nathanappere

The parent comment is about adjusting pay for productivity, what is the reasoning behind "in a world where you compensate for productivity, you'll have to pay more employees that are less productive"?

Interesting use case.

Some quick feedback on the homepage: - "predictable": serverless is not exactly "predictable". On the other end, having a set number of instances is :) - "low cost": this assumes specific patterns of usage under a threshold where it's profitable to use serverless. That's entirely "use-case dependant" & a bit misleading. - "constant speed": so you're saying that the worst case scenario is bounded to 150ms, no matter the data set?

Pointing these out because it makes it read as "yet another DB vendor with some unrealistic claims" :)

If you prevent CEOs from selling their stock for the next X years, sure.

You mean costs the same and as a result is entirely a function of how well you're paid in order to be able to afford better coverage?

I never understood why this would not apply to any investment, as it always distort the market.

Looks interesting, curious to try it with our "assets-centered" workflows!

One of the issues here is about Uber getting away with paying the same type of taxes than other businesses, which creates an unfair situations with other companies & a race to the bottom for workers rights.

If they are not required by law to provide the reasons and some context, it is difficult to assess if they are discriminatory or class protected.

Reductions. How do you ensure in other languages (or in other VMs) that every actor will get the same cpu time and that none can block the others? That's a tricky problem to solve as a lib.

"So the net result was that this optimization really resulted in an extraneous recvfrom call per request, which returned EAGAIN. What was I thinking?"

That only happens if somehow you have received data that is EXACTLY the size of the buffer you send to recvfrom. If you have read less, you know that there is no need to call it again. If your buffer is full, then odds are there is more data to read so the next call won't be useless.

I suspect you actually had an extremely small proportion of "extraneous" calls.

(So all Kenya needs to do is create a colonial empire and exploit other countries ?)

Interesting!

This looks like a demo day project of Le Wagon, a french bootcamp in Paris (I found a link from their demo day https://youtu.be/hHPOV8vz5Q8?t=503 ).

Where you part of that bootcamp too ? (It seems to reuse the same "Nul n'est cense ignorer la loi." gimmick!)

Number 7 is exacty what you want: http://www.nirandfar.com/2012/08/the-webs-slot-machine.html

For the javascript part: https://github.com/fnando/i18n-js

Agreed!

Yes, but it's like a thousand other things that can make you a better developer.

It has always seemed quite useless to me to try to learn as much theory as you can before starting to actually do significant projects. You should learn by doing those project, and when it appears that you need theory to make it better, then you learn it.

I think there is a huge parallel with the lean model for CS teaching, programming yourself should always be the starting point.

Sure, not the same quality of education for programming though. Since the two are free, I really see no reason to choose university over 42 now.

Well I don't know many schools where you only do projects, classes are absolutely optional, you organise your time exactly the way you want, you don't need any kind of diploma to enter. The only mandatory part is being present for the project defense.

Never saw this kind of teaching in any US university / school, so it's not that far fetched to call it "revolutionary".

It certainly is in France, especiall (Epitech has been doing that for 10 years but you had to pay something like 40k which for us is a lot for education, and also you need a high school diploma, which is mandatory if you want the bachelor/master diploma delivered by the school to be officialy recognised by the government.)

Epitech student here (42 is kind of a fork, the whole managing team is Epitech Alumni). So far they copied everything from Epitech and it seems to work in the same way so:

1) Mentors should be available at 2 in the morning. Not that they have to (most of the time) it's just that you mostly interact with teaching assistants that still are in school working on their projects too, so they usually pull the same kind of hours than you.

2) You have really tight deadlines and a huge amount of project so it's not unusual to have to work 15 hours. But the whole thing is just about learning to manage crazy constraints, you deal with it the way you want. If you choose to work 15hours a day an manage to be efficient then great, but it's more about learning to be organize / be efficient.

Why fake problems ?

Rewritting a shell / ircd / httpd seems like a really good place to start to learn about the unix system / network / versioning systems / working in a team / learning how to find answer to things you know very few about. That's part of the program.

Well I guess that when you have a good story it is a shame to let facts get in the way.

French unions have no say in the curriculum of a private school (and it would make absolutely no sense, lot of fantams about french unions there I think!).

Private school can teach pretty much what they want. If they want they diploma recognised by the government then they must comply to some policies of the ministry of education, but it is up to them.

Epitech Alumni here (42 is basically the same thing but free).

1. Studied in CSULB / UCB, you really don't do enough of them. More theory than practise when you should probably seek theory when you're confronted to a problem that needs it.

2. You really don't need to know about automata theory, linear algebra or whatever to be a good developer. Brings me back to what I said in 1.

3. Actually the paper doesn't depicts the reality clearly: you have a lot of teacher assistant always available to help you, it's just that they are not allowed to gives you the answer, only to help you search in the right direction. When you encounter something tricky where finding documentation on the internet becomes an issue then you can talk about it with them. The system works quite well in Epitech. I have been a C & C++ assistant there and the level amongst the assistants is really good, and you could never hire the same number of qualified teachers (if you could find them). Having people that worked on the same project themselves quite recently (like 1 or 2 years ago) makes it that you will always find someone who knows what he is talking about.

To teach CS this is really the best system I have seen so far, and it teach you to never get stuck, never trust blindly what someone says, and always go find the answers by yourself.

Epitech Alumni here (42 is basically the same thing but free). The idea is rather that if you start from low level programming it is not difficult to go up, and you can pretty much do anything after that. I do think that starting from C/C++ & having a good understanding of how any interpreted language work makes you a better web developer. Not that it guarantees it / is absolutely necessary, but it gives you a broader understanding, and that's what the founding team from 42 believes too.

Quick question, how much is given to Watsi (what is the profit exactly) ?

What would be the point of making it a standard ? It's already one de-facto. I prefer to have other solutions not trying to be devise but rather solving auth problems in their own way. Also, the lack of documentation is often balanced by the great readability of the code (Which is extremely clean / straightforward for devise & omniauth & most of the strategies)

http://requirejs.org/ addresses the same problem (AMD style rather than nodejs require style)

Seems like a very civilized way to deface a website, you have to love how GitHub changed the game.

A coffeescript version would be really welcomed even if it is a pain in the ass to maintain two version of the same source code.

I was just about to ask this question, but there seems to be some differences: loading of remote data (I believe in chosen you have to hook yourself the ajax part), the infinite scrolling is also a good idea, and the drag & drop sort was missing.

Just checked it and you're right. I believed I had encoutered the behaviour I described but I do not remember in which context. Anyway glad I learned something!