NetKernel News Volume 5 Issue 2 -Data and Reality, Back to the Beginning Series
search:

NetKernel News Volume 5 Issue 2

January 24th 2013

Catch up on last week's news here, or see full volume index.

Repository Updates

The following updates are available in the NKEE and NKSE repositories

  • layer0 1.101.1
    • Update to NKFAsyncRequestListener so that java.lang.Errors as well as java.lang.Exceptions can be intercepted if desired by endpoints
    • Update to endpoint base classes to change semantics of declareUnhandledExceptionsExpired() functionality. Now resolution failure exceptions will not cache regardless of the setting of this flag.
  • module-standard 1.61.1
    • BranchMerge overlay fixed so that nested BranchMerges (that is BranchMerge inside BranchMerge) will generate the correct metadata for the space explorer views on the management console.

Please note these updates are published in the 5.2.1 repository.

Tom's Blog: Back to the Beginning Series

For some time now I've been meaning to highlight an excellent set of articles that Tom Geudens has been writing.

"Back to the Beginning xxxxx" in which Tom trys to step back in time, to cast off the expert's cloak, and present NK/ROC from an absolute beginners perspective. He starts with a small set of assumptions...

  • You want to take a look at NetKernel.
  • You have succeeded in installing NetKernel in a directory that I'll from here on describe as installationdirectory.
  • You have an instance of NetKernel running and when you point your browser to http://localhost:1060, it shows the NetKernel Backend Fulcrum.
  • You are able to extract a directory from a zip archive.

Here are the articles in chronological order...

The last one, on logging, nicely sets up a very interesting technical development that Tony and I have been discussing... I'll leave that as teaser for now...

Great stuff Tom - and you still found time to get the Belgian government open data services out (Belgian not Dutch mind. They're going for Open Linked Data on NK in a big way in Benelux)

NetKernelROC Community Update

Things have been picking up pace in the NetKernelROC community. In a recent informal conversation, Randy mentioned a few of the activities that have been completed in 2013 and other projects that are underway or planned for 2014. With all this activity it seemed like a good idea to step back and let the community take the floor to summarize it...

I am writing on behalf of the NetKernelROC community to let you know what we have accomplished in 2013 and what we are planning for 2014. But first, a little background about the community.

After the last NetKernel developer conference in Fort Collins, Colorado, Chris Cormack and I decided it was time to build a foundation for a more formal NetKernel user community. Chris contributed his own MacMini and moved it to a co-location facility to host the forum (http://www.netkernelroc.org/nk4um/), a community Wiki (http://confluence.netkernelroc.org/), a community Apposite server hosting our NetKernel packages, and the Atlassian set of software tools.

During 2013 the level of interest in and projects using NetKernel increased significantly. While I cannot speak for everyone working with NetKernel, I can report on the NetKernelROC group activities.

We began the development of a new NetKernel portal (running on port 8088) to support features and functions that support the development, deployment, and management of NetKernel system. The first feature of this portal is a layered module management system with a Web UI, REST API, and a NKP accessible set of low level services. The module management system supports the publication of individual NetKernel modules for either production or development as an alternative to the Apposite package management system.

We implemented and published a plugin for SBT that supports NetKernel development. It supports the rapid creation of NetKernel modules, the creation and removal of IntelliJ project files, it communicates with the REST API of the module management software to put modules in development into a running local NetKernel instance, it reports on the status of the local NetKernel instance, and it creates an Apposite package for a project. The plugin leverages the Maven server system released by 1060 Research to access core NetKernel modules.

A NetKernel server was created and located at http://maven.netkernelroc.org to host a publicly available Maven server system.

We have refreshed our tools to include the latest set of Atlassian releases and we are using Stash (at http://stash.netkernelroc.org) for community projects. We are slowly migrating community projects hosted on a variety of systems to Stash and we anticipate having a steady stream of package publications through 2014.

Ansible (http://www.ansibleworks.com) support is being written to create, deploy, manage, and monitor NetKernel instances.

For 2014 we plan to solidify our NetKernel infrastructure. We need to move it off of Chris' computer and onto a new set of computers. The NetKernelROC community is focused on two things for 2014. The first is improving the developer experience with NetKernel. The second is developing a set of enterprise management tools to allow us to quickly develop, deploy, and manage large scale NetKernel systems.

Randy, Chris, Mircea, Radu, ...

Status Check / Data and Reality

The New Year has been flat out for me and consequently I've not had a lot of spare bandwidth to devote to newsletter content.

A week ago I gave an "Introduction to ROC" talk in Oslo which went very well, with several people actually asking for even more discussion of the philosophy behind ROC. (Imagine that, more philosophy!). Such was the interest, that I've been invited back to give a full day workshop which we're in the process of working out a suitable timeframe.

In another beer-fueled conversation, I was able to offer (and have understood) the most concise description of ROC possible...

Resources are abstract. An address space is a Hilbert-Space of potential states of resources. The operation of ROC is exactly equivalent to making a measurement of a quantum system and having it collapse to a measured quantum state (representation of the resource).

...but I definitely don't recommend you use this as your introductory pitch for ROC (although the beer might help).

On literally half a dozen occasions in just the last two weeks, I've been involved in discussions in which people have told me about the lamentable state of some organisation-or-others data and the rigid and fragile software stacks they've witnessed.

More than once I've had opportunity to explain how ROC's ability to accept tolerant representation of state means that ultimately NetKernel is a system that provides an impedance match between the operational state of an information system and the reality of the physical world for which that information state is required.

In short NetKernel works well, not for arbitrary technical reasons, but because it encompasses a deep fundamental approach that recognises the approximate relation between data (operational state) and reality. And that software solutions will always ultimately break if you don't introduce tolerances that allow for the fact that reality is a moving target.

Several times in order to validate this statement, I've referred to William Kent's 1978 masterpiece "Data and Reality" - a book I've mentioned here on and off over the years.

I last read the 1980 2nd edition in 2010, thirty years after publication, and was struck by how it is completely relevant today and provides a vital philosophical basis for comprehending the subtle evolving relationship between data (state) and the realworld.

Anyone working in software should have read it. Very very few have.

Since I've been referring to it in conversation I went looking for the old PDF eBook copy I bought and discovered it had been removed from sale (out of print). Instead I discovered that I'm not alone in admiring Kent's work - in 2012 a revised 3rd edition was released edited by Steve Hoberman. I just ordered my copy...

Data and Reality, by William Kent

I'm looking forward to re-reading it. As I recall, around about chapter three he makes a strong case for software systems of the future to be built in a way that will seem pretty familiar - something that sounds a lot like NetKernel...


Have a great weekend.

Comments

Please feel free to comment on the NetKernel Forum

Follow on Twitter:

@pjr1060 for day-to-day NK/ROC updates
@netkernel for announcements
@tab1060 for the hard-core stuff

To subscribe for news and alerts

Join the NetKernel Portal to get news, announcements and extra features.

NetKernel will ROC your world

Download now
NetKernel, ROC, Resource Oriented Computing are registered trademarks of 1060 Research


WiNK
© 2008-2011, 1060 Research Limited