CollegeWhere™ 
How do you match a student with the best college that they are likely to be admitted to.
That was the problem that CollegeWhere set out to solve. It was a problem we solved. CollegeWhere’s innovative technology won it a Product of the Year
Award from Family PC-- plus the undying gratitude of thousands of anxious parents.
CollegeWhere™ was a six month effort, with an additional 4 week revision that resulted from its successful introduction at the E3 tradeshow. We
delivered a fully tested, shelf ready, shrink wrap product in volume, on-time. It was the first application built with the Wolf Rock Application
Generation Technology, so we made it with at least half an hour to spare.
Although CollegeWhere was a shrinkwrap product, the technology we created to accommodate the huge database and integrate it with a
user-friendly interface formed the basis for a number subsequent web based and internet data exchange models.
CollegeWhere contains an expert system built on a neural database. Wolf Rock buit both the database and the non-linear dynamics based
expert system engine. The decision support model was a product of a simulation and analysis effort that determined the criteria, on a college by
college basis, used to make the admission decision. The database was based upon a set of attributes that normalized college capabilities to a statistically significant level.
|
|
|
Key Project Data
|
|
Project Duration
|
6 months+4 week revision
|
|
Key Technologies
|
Human Factors, Database, Modeling, Simulation
|
|
Major Technologies
|
Predictive System, Non-linear Dynamics, Compression, Statistical Analysis
|
|
Key Success Factors
|
Finish Date, User Satisfaction, Accuracy, Support Metrics
|
|
Surprises
|
The level of enthusiaum of the CompUSA buyer.
|
|
Client Platforms
|
Windows 3.1, 95, 98, Mac 7.1
|
|
Development
|
Progress, Wolf Rock AppGen, Borland C, Turbo Pascal, Visual Basic, MS C.
|
|
|
|
Unlike US News, which seems to believe that a 20 point variance in SAT scores, and a small difference in student faculty ration, which varies from year
to year, can de used to create an absolute ranking, we normalized the schools data using a four year basis to create groups with statistically significant differences.
To ensure the level of quality required for retail volume we constructed our AppGen code generator to produce code for the thousands of complex data relationships. The
result was a manufactured application with very high quality metrics.
|
|
With megabytes to spare |
|
We developed technology to insure that he entire application, including a 3400 page college info book, and all of the databases fit on a single floppy.
When the first CompUSA order came in it was for a CD-Rom version. We had to take the engineers shoelaces and belt away.
|
Now that’s quality |
|
Our President’s home phone number was put on the right on the box as technical support. In hind sight it was an amazing feat of bravado.
|
Breaker, breaker 19 |
|
We
had 4 weeks to produce the final artwork, kits, updated software and manufacture the final packages. We ran right down to the wire and in order to shave time we even rented the truck to pick up the cartons of product and deliver them to the distributor
|
|
|
