BOWIE, MARYLAND – MAY 29: Local residents wait in their vehicles as they line up for a coronavirus … [+]
Getty Images
Software coding, like life, is difficult. But that’s what’s appealing for so many software application developers; they relish the chance the learn the details, syntax, nuances, form and structure of coding platforms that enable them to create software application functionalities that users will ultimately find useful, enjoy and perhaps even occasionally marvel at.
But coding is in fact so difficult that we don’t have enough software developers on the planet to deliver on the seemingly insatiable demand for apps (and all the custom-build optimizations and customizations) needed in the typical modern enterprise IT stack.
Low-Code/No-Code
The last half-decade in particular has seen the rise of Low-Code/No-Code software tools to help try and circumvent some of the challenges developers face. Looking at Low-Code in particular, these are tools that provide templates, references architectures and various types of (often best practice approved) shortcuts for software engineers to use at the coalface of programming and development.
Low-Code still takes a lot of coding and Lower-Code would perhaps be a more accurate label. Simpler still is No-Code, which are platforms designed to abstract away the complexity of the code beneath and give what are often non-techie business users the ability to use heavily graphical drag-and-drop interfaces to create software functionalities.
Both techniques are useful, but both may be partially or potentially responsible for the development of less than perfect software if the applications they are used to create are not a) properly policy-policed by the organizations using them b) properly secured and locked down as per company and corporate compliance guidelines and c) properly tested before they form part of the live working fabric of an organization’s IT stack.
Tracking & testing
This latter testing element is what keeps Copado up at night. The firm is a DevOps (developer + operations) platform that works to power enterprise software, specifically within the realm of Salesforce deployments.
The company runs an annual analysis among developers and quality assurance professionals to collect hundreds of data points and attempt to provide some insight into the software testing practices of what are billed as the ‘most sophisticated companies’ using Salesforce.
According to Pat McQueen, senior vice president of growth at Copado, testing in Salesforce becomes more critical as the ecosystem of custom applications, managed packages from the AppExchange and integrations are involved. Not only do organizations have to adapt to changes when it comes to their Salesforce deployments, but they also have to track integrations with other systems that have their own release cycles.
“With this rapid innovation comes the need for scalable end-to-end testing. A recent Forrester report confirms organizations using Low-Code platforms like Salesforce’s Low-Code tools can’t afford to ignore automated testing,” insists McQueen and team, in a press statement.
The problem at hand here may stem from the fact that so many companies are still engaged in traditional testing methods – such as manual testing or script-based solutions – when they could be plugging into the new era of autonomous system controls that offer a more automated approach to testing.
Payback time for technical debt
The consequence of this ‘automation oversight’ (not a de facto industry term yet, but it could soon be so) is that the quality of software releases suffer since there isn’t enough time to test changes before deployments. Defects found later in the release cycle are almost always more expensive to fix due to the technical debt incurred through rapid deployment, which needs to be ‘paid back’ in code fix time much like any financial debt requires payback.
“In a world where Low-Code/No-Code platform usage is coming to the fore, companies are clearly finding that automated testing is the best way to move fast and not break things,” said Copado’s McQueen. “Teams that prioritize their testing strategy outperform those who don’t. An investment in automated testing enables business agility which leads to long-term commercial success in a competitive market.”
The central message from Copado is that manual testing isn’t scalable and is expensive. This means that software production defects can have a massive financial impact on companies. In its own research, 41% of organizations don’t have enough time to fully test all changes before a release due to time constraints that force teams to reduce the scope of testing. As many as 92% of firms experience production issues (defects uncovered) each year due to inadequate testing.
Copado suggests that the cost of not adopting automated testing is a much higher incidence of production failures despite greater investment in the testing team. Teams using automated testing experienced less than half as many production failures per year as a result of releases (a median of 3 per year instead of 7).
Teams using automated testing release 50% more frequently than those relying on manual testing (34 times per year on average, compared to 22 times per year). Teams using automated testing were 50% more likely to complete all of their test plans for each release (67% of teams using automated testing reported completing all their tests for each release vs. only 45% of teams relying on manual testing).
Self-serving survey stats
Copado’s survey stats are interesting enough, but there isn’t an IT industry on this planet or the next one that isn’t engineered to deliver some sort of self-serving payload designed to reinforce the central technology proposition of the firm behind it.
Copado thinks that there’s not enough automated testing going on at the Low-Code end of the Salesforce platform’s software development and implementation spectrum. Hey guess what? Copado sells DevOps playbooks with high-grade business process integration know-how suitable for use in Low-Code SaaS platforms that need robust security and governance. Imagine that… how did the company think to assess the market on this issue?
Sarcastic naysaying and corporate gloss removal notwithstanding, Copado has a point i.e. if we are going to use Low-Code (and even No-Code) efficiencies to create software applications in the new world of AI automation advantage, then we should also be applying those autonomous accelerators to the testing function too.
Source: https://www.forbes.com/sites/adrianbridgwater/2022/01/12/why-low-code-isnt-no-test/
