Filtering out a PITA developer in an interview

In the last couple of years I had been unlucky to work with a developer that I really hated. He wasn't a bad developer. In fact, he made some decent input into the codebase and had come up with some very good ideas. However, he also had some traits that were absolute roadblocks for me.

• he was chatty: he loved to turn to me and distract me whenever a ghost of a question started to form in his head (I had to attend the birth pains of this question being phrased);
• he didn't know to hold personal distance: he preferred to grab attention by touching you on the shoulder while sitting behind -- not something I like, especially when I'm concentrated on something and not expecting this shit;
• he was slow: he talked slowly, he understood slowly, he browsed through the code fucking eternally, typing with one finger per hand and constantly slowly switching between mouse and keyboard; this slowness wasn't something you would notice instantly, but noticed very well after a 15-minute discussion if you wanted to get rid of him ASAP;
• he loved to argue for the sake of the argument: on the one hand this and on the other hand that and fuck it's all so dialectic that we should probably not do anything;
• he talked through every fucking smallest detail as if he was a novice instead of just discussing the main bigger parts and starting coding and then it will work out of course;
• in the process of these eternal (initially that was *days* until we figured out what a pain it was) discussions he would forget things from 10 minutes ago and you would need to revisit them;
• he loved to bitch about what a mess the existing (not written by him) code was; it's not that he wanted to offend, just I think because in his world there was smart him and dumb others, just as a principle, not you or me, but oh my god what a shitty method this is!..
•  it was normal for him to ask a colleague of mine to give him a lift and during the ride to comment that the car was too small or the speed too low and how he could've done that faster if just walked; again: not because he wanted to offend, but rather because he did not care and (this is my subjective opinion) was a smoking stinking piece of shit;
• he had salary larger than any of us, and at the same time he had no passion or enthusiasm for the project as well as no self-dependence, and the only times he e.g. overworked was when he figured out he would get yet more money this way.

I could continue this list eternally. Just to depict the level of my frustration from working with him I'll say that his presence made an otherwise perfect and thrilling startup project almost intolerable for me, and that working with him was the worst experience of 2012 for me, and that was a year when we e.g. split up with my girlfriend, so that should put it into some perspective.

Lyrics aside, I was often thinking if it could've been different -- if he could've somehow be filtered out by the interviews. I'm probably selfish here, but on the other hand what had his presence done if not destroyed the atmosphere in the group? Also, I saw him constantly doing things slower than they could've been, and when we later went to different projects in the same company, he failed to deliver on time by a factor of 10 maybe (weeks instead of days and none was ready when he left for a vacation) -- so overall I'd say that I would pretty much insist on not having such an expensive while at the same time useless and toxic developer in my group.

At the same time, I think to the interviewers he must have looked pretty cool: aged 40-something -- but we have some very good developers of this age in our company, some of the interviewers included; with 5+ experience in Java and 5+ more in C++ and so on; intelligently looking, calm (as a turtle), communicative... 

First, if there was a real task to be done (at home, on the scale of a day e.g.), he would certainly have failed it. Second, a possible test for him would be to give him some well-enough-written code and make sure he would 1) understand what it does, 2) reasonably fast, 3) mostly on his own and without constantly demanding clarifications from interviewers. Then to ask him to do a change -- ensuring the specification is really clear -- and make sure he 1) does it, 2) doesn't drag interviewers into helping him or constantly reassuring him he's doing ok, 3) doesn't pretend his own block is caused by others (bad specification, insufficient tooling etc). Another option would be to give some ok code and ask how bad he thought it was, ask to improve (in reasonable time), then ask how good he thought his code was. Another option: give a complicated task and a tool (library e.g.) sufficient to solve it, then see how fast he would do the task and how much he would blame the tool for delays. Or maybe a not so hard task and a not so sophisticated tool with a bug or two -- and check if he will fix the bugs or blame the tool.

And then -- then there should be some mechanisms to not have people stuck in teams they are unhappy with for years. And for me the lesson is: trying to play nicely is not worth the effort while there are so many tasks ready to solve and developers ready to help. Sometimes a coworker is a pain in the ass because he really is, not because your attitude is wrong.

1. Thought we had a problem in production (serious but not urgent).
2. Decided to be fixing in the production branch (although envisioned it would take long which could have been seen as a clear warning).
3. Changed it substantially (redeploying to a copy of production).
4. Then the REAL, urgent problem came. And we had out production branch un-deployable to the production.

Should have branched production and worked in the branch from the start.

Code is a liability

My attitude towards lines of code as a metric has changed drastically during the last several months.

We all know LOC is a meaningless measure for productivity estimation (and here I regard as LOC any simple counting metric on the codebase: it can be "words", "whitespaces", "keywords" etc.). Well, what this metric can show is that you can e.g. sit down and churn out 5000 LOC per evening. The amount of LOC for a given task can differ across languages, so let's consider that LOC is already "discounted" for a language. Still, use of this metric for estimating productivity or project progress is questionable.

One question is that even "optimal" LOC differ wildly across tasks. Adding a cool new feature can be 1 LOC, and refactoring for the sake of beauty can be 10 kLOC. Tests "write themselves", so tests (if you include them into measurement) can add times more than the functionality they test.

But there's more to it.

Adding a feature by using a third-party tool and several lines of code is faster, more reliable and overall "better" than writing it all yourself (which is more LOC). Basing your product on some technology that allows you to make your LOC 10% compared to what it would be without using it is better in the same sense, too. It gives you competitive advantage. I feel that most startups are about it: having an idea that is useful for people *and* for which a technology stack exists allowing it to be implemented with few LOCs and few configs.

It doesn't only apply to the "start". For a startup (and IMO to any product) it's important to maintain the product focused by only adding the necessary features (to fight feature creep and scope creep) which should be as coherent technologically as possible. Common parts should be reused. Codebase kept small.

I started to understant the "functionality is an asset, code is a liability" mantra (the linked post gives cool examples of what I've been trying to formulate abstractly).

Is {0 LOC, 0 added value} better than {N LOC, much added value}? I claim it is if the added value isn't that aligned with the product and worse otherwise, so "0 LOC" is incomparable to "N LOC" in and of itself.

Then why do some people, me previously (I hope!) included, tend to view the "amount of change" as a metric where more is better? Reflecting over my own beliefs, I think there can be at least two explanations. First, I tend to think that there's some "inherent" to a problem amount of work to be done. Seeing big LOC change gives me an idea that something important was done. It's probably not the case, or even sometimes this "something important" can be an important overall degradation. Second, I simply believe that change is always good, and that no LOC means no change.

These days I think neither of the two is the case.

right/effective

As Amy Hoy would put it:

Do you want to be right or do you want to be effective.

History is full of failed projects that where “right”. So it will do you no good to be righteous about testing. It’s 25 anniversary of OS/2 so that comes to mind. Don’t forget other like Multics or micro kernel based kernel GNU Hurd.

I’ll let you in on a secret. They were not really better. They only seemed better on the first glance. Kinda like Titanic seemed unsinkable.

from here

software project complexity

You are always wrong about the complexity of the project.

You overestimate it, and have been from the beginning. I'm not speaking of the resources needed for implementation, but rather about complexity for understanding (and almost always more resources as a consequence).

It's only that you don't know or have the right tools and you have much more complex understanding of the project. Your design is overengineered. You build a framework for representing graphs etc when a 20-line algorithm would do. Some of the tasks you are solving are no tasks actually with the right toolbox.

If you knew its actual complexity, you'd laugh at how much time and effort is spent at the project, and also how many people work on it. You think — what? Due to some lack of time the way you do is maybe 20% more complex than it would be if only you had time for analysis? Bullshit. It's 10 or 100 times more complex. The problem is other people have an even worse idea.

Following K.I.S.S might be right: there are less variants among simple implementations, and yours is one of them.

Thoughts

Today I've found a sheet of paper where I've written some of my conlusions several years ago after implementing a relatively large subsystem of the system I was working on in a large company.

1. To learn you have to make errors. You have to praise error rather than be afraid of it. "If you speak less and less, you'll have to speak more and more" — Cortasar wrote in his "Exam" meaning that not speaking makes the language and your command of it and expressiveness theirof degrade, and losing its expressiveness it makes you build larger texts. So, the same applies to errors: if you do less of them, you'll have to do more later.
2. You shouldn't be too clever in trying to overcome a difficulty: a much more efficient way is trying to avoid it altogether. So, you need to think about the nature, the source of the problem. Perhaps *it* can be removed. Being too clever in overcoming problems, seeing at once a (long) path of fixing a problem can obscure a (short) path that bypasses it altogether.
3. Relying on data format (think XML) in many different places weakens ability to change (and to support): changes in the format will lead to changing all places of using.
4. Premature optimization is the root of all evil (see (2)).
5. Data that differ slightly (say, two similar formats) are dangerous. You have to repeat yourself or at least are very urged to (things differ). Alternatively you can work out rules of conversion, but that conversion will most probably break when some change is introduced. Say, we have one source of data and all just works, we don't even think that there's some part of the code that builds another format upon it. We change the source format substantially and things suddenly break.

Corollary

Oh, and one simple thing my ruminations lead to is: the prototype must be written by the same people who will write the actual application, or else there's no point in it: the actual developers won't learn anything from it.