Ondot
As a well-funded start-up company, Ondot Systems has created a mobile card services platform that delivers real-time insight that is easy to understand and appealing for consumers. Customers can view real-time activity on their plastic and virtual debit/credit cards, receive notifications quickly, and participate in customizing their cards to best fit their spending needs. The company is headquartered in San Jose, CA, and employs approximately 50 people based in the US and India, including a software engineering team of 40 developers.
Business Need
As is the case with all funded start-up firms, investors look for the venture to deliver initial products to market as soon as possible. The software developers at Ondot thus constantly work under tight deadlines when building and testing new code.
“When we first started the company, we conducted ad hoc, ‘over-the-shoulder’ code reviews,” says Ondot President Dr. Vaduvur Bharghavan. “As our product development pace increased, we reached the stage where we needed an automated code-review solution to help us catch the errors that developers are prone to make when working under pressure.”
In addition to helping identify coding errors earlier in the development lifecycle, Ondot also wanted to streamline the code review process so developers could spend more time on creating and improving code. “Unless the review process is well-documented and developers have a clear idea of who provided each piece of specific feedback as well as the time of the feedback, they can spend a lot of time just trying to compile and sort through the feedback,” Bharghavan says. “We needed to reduce the code review project management time as much as possible.”
An additional code-review challenge Ondot needed to address is that some software developers work in India while others work in the US—in two vastly different time zones. “Typically, we have multiple people working together on code for applications in each of our four sub-systems, including the lead developer and two or three people to review their work,” Bharghavan says. “With each team including at least one person from the US, business-hour meetings are not always possible for the entire team at the same time.”
Targeted Solution
To address the Ondot need for an automated code-review solution, Bharghavan immediately turned to SmartBear Collaborator—without giving serious consideration to any alternatives. “The software development team at the previous company I founded (Meru Networks) deployed Collaborator,” Bharghavan says. “And the results proved very successful in identifying coding errors and helping facilitate the overall review process.”
Given the solution’s success at his previous company, along with its ease-of-deployment and low cost, the decision to deploy SmartBear Collaborator was an easy one for Bharghavan. As a full-featured code review solution, SmartBear Collaborator is the only code review tool that also enables reviews of other project deliverables such as user stories, test plans and user documentation.
SmartBear Collaborator also transforms code reviews into easy and efficient tasks for developer teams, which leads to greater adoption of code reviews as a best practice and helps programmers improve upon code development. Collaborator also allows other developer team members to collaboratively provide feedback, and for the author to track and take action against each comment.
Reinforcing the Ondot decision to deploy SmartBear Collaborator was the experience of several members of the software development team who had also used Collaborator in the past. The solution deployment proceeded smoothly, and the team began testing applications immediately.
“We believe in repeating solutions we have had success with at previous companies,” Bharghavan adds. “This approach gives us a level of comfort in the technology partners with which we choose to collaborate. SmartBear fits the model perfectly by providing reliable software testing solutions and continuously striving to enhance the capabilities.”
Approximately 30 Ondot software engineers use SmartBear Collaborator in conjunction with the company’s Java-based development environment in which Ondot relies on SVN (Subversion Version Control) for source-code control. Collaborator also supports Bash shell scripting and C++, both of which are key components of the Ondot development environment.
“For start-up companies under pressure to bring their software products to market as quickly as possible, SmartBear Collaborator is a necessity. The solution helps enforce code reviews and detects code errors that are likely to occur when developers work tight deadlines. ”
— Dr. Vaduvur Bharghavan, President
Benefits and Results
As expected, with SmartBear Collaborator Ondot can now easily test code and track the progress of reviews as well as clearly delineate the comments of each review—even when developers from both the US and India are involved. Ondot now mandates developers conduct application code reviews through SmartBear Collaborator as a required process in the development lifecycle. The Ondot development system automatically blocks code check-ins until code reviews are completed.
“Collaborator allowed us to create this as a hard gate in our code-development process,” Bharghavan explains. “Although any system can be circumvented, our developers completely buy into the value Collaborator delivers and stay true to the check-in process based on their experiences with Collaborator. Giving us this capability is critical—even with a small development team, the pressure of a start-up environment can lead to check-ins before code reviews are performed—which can cause major problems down the road if bugs are discovered late in the development lifecycle.”
Ondot also now has code review traceability and metrics for measuring code review time, the number of bugs discovered, and other performance indicators. Management leverages this information to clarify any anomalies in code review time, which leads to improvements in overall code review efficiency.
“The metrics particularly assist us in working with junior developers to help them improve their code development skills,” Bharghavan says. “They can be prone to mistakes when naming or creating very large functions across areas such as quote styling, quote structure, and how functions are invoked. When in a rush and particularly with a distributed team, problems are more likely to happen, but Collaborator helps keep us on track."
As for the ultimate benefit for Ondot, SmartBear Collaborator is helping the company accelerate the timeline for bringing its products to market. “Nothing is more insidious to the development lifecycle than someone checking-in code under pressure right before a deadline without conducting a code review,” Bharghavan says. “If you then catch the problem down the road after more code has been added, fixing the problem becomes a lot more costly and requires a lot more time. A 10-minute code-review with Collaborator can prevent a multi-day effort down the road.”
Business Challenges
- Eliminate informal, ad-hoc code review discussions.
- Accelerate the code review process to assist developers under tight deadlines.
- Enforce completion of code reviews prior to application check-ins.
- Facilitate code reviews among development team spread across two continents.
SmartBear Collaborator Solution
- Offers proven track record of success and deploys easily to enable immediate use.
- Supports documentation of other project deliverables that support code reviews.
- Generates metrics to evaluate code review performance.
- Facilitates greater adoption of code review as a best practice.
Measurable Results
- Assists in bringing software products to market faster.
- Identifies code bugs early in the development lifecycle to reduce repair costs.
- Enforces code review prior to application check-in to reduce software bugs.
- Helps management advise junior developers on how to lower review times.