Recruiting the Top 1 Percent - Joel Spolsky | Hacker News Reader

Recruiting the Top 1 Percent - Joel Spolsky

inc.com

16 pointsdanw17 comments

I don't know much about Mr. Spolsky or Fog Creek, so perhaps I am entirely incorrect. I understand he has a lot of respect in the software community, but every time he writes about business it reeks of charlatanism. He seems to do well enough; but, he hasn't quite built an empire.

Being kind to your employees and your customers is a laudable idea, but it is not a business plan.

You also don't need world class intellect to build the software he is selling. This is fortunate, because I doubt his $75,000 could buy the soul of someone great who wanted to work on something important.

That 'world class intellect' gets stuff done so much more efficiently that it works out cheaper than several weaker coders.

How do you get "world class intellect" to work on a proprietary bug tracking software, though? I think it is written in ASP, too? (Could be mistaken, though).

"FogBugz is written in Wasabi, a very advanced, functional-programming dialect of Basic with closures and lambdas and Rails-like active records that can be compiled down to VBScript, JavaScript, PHP4 or PHP5. Wasabi is a private, in-house language written by one of our best developers that is optimized specifically for developing FogBugz; the Wasabi compiler itself is written in C#." http://www.joelonsoftware.com/items/2006/09/01.html

Writing your main cash cow in a private, in-house language seems like one of the riskier decisions you can make.

If they would have at least open sourced it and tried to get some sort of community around it it would have at least start to become less dangerous as the community grows.

But programming your main (only?) product in a language nobody except your current employees knows? Employee turnover must be even more of a bitch over there then it is in most companies.

I wouldn't want to work there when this "one of our best developers" quits/dies/retires

It depends.

I've got an impression that Wasabi is highly-specialized domain-specific language (just a wild guess from his article). If it is a general-purpose language it's worth to build a community and get feedback, but if it's a niche DSL, it's less likely that you get a growable community and it's more likely that feedback tends to be feature requests that aren't necessary for the original purpose of DSL.

And if it is designed well, the risk of the original developper's quitting is just about the same as having in-house libraries or workflow, I think.

Being kind to your employees should really just be obvious. Unfortunately it is not. Many people with rank in a company treat others terribly, simply because they can. Probably they don't know any better, and haven't considered the downsides. Joel does a really good job of packaging ideas from books like Peopleware and Mythical Man-Month in a way average people will respond to. No matter what his software company does, that's valuable service.

I don't know all that much about bug tracking software, but at my last company the trouble ticketing system they used was a monstrosity (Peregrine Service Center). Everyone hated it. It had a sluggish java applet client. It was an enormous resource hog, it was buggy, and had a horrid interface. It took forever to log in and logged you out after 15 minutes of inactivity (taking over your desktop to inform you of that fact). I'd guess this load of crap was costing the company a fortune compared to a shrink-wrap solution like FogBugz.

Here's a picture of the interface: http://www.goladus.com/_.jpg

This is a rehash of one of his earlier articles. He got rid of one of the points of criticism -- complaining about no good resumes, while hiring people who wouldn't have passed his resume criteria being that he goes after them while they're still in school. Catch-22.

Resumes are a horrible way to screen people.

Also, these days the top 1% are probably doing start-ups, or if they're employees, it's for Google or someplace exciting -- not hacking VB code in New York.

"Then I pose a software development challenge. For example, how would you implement a Web-based clone of PowerPoint? This gives me a feel for how smart they are, and if they know the basics of software development."

A trick question, of course. The correct answer is "you don't, since Google will simply release their own which will make yours irrelevant."

If you need a job to motivate you and be told what to do and how to do it, you're not in the top 1%. I think this is a way to pander to his future employees. :)

Isn't the top one percent kind of a large number? Maybe you want the top quarter of one percent? I mean, the majority of software developers probably work in IT doing zero innovation, not in a company that creates software that pushes boundaries.

I definitely think the number of developers who are creative, passionate, smart, fast, adaptive, and good is less than 1% of all software developers. Maybe that's accounted for because he's ignoring those who start their own startups, as they are too good for his 1% formula?

"Isn't the top one percent kind of a large number?"

It is. They want less than the bottom tenth of that. This is what would happen if JS were to interview anyone further to the right on the Bell curve.

JS: How would you fit in here?

Candidate: I would simulate the internalization of the values of the corporate workplace.

JS: When can you start?

"A limousine meets them at the airport..." etc etc - I don't know, isn't that cheating a bit? It's not like people will have the same grand life after being employed by Fog Creek?

Then again, that's probably just life (any kind of deals, including dating). But I would almost be a suspicious - if he is trying to sell the job this hard, there might be something fishy about it? Of course Joel is not just somebody, I am sure it would be very interesting to work with him, but I mean, if it wasn't him...

Sure, it's cheating, if you think that life is basically fair. But it's not. Making a good first impression is the smart thing to do, even if you stretch a bit beyond "everyday life." Just so long as you don't actually lie.

Maybe it just boils down to the benefits of advertising: we see the advertiser can afford the advertising, so he can't be a total loser.

That's a good way of putting it!

Even grad schools wined and dined me at the interview stage, and one school even sprang for the limo. I can guarantee no grad student expects to live like that, except of course when the next year's prospective students are interviewing. There's no implied promise that this is a typical day; it's simply what's done when you want the best.

"think about what happens to the 99 people you turn away. They don't give up and go into plumbing. They apply for another job."

I want to know if that applies to YC. Specifically, how many applicants keep reapplying at every round?

People keep saying that there are "hundreds of great teams" and that your chances of being chosen are correspondingly low, but it might not be really true.

Reply on news.ycombinator.com