to replace its current timecard and timesheet (paper) system with an electronic, web-based system to simplify data collection, to more accurately record hours worked for all its employees, and to provide personnel management tools for supervisors and directors. Several companies offer systems at varying levels of cost and sophistication (e.g., http://www.vitrix.com/ or http://www.kronos.com/ ).  These systems need to be compared in terms of cost, performance, and implementation.  It should be noted that at present ISD only tracks hours for student employees, and that the full-time staff may resist reporting their hours, so identifying an elegant, affordable solution will aid in both the acceptance of this new procedure, as well as ISD’s overall efficiency. Analysis and results will be presented to management along with a recommended solution.

The Photocopied Table-of-Contents (TOC) Faculty Current Awareness Service is a DBase III program. Faculty are sent a printed list of about 110 journal titles from which they select those journal titles they would like to receive on the service. There are 100 faculty on the TOC Service. The library staff photocopies the table-of-contents of the latest issue received in the library and sends it to the relevant faculty via campus mail. Faculty review the photocopied table-of-contents, mark or check off the articles they would like photocopied and return the marked copies to the library. The library staff then photocopies the complete article or requested information for the faculty member and mails it to his/her office. Annually the faculty on the service are requested to update their list and either add or delete titles. Also, when a new journal title is received in the library a sample of the table-of-contents is sent to the entire faculty on the TOC Service. Faculty then has the option of adding this new title to their profiles. The library staff then updates the DBase program, adding new titles, removing/deleting dead titles, and revising faculty profiles. Staff also generate reports as needed. The library does have a scanner.

This project is currently operational, primarily as a manual system. We would like to create a new database and migrate it over as a web-based service. The proposed web-based program should have the following capabilities:

•        ability for individual faculty to access (by password?) and revise and update his/her TOC profile
•        ability to click on title (link) and
•          be sent to title's homepage
•          view a sample of the table-of-contents
•          view instructions for authors
•         ability for library staff to send TOC faculty information about a new journal title
•         ability for sys admin to revise or update all faculty TOC profiles
•         ability for system admin to view all faculty profiles
•         ability for system admin to generate/receive reports

7.      Wilson Dental Library New Book List:

Monthly, a text file is distributed to members on the SCDental listserve listing selected new books, journals, and media received by the Wilson Dental Library. Using HELIX, the online catalog for the Health Sciences Libraries (http://hsl.hsc.usc.edu/webpac13/wgbroker.exe?new+-access+top.horizon), new books are identified and the bibliographic record downloaded as a text file. Using Word the librarian edits the file, for example deleting extraneous information. The file is saved as a Word file then as a text file and FTP'd to email. Using Pine the file is then distributed to subscribers on the SCDental listserve.

1. Propose to convert this file to an HTML format with ability to click on the title and to view the table-of-contents of each item on the list. The library has a scanner.
2. Project is currently operational as a text file only.

USC subscribes to many different databases that index and abstract popular and scholarly articles and make them available over the World Wide Web. In some cases these databases include the full-text of the articles as well. A major problem faced by students and librarians is identifying where a particular periodical title is indexed, the dates of coverage and whether it is available in full-text. Title lists are available from the vendors but are large files and can be difficult to use on the fly. The purpose of this project is to retrieve and consolidate this information into a searchable database that can be accessed via the Web.  Two examples of approaches to this problem are available on the web. The first is the jake project at Yale University:
           http://jake.med.yale.edu
This is a cooperative, open source based project that collects data from as many databases as possible and presents it in a searchable database. This is a valuable tool but can be confusing to users because it includes information on many databases that USC does have access to.

A second example is a database produced by Indiana University:
           http://www.indiana.edu/~libej/
This is a database of titles contained only in databases that IU has access to which is more functional for users.

These examples show good interfaces but do not show the difficulty of collecting and collating the information that is needed. The vendors’ title lists are of varying quality and differing formats and can change without notice.

Currently, users of USC’s electronic resources spend a great deal of time trying to find sources in fulltext rather than focusing on their research topics. Having a comprehensive, easily searched collection of fulltext titles would fill a serious gap in the tools available to researchers and students.

Create applications for auto-generating FAQ files for troubleshooting  computing problems. Integrate these with existing web content to create intelligent web forms that help solve computing issues.

Take high-quality video and manipulate to create web-streaming version. Have the delivered video "click-enabled" such that viewer can find a frame in the video stream and obtain a high-resolution still image from the server.

Devise a way to synchronize digital whiteboard output with audio files (i.e., from a lecture or meeting). Enable these files to be viewed and distributed on the web. Utilize speech recognition to create a searchable index from the audio track enabling users to find specific points in the audio/whiteboard files.

The Web has created myriad opportunities for people to disseminate digital media. While the creation of audio and video content is now easy, moving this material to the Web is often non-trivial. This project seeks to create a media "kiosk" that enables users to easily take both analog and digital media from raw material to web-delivered product. These stations will have both hardware and software solutions that create a simple approach to make rich media content available on the web.

At present USC has over 10,000 images with descriptive metadata available via the web in 5 Integrated Digital Archive (IDA) databases (http://library.usc.edu/uhtbin/cgisirsi/gtH7u812Pq/0/1/37).  Current capability only allows a user to view one image from one search at a time.  Researchers want to be able to compare images from different searches and save the results for further study at a later time.  This project seeks to create a “composer” button to add to IDA metadata record/image display screen to enable user to save the URL’s for images of interest from different searches and different IDA databases into a dynamically created webpage; user could then view images in any order and add notes to images.

This project requires the development of a client/user interface for its collection of digital images and text metadata.  The 1500 digital images the library possesses currently are stored on a university server and the descriptive metadata is stored in File Maker Pro on computers housed in the Slide Library.

The client would like to make the images and associated text data available via the Internet to faculty and students who have access through the USC domain name.  The purpose of the access is to provide the capability to prepare and save classroom presentations using this material.

To be usable in this context, the images must:
     ·        be available in thumbnail and 72 d.p.i.
     ·        load quickly, within 15 seconds or less
     ·        must be crisp and detailed at 72 d.p.i.
     ·        capable of manipulation to order the images for classroom presentation
     ·        capable of display in sets of  2 images
     ·        have a zooming capacity to move around the image and display details

The IDA-LA system is a complete digital archive of primary materials about the Los Angeles Area. The objects (items) in this archive have been scanned and cataloged using specially designed Dublin Core metadata records. The power of this system can be traced directly to these metadata records since they are used to support searches based on space, time, keyword, collection, and format. Space here refers to the geographic footprint represented by the item (e.g. the area captured with an aerial photograph). Time refers to the temporal footprint (e.g. when the photograph was shot). The keyword(s) are similar to the words and phrases used to access "regular" library materials. The format refers to the method used to represent individual items. Five formats – audiovisual, geographic, image, tabular, and textual data – are currently supported. Some items are also part of collections (e.g. photographs taken by one individual for a specific purpose) and can be retrieved this way as well.

 This system as it is envisaged and/or built has three broad components:

-    A database containing digital objects and their accompanying metadata records.

-      Ingest tools to convert paper-based and other archival material into a digital format and cataloging tools to build the appropriate Dublin Core metadata records.

-     Search and retrieval tools so that library patrons can search for and retrieve digital assets using some combination of space,  time, keyword(s), format, and collection name.

The current archive contains approximately 10,000 objects. These assets are managed by Dennis Smith (ISD-Information Technical Services) and are stored in a BRS database (http://library.usc.edu/uhtbin/cgisirsi/gtH7u812Pq/0/1/37 ). A new search and retrieval prototype is currently under development as part of a collaborative project involving the USC GIS Research Lab and ESRI based in Redlands, California. A variety of ad hoc software tools and procedures have been used to ingest these materials to date. The majority of the cataloging accomplished to date was performed by a single individual (Wayne Shoaf, ISD-Digital Resources Cataloging), although we envisage a fully distributed and web enabled system with hundreds of contributors adding objects and the accompanying metadata records (see related project description below).

This particular project has two primary goals: (1) the evaluation and specification of key database needs and functions, and (2) the preparation of a proposal describing a database design using some combination of ORACLE and/or ArcSDE. The proposed system must support multiple formats (i.e. numerous audiovisual, geographic, image, tabular, and textual data formats), support geographic data types (important for the dynamic mapping components that will be incorporated in both the ingest & cataloging and search & retrieval tool sets), and spatial indexing to facilitate rapid search and retrieval. Various compression tools (DejaVu, MrSid, etc.) may be needed as well as a part of this solution.
 

The IDA-LA system is a complete digital archive of primary materials about the Los Angeles Area. The objects (items) in this archive have been scanned and cataloged using specially designed Dublin Core metadata records. The power of this system can be traced directly to these metadata records since they are used to support searches based on space, time, keyword, collection, and format. Space here refers to the geographic footprint represented by the item (e.g. the area captured with an aerial photograph). Time refers to the temporal footprint (e.g. when the photograph was shot). The keyword(s) are similar to the words and phrases used to access "regular" library materials. The format refers to the method used to represent individual items. Five formats – audiovisual, geographic, image, tabular, and textual data – are currently supported. Some items are also part of collections (e.g.  photographs taken by one individual for a specific purpose) and can be retrieved this way as well.

This system as it is envisaged and/or built has three broad components:

-   A database containing digital objects and their accompanying metadata records.

-   Ingest tools to convert paper-based and other archival material into a digital format and cataloging tools to build the appropriate Dublin Core metadata records.

-   Search and retrieval tools so that library patrons can search for and retrieve digital assets using some combination of space, time, keyword(s), format, and collection name.

The current archive contains approximately 10,000 objects. These assets are managed by Dennis Smith (ISD) and are stored in a BRS database (http://library.usc.edu/uhtbin/cgisirsi/gtH7u812Pq/0/1/37) . A new search and retrieval prototype is currently under development as part of a collaborative project involving the USC GIS Research Lab and ESRI based in Redlands, California. A variety of ad hoc software tools and procedures have been used to ingest these materials to date. The majority of the cataloging accomplished to date was performed by a single individual (Wayne Shoaf, ISD-DIAA), although we envisage a fully distributed and web enabled system with hundreds of contributors adding objects and the accompanying metadata records.

The primary goal of this project therefore involves the design of a new tool that replaces the existing PERL scripts and tables with a more robust and versatile set of ingest and cataloging services. These tools should use JAVA scripts that talk directly with ORACLE (the preferred relational database choice) and use ArcIMS to support the use of dynamic maps, gazetteers, and geocoding for specifying geographic footprints in the metadata records.

Currently USC utilizes a UNIX mail platform with a multitude of (front end) applications (Exchange, Express, Pine, Eudora, etc) being deployed by various associates to access their mail accounts.

While the existing campus mail system is quite robust as a delivery mechanism, it does lack some components of other existing mail systems suchas; propagation, improved directory services, web based remote access, etc.In addition to some missing functionality, we also see significant change in our desire to be able to communicate anywhere, anytime, with more and more web based applications that transcend the boundary of the MAC and Windows based back ends components.

To meet the changing environment the University would now like to explore and understand the options available for enhancing a campus wide web based mail system with appropriate security.

Our desire is to explore this from both a technical and economic value perspective.

Network utilization is the process by which ISD monitors the health of USC’s Network and Unix systems. The premise is that if we can track and analyze system load and failures on a daily, weekly, and yearly basis, this data can forewarn us of potential trouble, saving an enormous amount of support time and helping to stabilize the Network and Unix platforms. 

We currently use an in-house written tool called sysmon (systems monitor) that introduced the systems utilization capability.We use it for gathering performance data, CPU statistics, memory usage, disk and network utilization (see ISD Systems Status at: http://www.usc.edu/isd/services/systems_status/ ).

We would like to look at the feasibility of expanding and improving it's current offerings. Since sysmon is not a standard, we will have to develop our own tools for viewing, analyzing and displaying which we have lacked the manpower to provide.The proposal is to analyze options and develop enhancements to the current capabilities.Possible approaches include: using SNMP (simple network management protocol); pursuing further development of sysmon; and/or packaging it with existing shrink wrap tools.There are currently tools on the market but they are system specific and narrow in scope. We're envisioning a tool that can not only track Network and system utilization, but also have the ability to analyze Clarify (our call tracking system) or other database systems on concurrent platforms (network billing, host inventory etc.). 

Most people estimate software cost and schedule by analogy to their previous projects' productivity, cost, and schedule. But some key cost factors may change for the next project. COCOMO II can help people to assess the degree of change and to adjust their estimates. But the full COCOMO II model requires them to specify may parameters to do this. 

This project is to build an extremely simple Web user interface to enable people to adjust their analogy estimates, just by identifying which factors will be changing and by how much. We will provide a sample user interface and a background paper describing the approach. 

The paper and Strawman screen sequence:   http://sunset.usc.edu/fakharza/EasyCOCOMOII/

1. Overview Description

An on-line accessible of capability assessments of Systems Engineering within organizations.The project must handle multiple assessment types, provide statistical summaries about the assessments (and present the results graphically), have multiple levels of user access capabilities, and processes for submission, review and adding comments to assessment submissions.Each assessment submission has four types of information:Assessment information from an existing (paper) form such as those of EIA731,2) Observations,3) Findings (Strengths & Weaknesses),and 4) Results (=score).There is one existing analogous system, from the SEI for software capability assessments, for which we have design information.Also, there are also already existing (on-line) forms for capturing one of the assessment types information. 

2. Description Details

Multiple levels of user access capabilities:1) Control of Structure of Pages, access and Official DB structure; 2) Able to publish pages and modify links; 3) Viewers & Submitters, including1] Viewers (Anonymous but password protected) and2] Submitters who must identify themselves with submission; and4) the general public (answering the "What's this site all about?" question).

Statistical summaries including graphic representations of information:A controlled access portion of the SEI website provides examples of the desired/expected results.Access to that site will be provided to the team.

1. Overview Description

This project will build on an "out of the box" portal COTS package which provides a great deal of functionality.  The project will add customized applications for a flexible (to the Affiliate companies) way for their employees to register for access to the "Private Area", and tracking of user (of the "Private Area's) access and utilization of the site.

2. Description Details

Portal:  CSE has a product called "Hyperwave".  Details on it can be found at the following sites.

2.1. For white papers on Hyperwave, visit the commercial distributor's website at http://www.hyperwave.com/e/index.html and

2.1.1. navigate to http://www.hyperwave.com/e/products/his.html for an overview of the technical capabilities and access to the detailed white paper, also available directly at  http://www.hyperwave.com/downloads/e/his_techwhite_paper.pdf

2.1.2. navigate to http://www.hyperwave.com/e/products/hip.html for an overview of the "out of the box" COTS product and access to a white paper, which is also available directly at http://www.hyperwave.com/downloads/e/hip_whitepaper.pdf, or a portal tour, which is also available directly at http://www.hyperwave.com/e/products/hip_tour.html

2.2. An example of data plus meta-data display in a simple view: http://www2.iicm.edu/cguetl/papers.

2.3. European IST Prize: Hyperwave (with a short video)

The most distinguished European award for new products based on technologies for the information society. 3Grand Prizes of 200,000 euro each out of

URL: http://hyperwave.fast.de/ITP/Hyperwave
Last modified on: 15-Jan-2000 - 8K bytes - in English

2.4. Hyperwave Multimedia Presentation

Welcome to the interactive introduction to HYPERWAVE. Welcome message by Hermann Maurer (Vivo 870kB) First time users click here. Background,
URL: www.iicm.edu/hw_mm
Last modified on: 3-Mar-2000 - 12K bytes - in English
 
 

1. Overview Description

Working with a librarian who has been a client for other projects, and using the Hyperwave portal technology, develop a system to manage the content of the Center for Software Engineering's technical reports and presentations.  This includes capturing bibliographically correct meta-data, setting up multiple views for accessing the reports, and providing access statistics on a monthly basis.  Since there is already a large number of reports on the website, the project also needs to develop efficient ways to "ingest" each of them into the new environment.

1. Overview Description

Update the system and implement several of the "evolution requirements" identified for this project which was implemented in the last school year. The project documentation for last year is available at http://nunki.usc.edu:8082/team21b/.  The currently executing system can be seen at http://cse-class.usc.edu/cgi-bin/mbase/mbase.pl.  A major emphasis during construction (in the second semester) will be  regression testing using Rational's TeamTest.  If time permits, Rational's RequisitePro, adapted for MBASE documentation, will be used for requirements management and tracability.

1. Overview Description

Resume or restart work on an MBASE iPG (interactive Process Guide) authoring tool: http://nunki.usc.edu:8082/team20a.  The tool is to make it easier to develop and modify tools like the EPG you will be using this semester.  This project has quite a bit of database and user interface work.

1. Overview Description

Update the system and implement the (man in the loop) round trip "evolution requirement" identified for the EasyWinWin Viewer project  (called the MBASE-GroupSystems project) which was implemented in the last school year.  The project documentation for last year is available at http://nunki.usc.edu:8082/team19b.   The currently executing system can be seen at http://bsl16.usc.edu/EasyWinWin/index.html.   A major emphasis during construction (in the second semester) will be regression testing using Rational's TeamTest.  If time permits, Rational's RequisitePro, adapted for MBASE documentation, will be used for requirements management and tracability.

Client Name Dan Port, Ashish Soni
Client Contact: dport@ipal.com 310-578-8331

Project Sub-domain
Project Description

iPal provides internet wide taxonomy based matching services (see www.ipal.com) to the general public and personalized user-base notification for websites (e.g. new news articles matching a users specific interest from LATimes.com). Typically users create and maintain an “interest profile” by expressing their interests in particular items contained within the 300,000 item iPal taxonomy. On the other side “queries” are created which are Boolean expressions of taxonomy items attached to some form of content to be delivered to user profiles that satisfy the query. For example a query might be “/Sports/Basketball/Professional AND /Regions/Countries/Japan ANDNOT /Sports/Basketball/Amateur” and any user profile that has both the first taxonomy items but does not have the last “match” the query. After a user logs in their profile is instantly matched with all matching queries. Any time a user changes their profile or a query changes, matching is performed. For every match a “notification” is sent to the matching users regarding the content associated with the matching query.

 The matching service requires a client to receive notifications, edit interests, and submit match requests. Currently there are three clients – Windows, Web-based, and Wireless. A major shortcoming of these clients is that they simply display the notifications and do not enable the user to further process notifications once they arrive on the client. An advanced client would have an embedded scripting language that would allow users to further filter and direct notifications. For example not displaying notifications whose subject line contain a particular word, or directing all notifications that contain “High Priority” to a pager. For this project we would like to explore ways of making use of client side processing of notifications. This entails the following development tasks:

• determine desirable client side processing capabilities
• locate a suitable COTS scripting language that can accommodate the desired capabilities (e.g Perl, Rebol, TCL)
• build scripting capability using Rebol , Perl, or TCL into an iPal client (using of the existing clients or the generic Java client)
• implement example desirable client side processing capabilities using the scripting language

27. iPal Content Notification Production Assistant

Client Name Dan Port, Ashish Soni

Client Contact: dport@ipal.com 310-578-8331

Project Sub-domain
Project Description iPal provides internet wide taxonomy based matching services (see www.ipal.com) to the general public and personalized user-base notification for websites (e.g. new news articles matching a users specific interest from LATimes.com). Typically users create and maintain an “interest profile” by expressing their interests in particular items contained within the 300,000 item iPal taxonomy. On the other side “queries” are created which are Boolean expressions of taxonomy items attached to some form of content to be delivered to user profiles that satisfy the query. For example a query might be “/Sports/Basketball/Professional AND /Regions/Countries/Japan ANDNOT /Sports/Basketball/Amateur” and any user profile that has both the first taxonomy items but does not have the last “match” the query. After a user logs in their profile is instantly matched with all matching queries. Any time a user changes their profile or a query changes, matching is performed. For every match a “notification” is sent to the matching users regarding the content associated with the matching query.

Content notification tools are applications designed to support the creation and management of web site to user (B2C) touches. Such notifications alert users of areas on a web site that are of possible interest to a user based on their personal profile. Furthermore the items within an area have some form of time-value (e.g. “special today“, “event ends on”, news item “old” after, new item on, etc.), a particular location on the site, and a specific means to enter or access that area. These notifications have the following characteristics:

• Matching is followed directly by notification of all users that match (i.e. one-step, group touch). Any particular targeting is done in advance and applies to an entire user base.
• Notifications are non-interactive. Users are expected to respond by going to a URL (rather than “chat” or other form of user-to-user interaction).
• Notifications are expected to have a finite lifespan and be given a set expiration date. This does not preclude recurring notifications.

Typically content notification tools will take the form of “production assistants” that will be custom built applications that help content producing partners create notifications of their content during or prior to posting tailored specifically to their production process.  The output of such tools will range from suggesting taxonomy items for a producer to edit into queries, to automatically creating a queue of possible queries, ready for the match engine, perhaps requiring an editor approval. The editor could be housed at iPal or at a partner site.  The tool should also integrate with the partners existing editing content production process in a way that creates the least amount of overhead for the partner.

At iPal there are a team of “content specialists” that watch particular web sites for new things and then create notifications about them. At present each editor works with web sites that generally relate to a single branch of the taxonomy (e.g. for “Sports” perhaps would include “espn.com”) manually selecting the appropriate taxonomy category. We have been experimenting with classification products such as Autonomy and Rainbow that (after some training) can read content and provide suggested categories. These tools do not work well with a very large number of categories, particularly when most of the content is categorized in only a small percentage of then. What is needed is a production assistant that makes better use of Autonomy or Rainbow by including a web pages metatags and “pre-selecting” possible categories that a particular web site has used previously. In the event that none of the pre-selected categories are appropriate, the web pages keywords (available from Autonomy or Rainbow) should be used to select possible categories from the taxonomy. Any time an editor selects a category (either from the pre-selected or keyword search) the web-page should be used to further train Autonomy or Rainbow.
 

29.  Facilities Monitoring Project (From ISD)

The goal of the facilities monitoring project is to provide environmental monitoring for the University Computer Center Building.  This should be based on a stand-alone based system.

Phase 1 would be a specification phase where hardware and operating system are determined along with quantity and type of probes.
Phase 2 would be developing software to record the various sensor probes.
Phase 3 would be to add intelligent agents to make decisions on the status of the computing facility based on the sensor measurements and the rule sets for normal operation modality of the sensors.
Phase 4 would be to implement more intelligent agents to provide operations staff with suggestions as to the operational status of the various physical plant units (air conditioning, power etc.)

1. Station Data Project for the Web: Jean Crampon (Science Library; 740-4421; crampon@usc.edu)

2. Web Site Portal and Collaborative Filtering: - Candice Benjes (Norris Medical Library Room 102, (323) 442-1120, benjes@usc.edu )

3. Searchable Image Database for Pathology Slides: Maggie Wineburgh-Freed, (Norris Medical Library, Room B6;  (323) 442-1973; mwfreed@usc.edu )

4. Automated Link-checking for Electronic Journal Subscriptions: David Morse (Norris Medical Library, Room 201;  (323) 442-1134; dmorse@usc.edu )

5. Electronic Time Clock: Gameward C. Quan (ISD, Parking Structure D; 740-9013; gquan@usc.edu )

6. Photocopied Table-of-Contents (TOC) Faculty Current Awareness Service - John Glueckert (Wilson Dental Library, DEN-21 MC-0641; (213) 740-1439; gluecker@usc.edu)

7. Wilson Dental Library New Book List: Frank Mason (Wilson Dental Library, Health Sciences Libraries - (213) 740-1441; fmason@usc.edu)

8. Fulltext Titles Database: Amy Ciccone and Greg Fleming (Research Services, VKC Library – (213) 740-1958 or 740-2334; aciccone@usc.edu, gfleming@usc.edu )

9.  Smart Troubleshooting:  Joop Van de Wege,  (ISD Call Center, Jefferson 239; (213) 740-9400; joop@usc.edu ); Co-client: Deborah Hazzard (PSD 111; (213) 740-9270; hazzard) 

10. Streaming Video with Hi-Rez Push: Todd Richmond (Jefferson 203; (213) 740-4089; toddr@usc.edu )

11. Meeting Archive and Indexing: Todd Richmond (Jefferson 203; (213) 740-4089; toddr@usc.edu )

12. Easy Audio/Video Creation and Manipulation: Ruth Yarshapour (Jefferson 202; (213) 740-8454; ruthy@usc.edu )

13. Integrated Digital Archive Image Composer – Nava Herman (University Computing Center (UCC) 183; (213) 740-0785; navaz@usc.edu )

14. Access & Display of Slide Library Images: Howard Smith, Jane Cody, Mike Bonnet, Eunice Howe, Jim Steele (Howard Smith, (213) 743-4622, hsmith@usc.edu ; Dr. Jane Cody, (213) 740-6104, jmcody@usc.edu )

15. Integrated Digital Archive-Los Angeles (IDA-LA) – Database Design Component (Phase II). Contacts: Contacts: Sonia Chernobieff (GIS Research Lab, x7618, chernobi@usc.edu ), Wayne Shoaf (ISD-Digital Resources Cataloging, x4090, shoaf@usc.edu ), Dennis Smith (ISD-Information Technical Services, x2957, smith@usc.edu ), John Wilson (Department of Geography, x1908, jpwilson@usc.edu )

16. Integrated Digital Archive-Los Angeles (IDA-LA) – Ingest/Cataloging Design Component (Phase II). Contacts: Sonia Chernobieff (GIS Research Lab, x7618, chernobi@usc.edu ), Wayne Shoaf (ISD-Digital Resources Cataloging, x4090, shoaf@usc.edu ), Dennis Smith (ISD-Information Technical Services, x2957, smith@usc.edu ), John Wilson (Department of Geography, x1908, jpwilson@usc.edu )

17. Web Mail – Mike Pearce (Jefferson 207 & University Computing Center (UCC) 210; (213) 740-7824 & 740-7783; mpearce@usc.edu )

18. Network Utilization Tool: Russell Kaurloto (Jefferson 245; (213) 740-3013; kaurloto@usc.edu )

19. Easy COCOMO II
     Clients: Cyrus Fakharzadeh(fakharza@sunset.usc.edu), Barry Boehm(boehm@sunset.usc.edu )

20. INCOSE Capability Assessments Database
     Client [on campus]: A Winsor Brown
     Assistant Director, USC Center for Software Engineering
     Email: awbrown@sunset.usc.edu 

     Client [on campus] senior management:  Dr. Barry Boehm
     Director, USC Center for Software Engineering
     Email: boehm@sunset.usc.edu 

     Ultimate Client & Consultative support:  David G. Beshore, P.E.
     Systems Engineering & Software Development Process Mgr.
     Mission & Requirements Management
     Boeing Rocketdyne Division
      6633 Canoga Ave/ MS GB15
     Canoga Park, CA 91309-7922
     E-Mail: david.g.beshore@boeing.com

21.CSE Affiliates Private Area Portal

     Client [on campus]: A Winsor Brown
     Assistant Director, USC Center for Software Engineering
     Email: awbrown@sunset.usc.edu 

     Client [on campus] senior management:  Dr. Barry Boehm
     Director, USC Center for Software Engineering
     Email: boehm@sunset.usc.edu 

22CSE Tech Reports System
     Technical consultative client:  Julie Kwan
     Research Librarian, ISD - Business Library
     HOH 201, Mail Code: 1422
     Email: jkwan@usc.edu 
     Tel: (213) 740-4420

     Ultimate Client [on campus]: A Winsor Brown
     Assistant Director, USC Center for Software Engineering
     Email: awbrown@sunset.usc.edu 
     Tel: 213/740-6599

     Ultimate Client [on campus] senior management:  Dr. Barry Boehm
     Director, USC Center for Software Engineering
     Email: boehm@sunset.usc.edu 

23. MBASE Deliverables Manager
     Client : Ebru Dincel
     CS577 TA/RA, USC Center for Software Engineering
     Email: Dincel@sunset.usc.edu
     Ultimate Technical Client: A Winsor Brown
     Assistant Director, USC Center for Software Engineering
     Email: awbrown@sunset.usc.edu 

     213/740-6599
     Ultimate Client senior management:  Dr. Barry Boehm
     Director, USC Center for Software Engineering
     Email: boehm@sunset.usc.edu 

24. iPG Authoring Tool:
     Client : Hwan Dong
     CS577 TA/RA, USC Center for Software Engineering
     Email: hwandong@sunset.usc.edu 

     Ultimate Technical Client: A Winsor Brown
     Assistant Director, USC Center for Software Engineering
     Email: awbrown@sunset.usc.edu 

     213/740-6599
     Ultimate Client senior management:  Dr. Barry Boehm
     Director, USC Center for Software Engineering
     Email: boehm@sunset.usc.edu 

25.  EasyWinWin Viewer Evolution:
     Client: Mohammed Al-Said
     Research Assistant, USC Center for Software Engineering
     Email: AlSaid@usc.edu 
     213/740-4493
     Ultimate Technical Client: A Winsor Brown
     Assistant Director, USC Center for Software Engineering
     Email: awbrown@sunset.usc.edu 
     213/740-6599
     Ultimate Client senior management:  Dr. Barry Boehm
     Director, USC Center for Software Engineering
     Email: boehm@sunset.usc.edu 

26. Advanced iPal client: Dan Port.310-578-8331,  dport@ipal.com ;Ashish Soni, 213-749-3558, asoni@usc.edu

27. iPal Content Notification Production Assistant: Dan Port.310-578-8331, dport@ipal.com ;Ashish Soni, 213-749-3558, asoni@usc.edu

28. Targeted Advertising Prototype.: Dan Port.310-578-8331, dport@ipal.com

29.  Facilities Monitoring: Diane Eastman (Jefferson 102c; (213) 740-3030; deastman@usc.edu)