I had never used it before but will give it a shot tomorrow. I know it's not an issue of overusing RAM or disk accesses because top didn't reveal anything unusual.

I could do it, but not in Java.

It could be a bug in the OS or possibly some weird feature of the file system that you're running up against. I've never personally seen anything like that happen though.

It's probably an OS bug. strace reveals nothing, no difference whatsoever among processes between the 2 output methods aside from time. I also randomly encounter situations where I can't create a subdir under a dir unless I blow away the dir first. It's being run inside VirtualBox so it's bound to have imperfections like that. This is just a temp situation until we get our datacenter up and running next month, then we'll be on a legit Redhat OS. Startup company woes.

building a new PC today. Super hyped about it. In the past, I would have a master partition for a Linux install and dual boot to windows 10.

This time, I'm thinking just have the Windows install & maybe do either virtual box or docker for any linux dev environments. Now, I can get virtual box going no problem, but would it make any sense to do something like this in docker?

I wouldn't mind saving a little bit of hdd space if possible, but I dunno if docker would do that sort of thing. I know for VB I need to have 8-20 gb reserved for each linux install.

Thinking ubuntu for ruby development, and maybe lubuntu for js/node.

Thoughts?

Docker is probably most useful if you have certain specific requirements.

Like... doing dev for our product requires that you have a running instance of mysql, mongo, our web codebase, a celery worker, and a few other related things. You can make docker containers for different configurations and flavors of these things and mix and match them. For example when running unit tests I use a mysql container where the database is created at the start of the test run and destroyed at the end. For development I use a container that refers to a persistent store on my host machine. And so forth.

I doubt docker is going to save you any space.

If you're familiar with python, docker is more like a way to have lots of different configurations, the way that you can have multiple virtualenvs on python vs one big system wide python library. And similar to how you can spin up a virtualenv from a requirements.txt file, you can spin up your docker instances from a Dockerfile, which contains instructions on how to build the container.

The containers do tend to be minimal - you add to them only what you need. But you often end up with lots of images around so I think the space savings is probably nonexistent.

Ah, gotcha, cheers.

Yeah, I don't do any python atm, so that's not a huge concern. I've just had issues in the past having lots of installs and updates and mixing with windows and keeping track of which thing has what. All got very confusing.

Figure having separate, distinct VMs for separate, distinct purposes would be the way to go.

The python was sort of an analogy, docker is specific to python in any way or vice versa.

If you want to have several distinct VMs then I think docker is a good choice. The reason is that you can make containers that, say, all "start" with the same base of a linux install. Then you add on to those. In that case you may be able to save some space.

Also, docker emphasizes making VMs from scratch, in a way. You should generally kind of consider your docker containers as things that could be deleted and created from scratch at any time. That mindset means you always have a means on hand to recreate your environment.

I've started using docker for the build environment for my personal projects. I have been programming for 20 years so I have lots of old projects that have suffered from bitrot. They use older versions of libraries or depend on things that are no longer common. Rather than having to bring a project up to date just so I can fiddle with it or show it to someone, I can have a docker container that contains just what is needed to build that project.

Of course, now I'm dependent on docker, which will be super ****ing annoying in 5 years when it disappears.

got a new india developer, after 2 weeks he finally commits his first piece of javascript.

Seems ok(?) except x being 15 or empty string seems bad. And unless labels_show_histogram came from some other out-of-our-control weird source, I'd be comfortable w/o the "=== true" part. (Typescript ftw)

Also for the love of god 8-space tabs. I just cannot. It's like "where did the next line go?" (swivels head) "Oh way over here".

Fixed?

New proposed standard at my company:

I dunno guys, let's not rush into things.

This. I don't know why people try to code on a host, or on a non-production environment.

A+

Doing things that could be done in 1 line in 5 instead and comparing against true is just garbage. Weird thing is he came from my sister team. Did they not do code review at all? IDK I'm afraid to ask.

I've seen so much worse from our offshore programmers I guess I was just happy the code made sense at all.

Hardly. Grue is right, the original is pretty bad. In addition to the problems you fixed, should be using js naming conventions, should be using shorter, better variable names, for the love of god should not be using the monstrous underscore for a well supported js feature.

other mistakes which may be defensible / are more opinion based: should be using es6 arrow functions when appropriate (if you need full browser support, add a compilation step to your build process). inside arrow functions, use one-letter variables names and "i" for index -- local context usually makes additional clarity unnecessary in this particular case.
most likely, the data structures themselves are poorly designed, but it's impossible to tell from this short out of context snippet

Yeah I thought about the map one, but does every required browser support JS native map? Looks like IE 8 doesn't. http://caniuse.com/#feat=es5

It didn't really occur to me to rename variables that were created outside the block - as our Indian friend inherited those variables from someone else's code.

even if you have to support old browsers, MDN provides short polyfills for all of them that you'd be better off including than underscore. sure, maybe _ is already there from some legitimate reason, blah blah, it's always possible but i think incompetence is a safe assumption here

again, possible, but let's play the odds....

He said that was his first code ever. So he must have inherited the variable name.

I agree with you that native JS is better than underscore. But I wouldn't hold it against some developer like they're incompetent for using underscore. Maybe they've been working with underscore since before there was much browser support.

If anything I was impressed they know how to use map at all. Most of our offshore devs would just use a for loop and populate a new array. I'd be overjoyed if I saw one using .map or _.map

Did everyone else get this TWO TIMES in their email box today?

I'm interested in where they "came across my profile." Was this from the hacked email dump from LinkedIn?

I'm living in "no snow" USA and not a single item on this list is something I work on at all.

ha, fair enough. it's all relative.

yeah inherited from our crap.

been trying to replace an indian guy for about a month. Talked to ~5 candidates and I can't get one that knows what "map" is in the door. I've talked to 2 "superstars" over the past year or so who do know the slightest bit about functional programming but they vanish before we can hire them presumably for a better place. Everyone else for loops and pushes all the things.

Then the manager gets mad at me for rejecting them and I'm like "they can't even tell me what a CSS selector does, specifically, they wouldn't do well here". Like I ask a basic technical question and I just get total word salad thrown at me just hoping it sticks and I move on. Its hilarious and sad =/

Yeah every one of our offshore front-end devs lists 'jquery' as a skill. Which means they had a 2-week crash course.

We did eventually get some good offshore devs. But it took many rounds of them sending crap and us complaining and literally letting projects fail because of it (which sucks for everyone). Also my boss went to India like 8 times, sent some of us devs over there, basically made it his cause for a year to improve our relations with them and the quality of devs they provided. It was a huge effort.

One of the sharpest guys I ever worked with is an architect they provided fairly early on. If they could clone him they would take over the world. Most people like him leave the company. But he lives in some smallish city in India and has a sick Mom he needs to take care of or something - so he sticks around because they let him stay there.

I've also heard that most of the people hired by offshore companies are the 2nd tier students coming out of medium schools. Or something like that. They got a few very top people (like the architect I mentioned), then everyone else is average at best. The good/very good people in-between don't wind up as offshore devs for some reason.

What's been happening now is the really good devs go from offshore to onshore - which is held out as a big carrot for them. Then once in the US for a while they find someone else to sponsor their visa and move on - because 1nf0sys pays for ****. At least we usually get 2-3 years out of them between offshore and onshore.

The range of skill between candidates is amazing, I don't really get it at all. And their resumes are just garbage! It is literally a list of "skills" they supposedly have and then what the company did/does where they worked! I don't GAF about that, show me some code you've written. Why is that hard to communicate? blah blah different cultures I know.

I don't really do a lot of coding in my spare time, and my employer (and specifically my non-disclosure agreement) certainly wouldn't let me show you code from work.

If you want to see how they write code, perhaps you should assign them some sort of task to complete in between a first and second interview or something?

Wat? Isn't that basically what a resume is? I mean, I would probably get a little more specific about the skills I have and how they were applied in previous roles, but what exactly is it you are looking for in a resume?

I don't think you understand how few employers even bother looking at code or projects at all. It's not like we don't have github and can't go to graphs->traffic or open up gAnalytics and click a few links to figure this one out.

So, if no employers look at code, there isn't much incentive for applicants to work on code in their spare time or push code to a visible repo.

To make things worse, you actually have employers who don't bother looking at code at all, but look at your commit count. If the count isn't high enough, they figure that you aren't coding outside of unemployment time, and that counts as a strike against you. If there are no forks, they figure you are wasting time on a silly project.

And also, so much code in the spare time is just mind dumps of crap no one bothers to read anyways. Github is a wasteland of 10 LOC trash, which works against us who are really building out massive projects.