Interview with Rob Farley

After the reviews last week about Rob Farley’s chapter in the Deep Dives book, we talked over email a bit. I had the chance to sit down and ask him a few questions that Jeremy and I came up with. I hope that you enjoy them. What amazes me the most is the fact that Rob did not start as a database guy, not only did he not start as a database guy, but actually chose to pass up the opportunity to take a database class. Another item that really grabbed me was how Rob hires consultants that he has, see the product is growing so large that few know everything about SQL Server to the skill of being able to learn new things is just as positive and know the basics.

How did you get started with SQL Server? What is the draw to SQL Server, why do you like the product so much?

Funnily enough, I skipped the databases subject at university. People said it was easy marks, but I was more interested in other things. The things that fascinated me in those days were Artificial Intelligence, Natural Language Processing, things like that. And I liked to think I didn’t need easy marks. During my Honor’s year, I was offered a job in a Melbourne consultancy, and they said I’d need to know about databases, giving me a book to read about Oracle’s SQLPlus. It just clicked, with many of the principles relating to stuff I’d learned in my Pure Maths half-major – Set Theory, Number Theory, Logic.


So my first databases were Oracle, but I also had clients who used SQL Server. That was 6.0 back then. I was one of those consultants who seemed suited to being spread across many clients, sometimes working for six different clients in single day. Always some Oracle, but an increasing number of SQL Server ones as versions 6.5 and 7.0 were increasing in popularity in the late 90s.


All the while I was also writing code in iterative languages such as VB (versions 3 through 6) and C++, and didn’t consider myself a database guy. Everyone in the company did databases as part of what they did, and I didn’t appreciate it as a discipline of its own.


Years later, after having my career saw me venture into management and a continued generalization, I was still finding myself as the ‘SQL Expert’ in the company, and decided to specialize. So these days I do very little iterative programming, and spend almost all my time in SQL Server. I am very comfortable with the set-based paradigm, and I find this is probably the thing which continues to interest me – that difference between other IT people and the database guy.


Why did you pick the chapter or chapters that you wrote, what made that topic your topic of choice?

T-SQL has always had a special place for me. In consultancies of developers, it was my ability to see effective T-SQL solutions that differentiated me from my colleagues. I have written courseware on T-SQL, and consistently find that the gap between ‘competent’ and ‘expert’ is really large, but can often be addressed by an appreciation of the paradigm and a basic understanding of how to influence the decisions that the Query Optimizer can make.


So I wrote chapter seven about two main things.


One was a query that I came across that I suspected few people would really understand. It’s the main query from one of the Project Server Report Pack reports. I decided to describe a method for reading the FROM clause that would work regardless of how the query had been rearranged. After all, at some point we all get asked to fix a nasty query, and step one must be understanding it.


The other was about the fact that the Query Optimizer can ignore joins that aren’t used. This gives database professionals the opportunity to create views to encapsulate useful logic, while being confident that if an application doesn’t ask for all the columns, the QO will simplify it down a lot.


There are a bunch of other topics that I could’ve picked, but this seemed to make sense for one.


The other chapter I wrote was chapter forty, about indexes. In fact, in some ways it’s all the same topic, as all queries are really against indexes. I know we write them against tables, but really there’s no such thing as a table – they’re all just indexes (or heaps). If people had to write indexes instead of tables in their queries, the level of understanding would be very different.


But indexes aren’t just there for seeking and scanning. They also help the Query Optimizer decide how to run your query. Indexes can be used to constrain the system, so it’s easy to find ones that are “unused” according to the metadata about them (eg, sys.dm_db_index_usage_stats), but that still have a large influence over the performance of queries.



With respect to the quantity of new features being added into the SQL Server stack, have you found that the traditional DBA role is evolving in Australia?

Also, are you seeing new job titles based off that and are companies hiring managers learning about the new responsibilities and tasks with the broader stack that MSFT is delivering?

Certainly the Business Intelligence space has changed things, and is continuing to shift. A database professional is no longer just a DBA, responsible for backups and HA systems. I still find that few people consider themselves dedicated Database Developers, but the Business Intelligence Developer seems to be becoming a recognized role. I do wish that the industry recognized the need for dedicated database specialists much more than it does, as I keep finding that traditional programmers typically take the wrong approach to data. It gets more complex with things like LINQ, Hibernate, and other technologies, and even SQL features like StreamInsight, which I consider to be more of a .Net feature than SQL. The skills that a SQL expert has don’t automatically translate into BI, let alone some of these other features, which is making the SQL space far more complex than ever before. It’s a good space to be in, even if the days of being an expert across the whole platform are almost over.


I run a consultancy now, and in hiring people I need to look at their background, ability to learn, and where I can see them fitting into the overall platform. I can’t just hire a SQL guy any longer. I prefer to pick people based on their integrity more than any other feature though. I can always teach them how to think in a good way for MDX, or relational design patterns – integrity is the thing I want as a prerequisite.


Do you think there is a big culture difference between the US and AUS in the tech industry?

I actually don’t know what the US industry is like, but there are definitely differences between the UK and Australia. I grew up in the UK, and moved to Australia (Melbourne) as a teenager. I went back to London for a while, but returned to Australia (Adelaide) nearly eight years ago. I can assure you there are differences between Melbourne and Adelaide, let alone between Australia and the UK, so I’m sure there are big differences between here and the US.


What are your thoughts on technology debt in today’s rapidly changing environment with MSFT releasing major versions of SQL Server every 3 years now?


I think having regular releases is good, but of course people are so reliant on their database systems that few would mind if the data platform had less releases with a stronger stories around upgrading. Upgrading systems is still a major undertaking, and I think people often prefer to stay on older versions rather than risk their system for the sake of some additional functionality. Any time someone regrets an upgrade, the effort in improving the experience becomes that much harder for vendors.


I hope the people who are currently moving to SQL 2008 and 2008 R2 find it a positive experience, and that when version 11 comes around, the uptake is both larger and sooner.



Rob, it was great to talk with you, and I hope we have more opportunities to do more in the future.


0 Responses to “Interview with Rob Farley”

  1. Leave a Comment

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )


Connecting to %s


Enter your email address to subscribe to this blog and receive notifications of new posts by email.

Join 11 other followers