Monthly Archives: March 2014

How Dare You Change the Operational Parameters

This week has been full of annoyances. Changes in operational behaviour which have presented themselves as weaknesses in automation and tooling. I’m not talking about things we were too lazy to think about during design and development, I’m talking small paradigm shifts: 8-fold increases in data (mostly noise); new data combinations; and repeated execution of a human initiated process more than once a minute (for the last two years once a week).

We were actually quite prepared for the data increase. Our original design allows for scaling out. However, part of the data source is temporally sensitive and out of sequence handling is not an option.

We’ll get over these once we understand the reasons for the operational changes and what the new requirements are.

In a rash moment I thought it would be a good opportunity to totally reimplement one of the applications. Bad idea. The application is used by Project Managers, touches business partners (represented by our Supply Chain group) and involves our Accounts Team.

It turns out that not only do all these people have an opinion on the operational requirements for the application but also how it should be implemented.

I’m pretty sure I don’t mind producing a new application but I have no patience for the endless debate and lack of a view towards a common good. People are spoilt by software these days. They all want their own workload reduced to zero. In a world of tight budgets this is difficult to do.

Agile allows you to manage backlogs but they become endless. Every feature is essential.

We’ll eat the elephant one bit at a time and who knows perhaps end up with more than a series of compromises.

Bring back the single feature command line application.

Linking Data in the Meditteranean

I am just back from the European Data Forum and the Linked Data for Language Technologies Workshop. The co-location of these two events meant that many of the leaders in linked data and the digital representation of linguistic and knowledge concepts were in one place.

The presentation that stood out for me at the EDF was the second day keynote by Ralf-Peter Schaefer of TomTom. Seeing how they use their 9 trillion and counting data points to find patterns in, and make predictions for, traffic conditions was very interesting.

The LD4LT Workshop was very productive despite the virtually non-existent free, and totally non-existent pay-as-you-go, wi-fi connectivity for the whole duration of the two events. It definitely brought home to me the importance of connectivity these days. I convinced myself to leave my mi-fi at home – will not do so in future.

Presentations I took note of during LD4LT were about webLyzard and Rozeta.

I presented my three industry challenges and my hopes for how linguistically motivated linked data might help me solve them.

I used the travel time to learn about the AngularJS project. Pretty impressive stuff. I particularly liked the way the framework handles the binding and referencing of data within the rendered HTML UI without explicitly needing to use data-* attributes to store object id’s.

I was hoping to be able to document trials I have been doing with DITA and XLIFF+ITS using the XLIFF-DITA Roundtrip toolkit but I got stuck at the last hurdle and being able to figure out the issue would require my having a deeper understanding of the toolkit.

We finished our assessment of supporting XLIFF 2.0 in Ocelot. It currently looks as though we will introduce a layer of abstract data model façade objects between the Okapi XLIFF filter classes and Ocelot.

I must dedicate some time to my H2020 proposal.

I compensated for my lack of my regular 2 kilometre daily walk with a day long urban hike around Athens taking in the Acropolis (naturally), it’s museum (I’d like to point out that Thomas Bruce, 7th Earl of Elgin was Scottish) and the chapel of St. George at the summit of Lycabettus Hill which offers truly stunning 360 degree views over Athens as far as the Aegean Sea. Unfortunately I didn’t have time to visit Kastella Hill near Pireaus.

2014 Ramp Up

The beginning of 2014 has been a hive of activity: San Francisco, Boston, Luxembourg, team re-structure, and new research interests.

Until now a hectic travel schedule has meant generally more input (reading reports, e-learning videos, and thinking through ideas) than physical output. However, with the increasing prevalence of in-flight wifi and Google QuickOffice for iPad, I have been able to work more in real-time despite being on the move.

With the extended travel and inevitably longer periods of being sat on my backside, I wanted to get sorted with a Sit Stand Desk. This is actually quite difficult if you want to avoid spending several hundred Euro and not have to get Allen Keys and screwdrivers out. My solution is this:

Budget Sit Stand Table

Seriously, I’m delighted with it. It is portable, easily stored, is stable, spacious (450 x 1,350mm) and fully adjustable between 610-1,020mm. Total cost including purchase, transport and installation: €130.

One benefit of travel is of course meeting friends and business associates. San Francisco was the chance to socialize with our friends and development partners, Spartan Software.

Spartan Dinner

Kevin Lew, Yan Yu, Scott Schwalbach, Chase Tingley, Paul Magee, Chris Pimlott, Me.

Luxembourg was an opportunity: meeting many of the Multilingual Web LT Working Group. I hope to have an opportunity to work with these people in the future.

mlw-lt-final-review-team

Pedro Díez Orzas, Milan Karasek, Me, Stephan Walter, Jirka Kosek, Felix Sasaki, Dave Lewis, David Filip, Tadej Stajner.