The Sad State Of Programmers Part 1 : The Phone Interview

re: The Sad State Of Programmers Part 1 : The Phone Interview

Adam Machanic — Sun, 02 Dec 2007 22:48:25 GMT

Great post! Looking forward to more.

re: The Sad State Of Programmers Part 1 : The Phone Interview

Bart Czernicki — Mon, 03 Dec 2007 04:09:05 GMT

First, I hate these kind of articles. Because, while in spirit I agree with a lot of what you have to say...it has the I am "more well read than you" theme.

For example, the "truncate myth". First, a lot of people say this because the information on a lot of sites is INCORRECT. How can u hold that against someone? Basic understanding of the truncate vs. delete is enough for a jr. and even maybe a sr. developer (depending their primary role as a DBA or front-end developer). Unless your looking for an architect position ur NOT going to find someone that is going to be able to tell u answers to proper myths/answers especially now that someone has to be very well versed in a wide array of developer tools.

A PERFECT example materialized itself in this month's SQL Server magazine in the article by "Itzik Ben-Gan" who is a frequent author. Here is a direct quote from the article on read uncommited vs. commited:

"Most of this article's findings are fairly recent-in the past year or so--and my answers to the questions above before those findings were WRONG or INCOMPLETE."

---

A few years ago, I was asked on an interview by a similar personality "Does SQL Server support recursion". Well SQL Server 2000 doesn't support "true" recursion, but you have recursive triggers (limited by levels) and SQL Server 2005 does with CTEs (it was in beta then). The funny thing is the answer suprised them because they were expecting a hardy..."No".

re: The Sad State Of Programmers Part 1 : The Phone Interview

Denis Gobo — Mon, 03 Dec 2007 12:15:35 GMT

Bart,

I really though about it a long time before posting this. I did however try my best to make it as non acerbic as possible. I did not create a post which listed a bunch question with all the crazy answers I got.

I did not ask about rollback with truncate and don’t consider that as something that will influence my decision to hire this person. I did ask when can you not do a truncate, I expected the answer to be “when you have pk fk relationship”, only 2 or 3 people got that one right. I also did not ask about undocumented procs like the for each table proc (sp_MSforeachtable) or sp_who2.

Take the following question “How can you ensure that only values between 1 and 10 are allowed in a column?”

I got several answers

One person answered that all that should be done at the application level

Several people answered trigger

A couple of people had no answer

A handful of people answered constraint

The ultimate goal of this series is not to say look at all these n00bs but to try to provide some tips about the interview process itself. The 2nd and 3rd part will make that more clear.

Denis

re: The Sad State Of Programmers Part 1 : The Phone Interview

andyleonard — Mon, 03 Dec 2007 15:53:26 GMT

Denis,

I like the article and did not interpret your style as acerbic.

Since you mentioned recruiters, I'm curious how you attracted candidates for this position. Did you use recruiters?

:{> Andy

re: The Sad State Of Programmers Part 1 : The Phone Interview

mksql — Mon, 03 Dec 2007 16:31:52 GMT

Having been on both sides of that call, I try to follow the concept "It's not as much what you know, as what you can figure out." Phone interviews that test a candidates ability to recite Books Online are not terribly useful, but those that test _understanding_ and research skills are more telling.

Having multiple approaches to an example problem, or being able to articulate a research path to locate unknown answers is a skill I find is in the minority. I find plenty can precisely state single approaches as described in the documentation, but few can describe the path to take when the "common" approach fails.

I am not disagreeing with Denis... many initial interviews are about finding what on a resume is accurate, and what is "marketing".

re: The Sad State Of Programmers Part 1 : The Phone Interview

david burnett — Mon, 03 Dec 2007 18:54:35 GMT

In fairness to some of these people, I have had headhunters modify the resume I gave to them adding claims that I had experience that I did not have. Clearly, you don't deal with that headhunter again but you only find out they did this the hard way in the middle of the interview.

On Sybase at least through 11.9.2, you could not include Truncate in a transaction so you couldn't do a rollback. Perhaps, some of the people you interviewed came from the Sybase world.

re: The Sad State Of Programmers Part 1 : The Phone Interview

Denis Gobo — Mon, 03 Dec 2007 19:01:27 GMT

Andy,

yes most of them came through recruiters

re: The Sad State Of Programmers Part 1 : The Phone Interview

Hugo Kornelis — Mon, 03 Dec 2007 19:02:41 GMT

If I were ever hiring, I wouldn't mind people admitting ignorance. I'd much rather hear an honest "hey I don't know but if it ever comes up, I'l check BOL or use google to find out" than have someone make up an incorrect answer on the spot.

If a day passes by without me starting up Books Online, it's a day in which I didn;t work!

re: The Sad State Of Programmers Part 1 : The Phone Interview

Denis Gobo — Mon, 03 Dec 2007 19:04:46 GMT

David,

Most of the people had SQL Server experience since version 6/6.5

And like I said, I did not ask if you could rollback a truncate, they gave that additional info ;-(

re: The Sad State Of Programmers Part 1 : The Phone Interview

Denis Gobo — Mon, 03 Dec 2007 19:40:38 GMT

mksql,

during the face to face interview I actually have people write T_SQL on a whiteboard. A simple 3 column table with a primary key

re: The Sad State Of Programmers Part 1 : The Phone Interview

Denis Gobo — Mon, 03 Dec 2007 19:41:15 GMT

Hugo,

I know what you mean, every time I have to create a function I have to look it up

re: The Sad State Of Programmers Part 1 : The Phone Interview

Bart Czernicki — Mon, 03 Dec 2007 21:18:18 GMT

Denis,

Like I mentioned...I do agree with ur approach. My point was that there are sometimes people who ask interview questions like the "truncate myth" without understanding that information on the web/blogs has been wrong in a lot of places. The example I listed I read last week in SQL Server was a perfect example.

However, if you are taking the answers with a grain of salt that is fine. It is actually interesting because u can tell a lot as an interviewer/interviewee how difficult this person is going to be to work with depending on their answers or questions. I know I am always trying to avoid the "I am always right" personality

re: The Sad State Of Programmers Part 1 : The Phone Interview

dmarkle — Tue, 04 Dec 2007 12:27:10 GMT

Denis:

Your experience jibes with what I see when I do interviews. And I must say that you are right to call out people who believe in these myths -- someone who blindly reads something on a blog and without questioning or understanding it is not a good asset to have on a team, especially if that person takes their knowledge personally.

Why? Because sometimes it can be SO HARD to undo the damage that has been done by "experience". Sometimes, I think you should just hire recent college graduates and teach them yourself instead of hiring "experienced" DBAs. How many people do you interview honestly believe that every table must have an integer "ID" column as the PK? That's so much worse than the TRUNCATE myth for me, because the wrong answer can't be disproved in less that 30 seconds. And then, ask them why. "uh..., uh because it's faster".

And I do agree with Bart -- it can be rough when an interviewee asks a question about something like the TRUNCATE myth and doesn't follow up. That's why, when I'm interviewing, I make a point to interview the interviewer when I'm being quizzed. When someone asks me a question and checks my answer against something that's wrong, I need to know, because I am going to call them out on it. If they respond negatively, is that a place I really want to work?

Here's another question for you, to see if you have someone who's really into SQL Server: What does DBCC DBREPAIR do? (LOL)

Answer: Nothing, in Sql Server 2005. ;-)

re: The Sad State Of Programmers Part 1 : The Phone Interview

Denis Gobo — Tue, 04 Dec 2007 15:05:24 GMT

dmarkle ,

What does DBCC PINTABLE doe in SQL server 2005? :-)

I usually don't ask any DBCC commands since we usually developers only, I don't expect them to know DBCC commands or trace flags

re: The Sad State Of Programmers Part 1 : The Phone Interview

Adam Machanic — Tue, 04 Dec 2007 15:13:26 GMT

A long time ago I worked with an "architect" who demanded that every table have an IDENTITY primary key. Reason given: "if you don't, SQL Server will complain."

Personally, I've never seen SQL Server whine, moan, or mutter under its breath due to lack of an IDENTITY column, but apparently this guy had, at some point :-)

re: The Sad State Of Programmers Part 1 : The Phone Interview

Denis Gobo — Tue, 04 Dec 2007 15:18:20 GMT

>>"if you don't, SQL Server will complain."

yes..if you have duplicate rows you can't delete it from Enterprise Manager because it doesn't know which row to delete.....an identity will 'fix' that :-}

re: The Sad State Of Programmers Part 1 : The Phone Interview

Alexander Kuznetsov — Tue, 04 Dec 2007 16:14:55 GMT

Denis,

this is a very interesting post, thanks, but could you please not post any more interview questions? I found some of my favorite ones, and now I need to come up with something else.

re: The Sad State Of Programmers Part 1 : The Phone Interview

Kalen Delaney — Tue, 04 Dec 2007 19:44:14 GMT

Adam -- When parameter sniffing has my SQL Server use a NC index to access every row in a huge table, or when I have shrunk my log down to a couple of MB and then do a large SELECT INTO, the harddrive on my laptop makes a nasty sound that I always considered to be my SQL Server whining and complaining ...

(However, it never did that due to lack of an IDENTITY Primary Key.)

:-)

Kalen

re: The Sad State Of Programmers Part 1 : The Phone Interview

David Markle — Wed, 05 Dec 2007 03:37:09 GMT

Oh thank God Joe Celko hasn't discovered this site yet...

And Denis, the answer to your question is, of course, that DBCC PINTABLE puts your whole table in memory, which is what you'll need if you forgot to put an INT IDENTITY primary key on your table. But not even that will save you when the time comes to write your queries -- if you don't have an INT IDENTITY column on each table, it's going to be mighty hard to write your queries in the Query Designer -- where will the lines go??!

OK, enough of that. Now for a classic that I encountered 2 years ago. I worked with this DBA/developer that had > 12 years "experience", and I was looking at their code, only to see something like this in a proc called something like (ha) sp_InsertWhatever:

DECLARE @whateverID INT

SELECT @whateverID = MAX(whateverID) + 1

FROM whatever

INSERT whatever (whateverID, col1, col2 ...)

SELECT (@whateverID, col1, col2 ...)

Ah, and not even the common courtesy of using a transaction as a "reach around". I asked this True Champion why they didn't at least use an IDENTITY on the whateverID column:

"Because it's not Best Practice."

I love it. By the way Kalen, this person had your book. It was in fantastic condition!

re: The Sad State Of Programmers Part 1 : The Phone Interview

Denis Gobo — Wed, 05 Dec 2007 09:21:21 GMT

dmarkle,

DBCC PINTABLE on SQl Server 2005 does absolutely nothing, you can run it, nothing happens

re: The Sad State Of Programmers Part 1 : The Phone Interview

dmarkle — Wed, 05 Dec 2007 11:47:32 GMT

I was being sardonic. It was a long day at work.

re: The Sad State Of Programmers Part 1 : The Phone Interview

Vadivel — Wed, 05 Dec 2007 18:27:41 GMT

I second hugo's comment :)

Btw, last year I wrote couple of articles on 'Tips to prepare for interviews' .. at ur lesiure do check out and lemme know your thoughts http://vadivel.blogspot.com/2006/07/tips-to-prepare-for-interview.html and http://vadivel.blogspot.com/2006/10/tip-to-prepare-for-interview-part-ii.html

Regards

Vadivel

re: The Sad State Of Programmers Part 1 : The Phone Interview

Denis Gobo — Wed, 05 Dec 2007 18:44:28 GMT

vadivel,

In your 2nd article under In Person add

Dress in business attire

I will cover that in part 3

re: The Sad State Of Programmers Part 1 : The Phone Interview

Vadivel — Thu, 06 Dec 2007 04:01:21 GMT

Denis,

Yep thats what I meant by "2. Present yourself with formals.". May be i would change the phrase as u have pointed out. Thx.

re: The Sad State Of Programmers Part 1 : The Phone Interview

Paul Nielsen — Fri, 07 Dec 2007 19:53:43 GMT

David, Joe Celko has discovered this site, and even commented favorably about my post on Columnar Databases, whcih worried me a bit.

The Sad State Of Programmers Part 2 : The In Person Interview

Denis Gobo — Mon, 10 Dec 2007 20:01:57 GMT

This is part two of a three part series. Part one was about the phone interview , this part is about

re: The Sad State Of Programmers Part 1 : The Phone Interview

Geoff — Thu, 13 Dec 2007 15:05:14 GMT

Hey Denis,

I interviewed with you, primarily to shake off the interviewing rust. I can't speak for everyone, but asking you to repeat questions was a function of not quite being able to understand you, esp. over the phone. I wouldn't interpret it any other way than that.

re: The Sad State Of Programmers Part 1 : The Phone Interview

Denis Gobo — Thu, 13 Dec 2007 15:15:46 GMT

Geoff,

This didn't apply to you, there were several people who asked me to repeat the question and I could hear the pages flip in the background.

Denis

re: The Sad State Of Programmers Part 1 : The Phone Interview

Geoff — Thu, 13 Dec 2007 18:44:29 GMT

Gotcha. No offense was taken originally. I have a thick skin. But glad it worked out for you.

re: The Sad State Of Programmers Part 1 : The Phone Interview

Ben Amada — Sat, 22 Dec 2007 14:57:07 GMT

Denis,

Nice post. Regarding the "SQL Server 2005" experience, I think your expectations are a little unfair. I'm sure there are lots of developers whose first experience with SQL Server was with SQL 2005 or at least some of their more recent projects have been with SQL 2005. For these developers (specifically ones that never used SQL 2000), what do you expect them to put on their resume? SQL 2000? That would definitely not be accurate. Because these developers haven't used any of new functionality that came with SQL 2005, they can't list SQL Server 2005 on their resume? I can't think of anything that would need to be done in a database to support an application that has to be done using the new functionality introduced in SQL 2005.

This same argument can be applied to other technologies, such as MS Office. If a person has used Office 2000, 2002, 2003 and most recently Office 2007, but has never used a feature that was introduced after Office 2000, does that mean a person can only list Office 2000 on their resume?

Anyhow, just something I wanted to throw out there.

Thanks,

Ben

re: The Sad State Of Programmers Part 1 : The Phone Interview

Denis Gobo — Sat, 22 Dec 2007 15:12:26 GMT

Hi Ben

All of the people we interviewed had at least SQL 2000, most of them had also SQL 7, some of them had also SQL 6x and one or two had SQL 4X

When these people list SQL 2005 and then can't give me one new feature that they like besides the CLR then I find that a little strange.

If a person started with SQL 2005 then they would not have a minimum of 4 years experience in sql server, however if that person had other RDBMS experience I would definitely set up an interview.

>>If a person has used Office 2000, 2002, 2003 and most recently Office 2007, but has never used a feature that was introduced after Office 2000, does that mean a person can only list Office 2000 on their resume?

The problem is that the person can't answer the questions about the new versions and would look like he did not use all the newer Office versions. That said with Office 2007 it is pretty hard not to use the Ribbon

The Sad State Of Programmers Part 3 : General Tips

Denis Gobo — Thu, 27 Dec 2007 17:29:11 GMT

The Sad State Of Programmers Part 3 : General Tips This is the final part of this series. You can find

re: The Sad State Of Programmers Part 1 : The Phone Interview

Jon Nolan — Sun, 30 Dec 2007 17:49:42 GMT

>>One thing I never understood is the fact that it takes a person one minute to answer a question. You either know or don’t know the question.

I'm usually nervous during an interview, so I don't do very well in general, but in particular it almost always takes me a minute to answer questions, even when I know the answers.

re: The Sad State Of Programmers Part 1 : The Phone Interview

mos — Wed, 20 Feb 2008 23:25:08 GMT

I'm going to give a phone interview tomorrow to someone who will be doing some SQL and some C# work, and I stumbled upon this page looking for some SQL question tips. But having read it all, I'm wondering why people look down their noses at naming something "sp_Whatever" (I am not a SQL programmer, but our DBA does name things that way).

What did you do to master SQL?

Denis Gobo — Fri, 16 May 2008 16:23:17 GMT

After my The Sad State Of Programmers Part 1 : The Phone Interview series I got an email from a person.

re: The Sad State Of Programmers Part 1 : The Phone Interview

TC@Kin — Tue, 14 Oct 2008 08:38:59 GMT

Thanks for your post!

Actually I agreed with the opinion came from 'Bart Czernicki'.

Correct and complete is required.

If you are in hurry, ask to them stop.

re: The Sad State Of Programmers Part 1 : The Phone Interview

nasherooy — Tue, 25 May 2010 04:43:05 GMT

nice blog brother