2nd interview at IBM tomorrow. Sounds like it's an all day affair. Wish me luck. It would be really nice to get a written offer and not have to stress about finding a job and just be able to focus on my last semester.

Well, yeah, I used PowerShell once.

I've got an unused Linux distro installed. Any chance shellshock could git me?

Good luck! Let us know how it goes tomorrow!

well yeah that's what happens when you use arrays like that assuming length is mutable

It seems I suck at sarcasm.

Damn shellshock probably undid >1 year of Linux lobbying I did. ****ing RMS with his "blip" remark still has me tilted.

Hah, I figured you were joking but I really do think windows could do cool stuff.

Actually it was vectors but I'm guessing I was waking on memory due to the fact I was using the subscript operator vs. the .at()

The interview was very interesting but not at all what I was expecting.

They didn't ask me a single technical question.

I interviewed with three different managers that each had openings in their areas, along with developers in each department.

Most questions were about school projects, teamwork, personality, etc.

The development environment also was nothing like I expected. Every developer has their own quiet little office, and most of them seemed to only use one monitor. I was expecting areas with cubicles and for most developers to have multiple monitors.

All three jobs seemed great and I would love to be offered any or all of them. Developers can wear pretty much whatever they want, work whatever hours as long as the work gets done, etc. Seems like a pretty sweet environment to work in.

Sounds like you passed the technical screening and now it's just them finding a department to place you

1 monitor is def. a leak. Even if you don't use the second one it improves efficiency. There's a pretty solid study on that iirc

Yeah HR questions are typical. Stuff like thoughts on teamwork, what roles exist in groups, how do you react if someone continues to turn in bad work etc.

where was this interview? or rather, where is the intel office located? and how many of the rest of us can you bring with you?

It was IBM, not Intel.

I am wondering if the lack of technical questions was because all the jobs were for doing development work that we don't really learn about in school.

Each of them are working at or near the kernel level doing DB2 optimization.

Sounds pretty interesting and IBM is a very good employer these days from what I hear.

Kato, looks like you got it locked in.
If they see you've got what it takes to learn and go, especially as it seems to be so specific that they can't expect an outsider to be familiar with the realm, they'll give you the tools and put you in a team where you can fill the gaps.
I'll hold back on congrats til it's been confirmed, but from my experience it looks like they do want you.

On a different note, I have been wondering lately where do small companies who do need some programming done, but not to the extent for filling full time positions look for candidates?
Do they all go the elance/odesk/et al. get-some-one-or-team-from-India route?

If that's the case, then there should at least be some work for someone like me, looking to do part time telecommuting work, to clean up the mess they get handed over, right?

Small sample size, but IME, companies that outsource and get screwed by a US or foreign company just takes the software and deals with whatever they are handed.

The company is not a tech company so they had no intention of hiring a programmer in the first place. They did not see the value of having an in-house developer and getting a dog handed to them just confirms their opinions and fears. Convincing a non-tech of your value to their company will be a steep uphill battle since you'll have to convince them, with no way of vetting you, that you are legit and can offer real value to them. They may wonder why you think you, an individual, are better than the team of 10 programmers they hired for a great price. Why would they take a risk? If you drop out, who replaces you?

Ah, interesting. Even if several features do not work as intended?

Wait, looks like I have expressed myself rather poorly. Not looking for employment but to fix/improve the botched product they got handed on a fixed price project basis. If I can't fix it, I don't get paid (in most cases, I'll know as soon as I dissect the code).

I was wondering where these types of companies go looking for help so I can advertise accordingly.

I wish I could speak for all companies, but I can't, so I don't want to offer advice. From what I've seen, they don't look for someone to fix broken software; they keep going back to the people who made the POS* software to have them fix the software, which is sadly a large part of the "liaison" jobs I've had over the years. These are total scumbags, and I've had to bite my tongue on many occasions, though I've failed to do so 100%.

If I was to start anywhere, I'd say that the big issue will be money. They don't want to spend more than $2000 on a product, and they won't be happy to drop more on the software, especially if it isn't a sure thing. These companies don't want to be a "technology" company, which is fine, and it is a business decision that makes sense in many ways. Just note that you are basically a sales person more than a tech person with these companies.

I would say just find a bunch of companies with crappy websites and see if they are interested in upgrading to test the market and see what happens. Internally, they are likely using a bunch of closed-source, fee-based crapware to run the rest of their company, so you won't be much use there.

I suspect that other posters here have a larger variety of experience in this domain (shoelace?).

*Hopefully obvious I didn't mean "Point Of Sale."

I've got 200k records I need to match spread over a dozen files, each with 200-300k records in them (an ingest process silently failed on batches of records), currently I'm zgrepping for them which doesn't seem to be very fast, but was quick to write. They are date based (bill end dates) but it's not going to be a perfect match. Anyone got any faster ideas than best guessing batch the records to be found by the date range of the individual files?

Does this mean you're running a separate zgrep operation for each of your 200k records? Am I wrong to assume that there won't be any caching between runs? Could you write a simple program to load your dozen files into memory and then scan that? If it's still too slow you could try throwing more threads at it. Assuming your records aren't prohibitively sized of course, but even then you could load say one file at a time, do a search on that and then go onto the next file?

We weren't expecting near this many records so hadn't spent enough time thinking about this step of the process. The list of 200k records was generated doing some map reduce stuff on our hadoop cluster and the guy who wrote it agrees he could have put out the entire bill record instead of just the hash he was using for the MR stuff.

Why do almost all restaurant websites suck? When I visit a website for a restaurant, I only want to know a few things.
1. Where are you
2. When are you open
3. What is on your menu (not pdf download)
4. How much does that stuff cost

Surprisingly few restaurants pass this test. Many of them do have flash and autoplay music though!

There was an interesting discussion on HN about this subject a while back.

The dirty secret of the restaurant business is that restaurants, aside from a handful of brutally efficient places, work in the red. They tend to run for about 2 years, the bank denies the next loan, then the place changes owners. They can't afford decent talent to make their websites.

Restaurants are looking for a skeumorphic "experience." Why? They don't know any better and with no talent, no one is telling them better. Restaurants don't have time to be working with their website. The PDF is given to them by the menu designer and they are fine just uploading a PDF and forgetting about the site.

Apparently, there are startups out there attempting to solve this issue, though I imagine this is like drawing water from a rock.

Perhaps they don't need better websites, I'm not sure how many more people through the door they'd get with say having a non PDF menu, or even listing opening times. I imagine most people searching would be already motivated/interested in going so will put up with that. Capturing a potential customer on the phone making enquiries probably has better conversion rates as well.

Perhaps having a crappy website gives it more of a 'this isn't such a known place' sort of feel.

And maybe obviously listed pricing could work against them some of the time.

I think there's probably quite a lot of factors at play there. If it does get more people through the door, in the UK a site like that might cost you in the region of £1k to get it done nicely. If you make £10 profit per head avg, you'd need the website to get 100 extra customers from the cost of changes which might take a few years.