There’s little doubt as to whether Pokémon Go has entered into mainframe shops. The game – and discussion about the game’s development and support infrastructure – are omnipresent. But has Pokémon Go captured the imagination of tomorrow’s mainframe developers?
“Don’t Need Your Stinkin’ UX”
In a post last September, Ben Wilson bristled with contempt for what he called “Consumer UX (User Experience)” . Unlike consumer UX, Wilson argues that enterprise UX:
• Automates more complex domains
• Has more complex screens
• Has a greater number of screens
• Has a smaller per-screen budget
He worries that enterprises are determined “. . . not only to web-enable and cloud-enable [mainframe applications], but to all-out upgrade them to consumer-grade experiences.” Wilson believes that the mega million-line enterprise app may become a historical curiosity, like sending a person to the moon.
But over at Virtel, the watchword is Steady as She Goes. Virtel offers tooling to replace the UIs that support the early 1970’s era 3270 display terminal format. The 3270 screen is iconic, frozen in computer history as the classic “green screen.” A legacy app, it was once assumed, meant a green screen app. (Since then, “legacy” has shifted to mean any app that cannot be web-enabled, or is simply old or gnarly to maintain).
It turns out that the layered approach embodied in the original 3270 design allowed Virtel to offer solutions that left substantial portions of mainframe apps intact. Virtel’s proposition was this: tweak the back end of your CICS, IMS and other host apps a little bit, and we’ll give your users a browser-ready experience on the front end.
Virtel touts the superiority of their modernization approach, but IBM believes users should consider its Rational Host Access Transformation Services (HATS) tool. HATS offers enterprises still knee deep in green screens a way to step out into web, portlet, or mobile device user interfaces. The Company believes that the rules-based transformation engine in HATS is powerful enough that developers can make some applications browser-ready within a day or two.
For newly developed z/OS applications, IBM offers fully web-ready development tooling, such as the WebSphere Application Server and Mobile Toolkit. IBM positions the product as SOA-ready, supporting a “best-in-class Ajax development kit.” WebSphere apps can be connected to various webservers, including z/OS webservers through a WebSphere Web Server Plug-ins.
There’s plenty of z/OS webification tooling to go around.
Unintended Consequences: User Expectation
Updated screens will likely produce unanticipated user expectations.
IBM’s recent announcement of zEvent illustrates the primacy of mobile apps for monitoring and management. As operational intelligence becomes more user-centric, the focus will likely move from simply managing response time during peak periods to managing for resilience.
In addition to zEvent, IBM proposes expanded use of SMF data, especially for operational decision support. IBM’s Tivoli Decision Support (TDSz) tool is positioned to “help improve operational planning, cost management, responsiveness, and decision-making process in organizations to effectively manage economic performance of the IT investment.”
Syncsort’s Ironstream acts as an even more aggressive collector of SMF data, delivering low level performance and security event data in real time to Splunk.
Granularity: Going Down
Achieving lower granularity in systems management will mean different things to different users. But here are a few examples:
• Marketers about to launch a campaign want to see real time trends from sales triggers, a service offered by outfits like Cognism.
• When HR fills a job requisition, it wants its HRMS to integrate salary history for this and related positions, as well as access petabyte-class data from ADP’s anonymized repository of worldwide salary history.
• Manufacturing engineers once content with 36 months of log history on high value factory floor equipment now want lifetime data to improve predictive maintenance.
• Public safety officials once content to receive low bandwidth in-roadway sensor streams want to see both real time sensor information combined with live video on demand. They also want systems designed for cars and trucks to be bicycle-aware.
Mainframers know that z/OS has the capacity to satisfy these user expectations on the back end. Capacity and scalability can be had. That part’s not in dispute.
Less easily resolved, though, are the demands placed on mainframe developers to create responsive interfaces to complex backend systems across desktop, tablet and mobile platforms. Screen-switching, once an accepted practice, must be replaced by portlets, greater interchangeability of web components, and an API-driven ecosystem. SOA is moving toward microservices.
Expect Still Greater Expectations
The use of a standardized browser interface creates user expectations across a broad spectrum of functions.
Cut and paste Users expect cut-and-paste to work across media types, even video and recorded phone conversations, or the threaded history of all interactions with a customer. Gestures for cut-and-paste are relatively new, but already ubiquitous on mobile and moved onto desktops with Windows 8 and 10.
Menus Users have expectations of where to look for menus. They have learned to look for the “hamburger” icon, but may not know where to find 90’s era pulldown, first-letter-oriented menus.
Swipe to Remove Users don’t mind taking responsibility for computing resources, but they have certain expectations about how to access resource management tools. Think iOS fast app switching with swipe, or Windows down-swipe.
Accessibility The pinch style zoom feature is helpful for the visually impaired, but has also become a de facto standard for image manipulation. This expectation will extend to visual analytics.
Speech-input Most apps on the major mobile platforms allow for speech input in textboxes. While adoption has been slow (except for in-vehicle Bluetooth settings), that’s changing quickly.
Search Users expect to be able to conduct free-form search, even with poorly formed search queries, and to do so over vast datasets.
These demands place new requirements on mainframe UX.
With the Ease and Confidence of a Pokémon
In his piece, “What Does the Mainframe Have to do with Pokémon Go,” Compuware’s Michael Siemasz recently challenged the z/OS community to embrace change:
“Mainframe companies can take a few cues from the innovative steps taken by Pokémon and mainstream the mainframe. This means including the mainframe in their broader enterprise Agile/DevOps processes by enabling IT staff, regardless of their familiarity with the mainframe, to do mainframe-related tasks with ease and confidence. To do that, companies must provide the next generation of developers with the right culture, processes and tools.”
Good advice, but only if developers have a good sense of what to expect from users. In a self-service, knowledge-oriented enterprise much will be demanded of tomorrow’s apps.
If mainframe solutions fall short of user expectations, the reason won’t be the mainframe’s inability keep up with the computing demand. It will be the inability of mainframe apps to adapt to user expectations.
OK, so some users expect petabytes of analytics to be filtered in a swipe of the hand or a few words spoken to a search engine.
Mainframe apps have always excelled at performance. Now they must compete on usability, too. Even if it means a self-driving, self-collecting version of Pokémon Go running on a mainframe to make the point.
If only to be able to say, Augment that global geospatial reality, baby.