Suzzer I think this has a lot to do with companies being really picky and cheap more than anything else. I don't know anything about front-end JS development but I'm guessing your code was fine. I don't give much credence to the notion that there is a severe shortage of software developers. A shortage of inexpensive software developers sure but companies are still picky especially about hiring very senior developers.

Companies make lousy hiring decisions constantly. I could go on and on about how ridiculous the process is in many instances. Asking you to submit a "homework" assignment has it's share of problems. I was given a "homework" assignment about a year ago. The assignment was too vague as it could have been interpreted in a lot of ways. Maybe the assignment you got was vague, you made certain assumptions that they didn't make and that was that.

Crap I forgot to put Company X. Oh well. **** em.

Yeah without a doubt it was vague. It literally said:

Which I explained all of in my email, how I would extend it to handle these things. Well apparently they wanted a framework instantly ready to handle them with no refactoring, based on the vague incomplete requirements above. Oh yeah, I never got to explain anything because they never asked me about it. I'm not even sure they read the email based on their comments.

I got dinged because they don't think I know how to convert a singleton controller into an object that could be inherited and instantiated. I explained in the email that I would factor those into a base object, but right now there are more differences than commonalities and it didn't seem necessary.

I also got dinged for an async race condition where the user could select a couple things in rapid succession and the second thing could come back first. Totally legitimate, but something that I think any developer could miss on their first pass.

And stupid **** like not using namespaces and not concatenating my JS. I guess I was supposed to deliver a production app.

And i got dinged for using the DOM to keep track of which item is selected. I mentioned in the email I wasn't nuts about that, but it worked and was super simple from a code POV, so I kept it in. I'd say that was worth half a ding. Again something very easy to use a tracking object instead of looking at the DOM. Just messier code.

Basically I developed this exactly the way I would in the real world - keep it simple with an eye toward future requirements. But don't add a bunch of layers of abstraction for unknown or incomplete requirements until they become clear. It's a lot easier imo to refactor something simple to add another layer of abstraction, than undo a mutli-headed hydra with more layers of abstraction than it needs. I guess they wanted the exact opposite.

I also never claimed to be an SPA expert, but that I am good when you throw me into the deep end on new stuff. Well new stuff apparently means tons of different devices but we want a developer coming in who knows a ton about how to design an SPA designed a very certain way (which any two-way data binding framework does for you anyway). But now apparently I am poisoned from even the node team, who I had talked to, because of this.

Grrr still butthurt obviously. Eh **** em. I'm happy where I'm at. Our technology stack is ahead of them anyway. They're just converting to node and don't even have automated testing.

Key question would be whether it was a third party recruiter or the company itself. My experience with linkedIn has been that companies cast a wide net in recruiting people on LinkedIn and they just filter people out as they go along.

What I think is crazy is that with LinkedIn companies could actually communicate much more effectively with candidates than when they post an ad and have people respond by emailing their resumes. My experience has been that companies show an interest then ask you to send them your resume then go through pretty much the same process as if they had posted an ad. I recently was contacted by some company. They asked me to email my resume. I responded that my resume was online already at LinkedIn (which the could see and I assume the content was the reason they contacted me) and that they were free to contact me with specific questions. That just didn't fit in with how they wanted the process to go. Oh well when I need a job I'll probably act differently.

My name is shared with some utter whack jobs who pop up way north of "me" in any search (think magic psychic crystals and alien conspiracy theories). I have my photo in one place, but even that may not pop up my image. No FB, Twitter, LinkedIn or other social media.

I once did a search on myself. The results matched my city history, phone number, and other things. I also discovered I had over 10 arrests for all sorts of ****, which is strange considering I have no criminal record.

Not a challenge, but most interviewers would have to be pretty damn bored to figure out who I am. It took me several hours to locate me and, of course, I was cheating. It is scary to think that they do a search on my name and find a bunch of wrong information and judge me on it.

Suzzer, if you're interested in moving to NYC I have a job posting you might find interesting.

I will say from your description that I don't think X is being that unreasonable. Literally a ding or two is probably all they need to disqualify a candidate. They have a ton of candidates so they are in a position to care more about their false positive rate than their false negative.

Yeah but they reached out to me and they know I have a fairly lofty position at a big entertainment company with a very visible website. It's not like I just walked in off the street. I think you give someone like that every benefit of the doubt and don't just leave it up to some random engineer who may have a very specific way of doing things, might be having a bad day, doesn't want his job jeopardized etc. I've talked to 4 people at the company and already had a phone tech screen. They spent a lot of time on me already. Then they just reject the assignment out of hand w/o talking to me.

I really feel like something weird was going on with this. My gf works with someone who worked at X who says there's something nefarious going on with their tech screens. She's gonna fill me in later. FWIW candybar looked at their comments as well and wasn't very impressed. I will let him fill that in.

Nah I don't want to move. I was just willing to consider it for this chance. Honestly after all this I'd say no if they somehow come back and change their mind. Really bad taste in my mouth.

It sounds like they asked suzzer to do x, y, and z expecting him to do it a certain way without asking him to do it that way or even hinting at how he was supposed to do it. It's not like there's only one way to make something work. Seems pretty unfair to ding him for that imo.

Ignoring Suzzer's specific situation because I don't know enough - I'll say that we ask a very general homework assignment and also look for some aspects of it to be done a certain way without specifying/hinting at what we want.

Sometimes thats the best way to see if someone really understands the complexities of the problem.

Edit: I say this in the context of recruiting - which means that the 'best way' isn't necessarily a highly accurate thing.

Also, because I like recruiting bonuses, this goes for anyone else that I know well enough on this forum that is interested in a programming job anywhere from front-end to back-end to data-science to dev-ops.

How the hell did you get into that? That's way beyond me man. Grats.

Believe me though, I wouldn't worry about your name, I'm trivially easy to find and trainwreck from my postings here and people have plenty of cause, and nothing's ever happened to me. Your name's pretty common too.

Company X has one of the worst JS websites I've ever used, from a user standpoint. Nothing ever works for me and it's constantly changing. It's extremely clunky to boot. From my totally irrelevant opinion, I wouldn't weep too much if i were you.




Anyway thanks adios. I appreciate the support, this has been kind of a rough week for me. Positives though - my barrage of texts got through, and the chick in my group who I thought was a terrible programmer actually implemented everything (and more) I wanted on my wishlist that I couldn't get to, and drastically improved the visual layout of the program. I am glad I took a few days off to let the other programmers catch up before I did too much. I quickly fixed some of her more serious bugs and am no doubt gonna have to restructure things, but I'm amazed at how her ability to make things look nice and make components work well together complement my ability to code the business logic. This project's gonna be fine.

Also, it's really fun building something finally. I've been in programming stasis for a year, I'm afraid to write more than 10 lines of c++ code for fear of either my computer blowing up or my teacher marking me off a letter grade for forgetting a const modifier on some totally irrelevant function. Programming in java is therapeutic for me right now.

Yeah that's not a bad solution, was thinking it'd be inefficient with lots of tables but if it's only needed on the admin end should be OK. I think it probably wont just be limited to admin end in future so probably going to go with other way.

Just released new arcade in beta (sound warning):
https://www.scirra.com/arcade/action...ficetrouble-15

Any feedback appreciated! Was surprisingly difficult to code. Still some features to go yet (comments!) but pretty much there now. Feels good to publish stuff been a while now.

They may have found someone they preferred more and just came up with a reason to reject you. Do you think it might be a possibility that they give you an assignment that is sufficiently vague such that they can use it as a reason why they don't hire someone? As far as all the screening, there's ample evidence that companies don't mind wasting your time looking for employees.

I guess it would be considered the driver.

Basically:
- we have an instrument hooked up through USB that constantly outputs data
- recently for a project, we changed that instrument to stream much more data (~4MB / second, up from ~1MB / second). At the same time the data format changed for this project.
- there was a previous program that I am modifying to accommodate the new data.
- previously the program kept up fine w/ slower instrument, but now packets are getting missed on occasion

I have the USB reads on their own thread. This is all with winusb functions. Next, I guess I'll look into giving the thread higher priority in windows, etc - but was just curious if anyone has experience here. (This is new stuff to me)

once the conceptual problem is solved, efficiency is a detail. eg, you can make it an actual table that updates itself to stay synced via triggers, or use views, or whatever based on your db and needs. one way or another you can get this "table" + speed without much effort

Seems like you are opening the windows USB driver and your program is running as a user mode application. All perfectly reasonable approaches.

Yeah not sure if this is a Windows problem or not. Problem with upping your thread's priority is that the Windows scheduler gonna do what it is gonna do IE prioritize threads the way it views as optimal. I mean upping the priority could work but it might not have that much impact.

This to me is an interesting problem, are you using Windows overlapped I/O?

The CreateFile() call uses FILE_FLAG_OVERLAPPED... but from there, I am not sure if I am doing the necessary steps.

(So thanks for pointing out overlapped I/O.. I am reading up now to understand what is going on)

Yeah, the problem as stated is not particularly hard, so they are judging you mostly by your ability to guess their intentions and fill in the blanks. And looking at their comments, I don't think whatever solution they'd come up with would pass my screen, if I were to be equally as harsh from my perspective.

Hiring should be done primarily based on capacity, character and motivation. It's hard to say for sure without knowing exactly what you do, but in general, technical screens should be used to judge the candidate's capacity and it's hard to imagine what capacity you're testing for if you're not stating up front what you're expecting. Generally, "let's see if the candidate can guess what I'm looking for" type of questions are used to test for social and cultural affinity - it's a nerd's way of seeing if you know the secret handshake.

If your technical questions are so easy that too many people are passing, make them harder or use other criteria. There's no real reason to judge people based on things that they don't even know are being tested for during a technical interview. What's worse, it's a way for interviewers to feel smug about their own abilities when they themselves aren't good enough to pass or administer objective tests that are sufficiently hard enough to distinguish between candidates.

If you're worried that the candidate's style of coding doesn't mesh with the existing team, give them an existing code base and have them add features and tell them that they will be judged based on their ability to follow the existing style guidelines and work within the constraints of the code base.

Overlapped I/O == Asynchronous I/O. This is what you want so that is good.

When you do your create file you specify an event flag as one of the parameters and use the appropriate Wait For Event call. Typically you wait on the event flag and an abort event event flag to kill the thread when it fires, ymmv. What you want to do, in my view, is have a separate thread simply waiting for the event flag to trigger, have a queue interface that synchronizes access to the queue making sure that the critical section is as fast as possible IE doing the least amount of work necessary, and have a separate thread processing the queue. Prioritizing threads within your running Windows process should work ok btw.

Edit: One more question (for now) are you doing some sort of graphic display with the data while you are collecting it? Writing data to a file? There are lots of things that can kill your throughput while collecting the data.

IDK doesn't sound like suzzer's attention to detail was very high here and I'd ding pretty hard for that too. No namespacing? i.e. you did a whole app with globals or inside of DOMContentLoaded? I wouldn't be too happy with that from a sr candidate either.

$ sed -i 's/Backbone/SuzzerVC/g' backbone.js

why reinvent the wheel amirite?

It wasn't all global. The namespacing scheme he used was reasonable for the size of the app he was asked to build - he basically left the top-level objects in global namespace. If someone had a problem with this in a technical screen, I wouldn't let that person interview anyone.

Again, the problem with allowing your interviewers to add unspecified criteria as they please is that it becomes entirely about tickling the egos of interviewers and not at all about fairly evaluating the candidates. Why not stop at namespacing? What about dependency injection? Idiomatic usage of prototypal inheritance? Did the candidate use IoC patterns properly? Does the code use promises to avoid unnecessary callback spaghetti? Is the code properly annotated jsdoc style? Sometimes you can't even win. Is the code is written cleverly using functional style? Ding them for not being straightforward and point out potential performance problems. Is the code written in a straightforward imperative performant manner? Ding them for not taking advantage of abstractions and writing verbose assembly code.

If you/other people have looked at his submittal and think its fine then he probably didn't want to work there anyways if they'll pass on him based on the above.

^ tyty - I'll start working through the event flag flow. My critical section is minimized already, but good point there too.

And yes - it does lots with the data.
- analyzes it
- pattern recognition
- records to file (I disabled this when troubleshooting)
- displays real time graphics (direct3D)

.. those processes are very heavy and are resource hogs too. So I will play around with bumping he USB thread and see what happens. Next step is to wire up 3x of these instruments - so figuring it out will become imperative.