Category: Learning

Jessica Kerr explains why software from large firms is so often terrible:

Software is hard to get right. And every time we don’t, customers leave.

Appointment scheduling that sends a calendar invitation “Join the Zoom” without a link. Checkout screens that delete my credit card number when I change the shipping address. Complete Order that comes back with “Please try again later.” Items can’t all be shipped, can’t all be picked up, and this is a maze to figure out (also Lowe’s). A shopping cart that pops up a generic error modal when any single call to the server fails.

Read the whole thing. A lot of this sounds like an incentive alignment problem: each sub-group within a large firm optimizes for its own benefits, but the sum total of those choices leads to a sub-optimal result for the firm itself, as in the case of bad software driving customers to Amazon.

Joey D’Antoni shares some sample job interview questions for hiring a senior DBA:

In my role as a consultant, it’s rare that I go on an interview anymore, though I occasionally get interviewed by a client, or interview potential DBAs for clients as part of my role. There are a number of these lists on the internet, but many of them are old and focus on trivia questions (there won’t be any questions on what’s clustered versus nonclustered index, but if you are interviewing for a senior role, you should know that anyway. I also like to focus on open ended questions, to gauge the depth of knowledge of the person I’m interviewing.

I will say that when hiring on the database engineer (i.e., development) side, the questions I love best aren’t trick questions; they’re experiential questions. For example, “Here is a common type of problem we need to solve. What would you do in this scenario?” And then we can dive in. As a quick example of one, “You’ve just taken over ownership of a database where most of the clustered index keys are uniqueidentifiers. Would you consider GUIDs a good clustered index candidate?” Then we can talk about yes or no, what makes for a good clustered index, and how you might go about changing it. Oh, and I’ll admit that my thought on this question has changed since Jeff Moden’s outstanding presentation on the topic.

Cole Nussbaumer Knaflic explains how to get feedback:

We recently kicked off a new 10-week course, which has been really fun to develop, because it’s both longer than our typical workshops and spread out over a greater amount time. Combining these aspects means that we get to cover more topics related to data storytelling and go into greater depth on each. We kicked things off with a focus on feedback, due to the important role this will play throughout the course, and the critical role it plays in our skill development and efforts to communicate effectively with data in general.

There’s some good advice in here.

Taiob Ali shares some commonly mistaken terms:

A list of items that people often get wrong. If you have a suggestion, let me know, and I will add them to this page.

On-premises and braces are the two I hear people get wrong the most. I had been afraid that the recognition of “braces” as the proper term had been so lost that I was like a monk in 9th century Ireland transcribing illuminations of worn tomes here.

Joe Celko offers up some thoughts on interpolation:

Interpolation is a mathematical technique which was popular before we had a lot of cheap computing power. The basic idea is that if you’re given a set of data and looking for a value in the same range, you can interpolate it to get a reasonable estimation for the value that is not actually in the set.

If you can find an old calculus, finance, statistics or algebra book, they had lookup tables in the back. Remember that the only computational tools students had back then were pencil and paper or a slide ruler. If you wanted to use a pencil and paper, you had to know what formula to use. If you use the slide ruler, you can only have three decimal places in your answer (yes, there were a couple of over-sized specialized slide rulers which could go as high as four or five decimal places. They were very expensive). But if your slide ruler didn’t have a particular function you were trying to compute, it was hard to get even the three decimal places.

When you try to approximate a value outside the range of your set, that’s called extrapolation. It’s a different topic and requires a slight leap of faith.

Interpolation is a key part of regression techniques. Read the whole thing.