Last year I blogged about an initiative to take TitanDb to the Apache Foundation as an incubation project. Yesterday, the next step for TitanDb was announced, when the Linux Foundation welcomed JanusGraph, a project which uses TitanDb as its starting point. The current documentation will look familiar to anyone who has investigated TitanDb, but the news will be welcomed by the TitanDb community. So it's worth keeping an eye on the JanusGraph website.
A few weeks ago, IBM announced Watson Workspace, the final name for Project Toscana, and its API Watson Work Services. The product itself has similarities to Slack or Microsoft Teams, but this post is not about discussing a comparison of the products. It's about the API backing it.
Watson Work Services is a REST API that uses GraphQL, a method of querying and posting via REST that focuses on configurability. Whereas traditional REST services have fixed endpoints that take fixed parameters and return fixed data objects, GraphQL is a sort of "API-as-a-service". You call an endpoint, pass a JSON object determining what elements of which data you want to return, include any dynamic variable values. The queries are passed to an engine at a REST service endpoint which parses the JSON passed, replaces any variables with the dynamically passed values, and returns just what the application or user asks for. This may include data from what, in a traditional REST API application, would be from different REST endpoints. For example, to get members and messages from a space, you might need to make a REST call to get the space ID, then another REST call to get its members, and a third to get the messages.
Over the last few months a lot of what I've had to do has come from combining frameworks. Usually one of those frameworks is Vaadin. But whenever you're combining frameworks of any kind, it usually means some content is pre-configured, which may conflict with settings in another framework. If you're not familiar with the framework and the technologies in use, it's a lot like looking at hieroglyphics without a Rosetta Stone! The result is a lot of learning on the job.
Earlier this week I had problems with high CPU utilisation and had to restart my PC. I took the opportunity to bite the bullet and install some Windows Updates. What I saw brought my mind back to UX and coding of applications. For at least five minutes, the progress displayed as "100% Complete". It prompted me to issue the following tweets:
Developers, progress bar should never show 100% complete. It should be 0% of next process or gone. Windows updates 100% complete for minutes
— Paul Withers (@PaulSWithers) 18 October 2016
I've been involved in developing training materials, tutorials, videos, online documentation and even books. So I've gained a full appreciation of the effort involved in not only creating good documentation but maintaining it too. Rene Winkelmeyer wrote a good blog post today about developer experience and his points are very valid.
Over the last couple of years I've seen a number of approaches as I've dug into a variety of new technologies.
Since earlier this year when I started trying to get a better handle on the breadth of graph database options available for a developer, Titan has been an option I have kept in regular contact with. It's fair to say there has been a lot of uncertainty about the prospects for Titan. But there have been some interesting developments regarding Titan during the summer. IBM Graph has reached GA on Bluemix, albeit with only REST access, which may not appeal to Java developers, particularly those familiar with Titan and comfortable with natively handling vertices and edges. And more recently there has been a lot of work on integrating Titan with ScyllaDb, which provides a long-term option for using Thrift as a communication mechanism between Titan and the backend database.
The bulk of my experience with application development has been building workflow-related rich client and web applications on NoSQL databases, typically IBM Domino. The challenge in the Notes Client was to provide dashboard-style displays and a good way to display documents for action by the current individual. Private views can be used, but impact database performance. So, typically, the approach is to display views that present a scrollable table of data. Domino's document-level reader security is then used to ensure only the appropriate data is visible. If data is archived appropriately, performance of the database is good enough for many reasonably-sized applications. (Of course, archiving is often omitted from scope of the first phase for the rapidly-developed application, and becomes a case of "out of sight, out of mind".) But with the increasing prevalence of web applications replacing Notes Client applications, the ability to display "my documents" and use structured searches to display a targeted subset of documents was much easier.
Communication & transparency are critical to lifeblood of any community, open source or closed. Seeing impacts in more than one community
— Paul Withers (@PaulSWithers) 28 July 2016
Before I started working with IBM's XPages framework in 2009, I was starting to use AJAX calls in web applications and starting to dig into Dojo charting options for an application. So not unsurprisingly, when I started with XPages I blogged quite a bit about Dojo charts and understandably chose to write the Dojo-related chapter and a half of "XPages Extension Library". I also contributed a Dijit Tooltip custom control and an extension to the Dojo Legend component, to allow more sophisticated formatting of the legend.
IBM Websphere Liberty is a widely-used modern lightweight server (less than 70Mb, I have several running on my PC) that allows modern Java EE development with Java 8. The server is free for development and build server, with limited production use (up to 2Gb of JVM heap space across all instances for the organisation). Web applications are deployed as standard web archives, which means it's easy to download and install demos of various frameworks and easy to use Maven. This is something I've fought with when developing OSGi plugins, because Maven is designed to manage and install dependencies be compilation. On the other hand OSGi best practice is to pull them from other plugins, defined in the MANIFEST.MF. At the very least OSGi needs the plugins adding to the classpath in the MANIFEST.MF. Tycho is designed to bridge the gap between what Maven expects and what OSGi expects, but it's not great. So a standard web application, sucking any jar files in via Maven, makes life a lot easier.