I definitely agree with you and like I said recently I think its really hard to interview for. I think I handle this in one of two ways in an interview:

1) If I'm hiring for a specific position where I know what skills are going to be needed I'll probably ask about a handful of specific features I'm expecting the future employee will need to use. If the candidate claims knowledge about any of the features I'll probably ask for specific code for one or two of them. Ex. Have you ever used triggers? Yes? Ok, how would I go about adding a trigger...

2) If I'm hiring a general-purpose developer (much more common) who just claimed to know SQL and I want to check on that then I ask about some of the things they've done with SQL. What features have they used? What's the thing they like best/worst? What's the most interesting project where they've used SQL? What challenges did they have? And so on...

Thanks.

Ya, I could stand to learn some more nuances about databases, but I am very proficient with a super fleshed out ORM, so I don't feel totally out of place. I think I'm decent at database design, but there's certainly stuff to learn there too.

I'll likely never interview for a big company so long as I live... wouldn't succeed in that environment.

If you like ORMs, you ought to check out Sequel -- it poos all over ActiveRecord

I should probably finish that course. I went through about 25% of it. I mainly wanted to learn more about joins. It was good.

Sometimes you need to step outside of the ORM. For example what do you do if you want to group up a list of things by a date interval? You would have to drop into raw SQL to do this because doing it with ruby would probably be slow unless you had a very small data set.

A good ORM like Sequel will translate your request into the proper underlying SQL code. It's pretty smart about it though not 100%, and for exceptional cases you can usually figure out how to make generate the SQL you want, although in this case you have to understand the SQL to begin with, taking away at least some of the benefit (though you still get things like portability and code-base consistency).

Unfortunately, the version I took didn't get into speed optimization too much. Sure, there's indexes and normalization, both major, but that's a small part of it all.
I think getting speed just takes opening up the admin and timing them down. It's quite incredible to see a 5 minute query turn into a 1/2 second query by simply rearranging the query or trying a different join.

I think the class focuses more on the algebra. Optimization is more like working with sets.

Do those of you who work at bigger companies have people dedicated strictly to user experience?

I'm working with an amazingly talented designer, who does a great job coding out the html/css as well. It's really nice, and we work nicely together. The problem is, he think it is his job to do all of the UX stuff as well... and he's not that good at it. Like, he gets defensive when I suggest stuff (6 person company) that would make the site easier to navigate, and whatnot.

It seems like UI and UX should be strictly different disciplines, and in the case where you don't have a UX person/team, it should be a discussion. Our relationship isn't like strained or anything, but I just think there's a misconception here that designer = UX person.

Our site looks great, but it's definitely hard to navigate in its current state.

They should be, and it should be.

And in general, I'd trust someone who imo was a great programmer to make good UX decisions more than I would trust a great graphic designer, although that point is debatable.

http://thehipperelement.com/post/479...rong-job-title

and also:

http://www.amazon.com/The-Inmates-Ar.../dp/0672326140

The question I was asked during my internship interview went like this:

Do you know what a trigger is?

Have you tried using them?

What is the main issue with them?

I didn't really like the way that question was a worded, but I began with the caveat that I only know the issue I ran into...

I sort of like that question because the syntax is different in every database. If you can answer that question, then it shows you've been burned, and serves as a good launching pad for mitigation.

I certainly wouldn't hold it against someone if they never used a trigger and of course, SQL is complex enough that trigger can be replaced with many other topics.

To be honest, I wouldn't be able to write one off the cuff right now, but I recall the steps. Postgres has terrible syntax.

When I worked at a bigger company, we did. Even now that I'm at a startup we usually send wireframes of big UI pieces to a sort of combo designer/UX person we contract with to get a nice design/UI.

In this case AR wasn't able to do it directly but the implementation is drastically different depending on what DB you use.

Postgres has specific data functions that mysql does not have, etc..

Interesting article. I mean our designer is talented, but I need to speak up in terms of getting everyone on our small team to help out with what makes sense. The site is basically going to launch in a very difficult to navigate state. It's fine initially, and we're going to try and get up a new dashboard before christmas... but still kinda lame.

I'm not great at UX either, but I seem to think about it more than he does.

I actually have a good practical suggestion for this problem.

You should avoid the question of "expertise" or "who's responsbility it should be" -- everyone thinks he's an expert at UX, and no one is.

Instead, pitch it as a data question. That's the great thing about UX -- it can be tested qualitatively. In many cases, you can frame the question in terms of funneling, click through rates, conversion percentages, etc. So come to as much of a consensus as you can, and then a/b test different ideas. The important point is to get everyone on the same page philosophically: these are questions that need to be answered with real data.

Yea, gotta use as much data as humanely possible. Users do some weird ****.

No one?

on the topic of ORMs, do any of the .net guys use things besides nhibernate? How do you like what you use?

I don't like most ORMs (they all seem to specialize in making easy tasks easier, while not making harder tasks any easier) so I don't have a positive opinion on any of them but Entity may have exceeded NHibernate in popularity.

Got a Pi today, will hopefully have it set up as a printserver soon so that I can stop breathing deadly toner :P

Then I dunno what else I want to do with it but I'll find some stuff to do (want to play with ownCloud)

I am waitlisted #1 today. So my chances are pretty good of getting in for that Java Programming 1

Good

Meanwhile you can check out (there's two parts iirc, just navigate back/search for Java)
http://pluralsight.com/training/cour...1-introduction

and (hopefully that's the right one)

https://www.udacity.com/course/cs046

I installed RasBMC on mine (a model A) and its incredible how well it works.

I'm not gonna lie...I'll be very tempted and eventually put Plan9 on it.

Learn something new every day, had never heard of Plan9

The PI is great, but I ran into 2 common problems (which I never had to deal with on other machines).

1 - you get a LOT of memory errors. For example, on mine, I cant clone a git repo over 100MB as it tries to load it all into memory first. Its the same when compiling some programs. Everything pretty much assumes you have at least 1GB of ram these days.

2 - installing/upgrading/compiling takes a LONG time. Prepare to make a lot of cups of teas.

Sweet thanks!

I got in and paid the fees already so I'm all set. I'll try and finish the python course that I already started a month ago then I'll start on your other links.

I'm getting ****ing tired of devs who think they have a productionalized product but "don't have time" to ****ing document some of the steps. ****tards.