Picasso Recommendation
Mission The School of Dentistry for the University at Buffalo (UB) uses a custom application called Picasso to manage their clinic operations. Developed by two engineers in 1997, it manages the entire dental process, from scheduling patients to submitting Medicaid claims. Our mission is to measure Picasso s technical architecture against contemporary standards and practices. The objective is not to compare Picasso with other software solutions.
Terminology / Definitions Executive Overview Detail Documentation 1 Includes: Business Process Description Goals and Objectives o Information Technology o Education o Research o Patient Care User Types Technical Overview o Application o Database o Network o Costs Feature Highlights User Feedback 1 Separate document
Technical Analysis Application Design Topic Rating Notes Development Process Process definition, from new change / defect report to implementation. User Interface Technology How the users interact with the data / system. Annual Costs Hardware The infrastruacture used by Picasso. Software The applications used to support Picasso. Employees People directly employed by UB SDM who develop and support Picasso. Architecture Extensibility Ability to extend the system. Accounts for future growth. Scalability Ability to support more transactions, utilization and throughput. Redundancy Ability to replicate core system components in an effort to ensure reliability. Data Backup Ensuring the data is safe from corruption. Security Protecting the data from unauthorized usage and data corruption. 2-tiered design appropriate for when Picasso was developed, but antiquated compared to current standards and practices. No code repository. Requirements captured in documents. No automated scripts to expedite regression testing. Restricted to Microsoft operating systems. Limited field experts. Inexpensive annual maintenance costs. Inexpensive annual maintenance costs. Significant salary expense compared to production value. Citrix and Sun Ray at the infrastructure layer. Citrix and Sun Ray at the infrastructure layer. Citrix and Sun Ray at the infrastructure layer. Significant effort made to ensure the data is stored and protected. Multiple layers of authentication. Application permissions managed by IT. Developers have access to HIPAA data.
Feedback User Topics 1. IT is unresponsive to user s needs 2. Insufficient reporting 3. Less than 100% paperless 4. Severe gaps supporting research initiatives Technical Topics 1. PowerBuilder is not a commonly learned development platform 2. Insignificant amount of depth in application expertise 3. No official defect / enhancement request tracking system 4. No code repository / review system
Recommendation Multi-Tiered Architecture o Streamlined deployments o Rapid fixes / enhancements o Easier troubleshooting o Improved scalability o Simplified security o APIs for external exposer Leverage Data Model Reporting Data Warehouse o Leverage Reporting Services Research Data Warehouse Utilize Business Rules Engine
Benefits Multi-Tiered Architecture Leverage commonly learned development languages to reduce the dependency on specific expertise Streamlined deployments Rapid fixes / enhancements Easier troubleshooting Improved scalability Simplified security APIs for external exposer Leverage Data Model Reduced development / support effort Reporting Data Warehouse Improved reaction to reporting needs o Leverage Reporting Services Research Data Warehouse Dedicated environment for research Utilize Business Rules Engine Reduced development / support effort No staff increase to support post-rewrite
Rewrite vs. Replace Risk ROI Cost Time Training Usability Support Performance Documentation Mobility Data Accessibility Re-Write Replace
Rewrite vs. Replace Risk ROI Cost Time Re-Write Must capture all existing functionality without introducing new defects, plus enhancements. More control over feature set / defect prioritization; once stabilized, support burden should be minimal. Increase in project cost, but will decrease over time as support stabilizes. Potential to gradually implement changes. Replace Certain to lose commonly used features. Typical 20% annual maintenance cost, will likely increase over time; cloud solution will incur subscription cost. Trial, purchase, training and implementation; annual vendor support cost + staff support cost. Trial, purchase, training and implementation; potential customizations. Training Minimal learning curve. High learning curve on new software. Usability Tailored to the business. Determined by the vendor. Support SLAs are tailored to the constraints of IT. Determined by the vendor. Can be negotiated for a price. Performance Managed by IT. Determined by the vendor. Documentation Incured at project cost, manageable over time. Determined by the vendor. Mobility Determined and managed by the business. Determined by the vendor. Data Accessibility Data elements and business rules managed by application. Unlikely to support all required or desired data elements.
Rewrite vs. Replace Packaged software will improve and standardize processes o FALSE: Offer new features, but also remove differentiating features currently used o o New software requires significant training Cannot underestimate the value of familiarity Packaged software will force process to comply with standard business practices o TRUE: Higher risk of losing market differentiating factors Packaged software will reduce costs o o o o FALSE: Annual support and maintenance costs for on-premise software, annual subscription costs for cloud software On-premise software Hardware and network costs Typical support and maintenance expense is 20% of license cost Additional licensing costs (database, application server, etc.) Cloud solutions Subscription costs vary, depending on number of users No control over software updates Requires external network connectivity Cannot place value on waiting for urgent enhancements and fixes
Rewrite vs. Replace Packaged software will allow for a reduction in staff o FALSE: Requires IT experts to: Troubleshoot issues Setup and configure the system Packaged software is easier to install, configure and maintain o FALSE: Configurations for server utilization must be made for on-premise software o Configurations required to integrate with other systems (MiPACS, Active Directory (AD), etc.) o Security model management Packaged software is easy to upgrade o FALSE: Regression testing often finds new defects o Integrations with other systems may fail
Roadmap Agile Development Process o Identify Functional Milestones Test and Implement Features as Developed Leverage Automated Testing o Close Collaboration with Users o At least 1 Year Effort Development Roles o 1 Program Manager / Business Analyst o 2 Data Warehouse (DW) Engineers 1 Engineer for the Research Warehouse 1 Engineer for the Reporting Warehouse o 1 Business Intelligence Engineer o 5 Software Engineers 2 Web Interface 2 Web Service 1 Business Rules Engine 1 UB SDM IT Software Engineer
Decision Making Does the process support the software? o Will the business change their process to accommodate software limitations? Does the software support the process? o Will the software change to accommodate business process changes? OR
Questions