Allan Wrobel is a senior software engineer at LexisNexis Risk Solutions and a long-time ECL user. Initially working with data operations, Allan now serves as an ECL developer on both Thor and ROXIE.
Allan has worked at LexisNexis since 2011 and was part of the original inception of the company in the U.K. With more than 40 years of experience in the industry, he has become a passionate ambassador for the HPCC Systems community and has contributed several video tutorials on YouTube for our users.
Flavio Villanustre, our VP of technology, sat down with Allan to discuss how he’s seen the industry change since he began working with databases in the mid-1980s, his ECL tutorial YouTube series and where how he sees our platform evolving in the next ten years.
With more than 40 years of experience in the technology industry, you’ve seen an incredible amount of change in the field. If you could go back and talk to your younger self, what advancements would you be most excited to tell him about?
Interestingly, as some of the technology goes, a lot of the fundamentals that we see nowadays in our modern world were already in place. For example, networks were around, but they were seen as point to point. You had precursors to SGML and HTML, such as IBM's generalized markup, which was all in place.
What's really interesting, I find looking back, was that the individual components for the technology explosion were in place and the genius of it all was the virtuous positive feedback cycle. There were all these technologies coming together in a way that no one imagined before. It's given me a job for life and a very interesting one at that. If I was talking to myself 40 years ago, I'd say, "Stay in this industry because it's so interesting.” It's been a roller coaster ride.
When did you first become involved with HPCC Systems, and what were your early impressions of it?
I have always worked with databases. In fact, I started out in a two-man firm where we had our own proprietary database. After that business ran its course, I moved over to HPCC Systems and LexisNexis, which felt like the natural next step from my previous firm.
What interested me in HPCC Systems was the vast size of it. It was one huge data churning industry that was just, in terms of magnitude, completely beyond my experience. Working at my previous firm, I felt like I had been tinkering with my old 1955 Ford Thunderbolt and then suddenly gained access to the Ferrari F1 team where I had full access to all of this expertise. It has been a pleasure working with people like Jo Prichard and other data experts, who I’ve learned so much from during my time at the company.
You’ve created a great series of YouTube videos on what you describe as the “funny little quirks of ECL.” How would you compare ECL to other programming languages? What makes ECL different from others?
Firstly, I love ECL. For me, it has just the perfect level of abstraction. I find myself thinking about the business problem and not the technology, which is quite a relief. This isn’t true in many other languages. ECL does have a high learning curve, but once you understand it, it becomes incredibly simple and just flows from your fingertips. I've never come across a problem with ECL that I haven't been able to fix intuitively.
I did find, however, that there were a few quirks that I had been struggling with and figured that if I was having trouble with them, I'm sure other people have in the past as well. For example, in HPCC Systems, with a concept of a builder window, you have actions and a file set of attributes and it's not that clear sometimes. This is especially true when they look the same in the IDE. It is hard to tell the difference between a builder window and a file of attributes.
That is what prompted me to create the YouTube videos – to help others.
If you could create a perfect programming language, what would it look like?
I think you'll be surprised by my answer, because one of my pet peeves is all of this proliferation of languages. You have Perl, Ruby and then Ruby on Rails, etc. There doesn't seem to be much justification to this proliferation of languages.
My perfect programming language would be English. Think about it. Businesses come up with a possible product, and they write it out in detail so it can be implemented by tech people such as myself. They have to get their mindset into that of a computer programmer and then that technologist has to implement it into the system. In reality, IT people such as myself are just middlemen. Any big business professional should be able to draw up their idea in a Word document.
There's enough faculty and preciseness in English to personally define any business problem you want. The problem set is too broad for any one language to handle and there is no such thing as a perfect language, but the best language would be English.
Where do you hope to see HPCC Systems in the next ten years?
I feel that querying is going to always become more complex. People and businesses are mining data by the petabytes and they’re going to ask increasingly complex questions about their data.
We had people on the databases I was working on decades ago who liked the fact that they could use “milk” as a search term, and it would retrieve documents on brucellosis, a disease that is caused by bacteria found in unpasteurized milk. The word “milk” didn’t appear anywhere in the documents, but the database would still retrieve information related to Brucellosis when you searched on milk. That impressed people. Creating that kind of powerful querying is going to be a trend in future. That's the race to win.
Want to hear more from Allan’s interview with Flavio Villanustre, VP of Technology, LexisNexis Risk Solutions? Listen to the webcast