Live data from Hacker News

Bitlean: Learn to code fast by competing against other developers

bitlean.com

31–40 of 41 posts

Re: Bitlean: Learn to code fast by competing against other developers

#31

Just had an Eats, shoots & leaves moment with that title. I was thinking, why does anyone need to learn to code fast?

Well, learning a programming language is slow process, our idea is to make it faster and help the skills to stick better. I can't remember all the cases in which I learned something and had to relearn it again after that - only if my process of learning was better, it might have sticked from the first time. :) That's what we are trying to do. I don't know if we will succeed or not, but we are going to at least try.

That makes sense from a pedagogical perspective. It's a lot like exercise though, if you're Arnold Schwarzenegger, you are not going to make sweet gains from writing a for loop.

You'll need to find people at your skill level to go against and this presents a barrier for a lot of us at step one.

Re: Bitlean: Learn to code fast by competing against other developers

#32

Earlier quoted context omitted.

Well, learning a programming language is slow process, our idea is to make it faster and help the skills to stick better. I can't remember all the cases in which I learned something and had to relearn it again after that - only if my process of learning was better, it might have sticked from the first time. :) That's what we are trying to do. I don't know if we will succeed or not, but we are going to at least try.

That makes sense from a pedagogical perspective. It's a lot like exercise though, if you're Arnold Schwarzenegger, you are not going to make sweet gains from writing a for loop. You'll need to find people at your skill level to go against and this presents a barrier for a lot of us at step one.

My intention is the application to provide developers with the possibility to go on a higher level. For example:

- Level 1: Beginners (really basic concepts, just explanation of the syntax) - Level 2: Intermediate (more complex tasks, that would require developers to combine several basic concepts) - Level 3: Master (where developers will have to be really skillful and the tasts would be much more complicated)

My idea is when someone is just starting to make him fail fast so gets over the fear of failure and just starts writing code and learning from others.

Re: Bitlean: Learn to code fast by competing against other developers

#35
This whole thread totally loses me: I don't get it. Here's why: To me, the coding is and for decades has always been fast, fun, and easy. Similarly for the design or architecture for projects up to, say, 50,000 lines of code and several servers sharing in the work and sharding, etc.

Piece of cake. Simple. No issues.

But there are issues, huge, humongous issues: The main issue is just and simply making use of the documentation, and object oriented classes, and other software developed by others. E.g., I've gotten -- found, downloaded, categorized, indexed, and abstracted -- well over 5000 Web pages of relevant documentation. At 10 a day, that's a lot of days.

E.g., learning SQL was easy, piece of cake, obviously built a lot on set theory in pure math (which I know pretty well and have used a lot), and I learned SQL reading Ullman's book over dinner at the Mount Kisco Diner during about two weeks. No problem. How the log file works? Fine. Entity, attribute, relationship design -- sure, easy.

Then the real problems: I tried to install SQL Server. What a mess. Maybe it got installed. Someone sent me a database, and I wanted SQL Server to recognize it. Nope, we're talking high end, world-shaking research here, unsolved problems of the universe, much like dark energy. As I recall, my little effort ruined the SQL Server installation.

So, I make some progress doing some simple things and then go for an update (general rule: never update unless totally necessary). The install asked a lot of questions with no explanations or references to what the heck might be the appropriate answers. Somehow I got two installs, using Microsoft's side by side (I've written lots of programs and they all run side by side as far as I can tell -- what's the issue here?), and the install ruined my system. SQL Server uninstall wouldn't. System repair wouldn't. So, the SQL Server install broke my installation of Windows. I reinstalled Windows and all my software to a freshly formatted hard disk partition. Then I tried again. Eventually I got the SQL Server install to work -- eventually. The mud wrestling went on this way.

Ullman's book, SQL, etc. were easy. Getting a good install was a barbed wire enema with an unanesthetized upper molar root canal procedure while being poked with a dozen red hot branding irons.

Then came time for the code of my Web pages to connect with SQL Server. So, I need a connection string. Yup, we're looking at a challenge at least 1/3rd of a Nobel prize in physics. It took a week, a solid week just to get a connection string that worked. Why it works, I still don't know.

Due to the times installs ruined my boot partition, I wanted to backup the partition so that I could quickly return to a good system after some install had ruined it.

So, I used NTBACKUP. It asked if I wanted to save "system state" but had no explanation of what was meant by that. They are talking about my options for Outlook? For the fonts on my text windows? Well, after reinstalling all my software several more times, I finally got good notes and experience with both saving a boot partition and being able to do a restore that would boot and be good. Hint: Yes, very much do save "system state" or the saved copy, restored, won't boot. Yes, that NTBACKUP does a volume shadow copy, that is, saves a boot partition while that partition runs, is super nice, but writing documentation on "system state" was too difficult for Microsoft.

Net, the coding is fast, fun, and easy. The difficulty is the documentation and workings of other code that needs to be used. And now we have to use a lot of such documentation and code.

Anyone have a way around the problem of such documentation and code?

"Learn to code fast"? That's less than 5% of the work. The other 95% is mud wrestling with bad documentation.

Re: Bitlean: Learn to code fast by competing against other developers

#38
post #20

I'd like to see a competition called "Learn to Code Slow" that encourages people to think and design good solutions and then implement them robustly.

And perhaps the winner is decided after the fact by a vote, based on metrics of efficiency, elegance, etc... I'd love to see a more natural, bite-sizable, exposure to other programming languages and/or paradigms in this way.

I like this idea. That doesn't devalue the idea of being able to efficiently solve relatively simple programming challenges

Re: Bitlean: Learn to code fast by competing against other developers

#39
So, there's two major things I would suggest changing:

1. Easy answers keep users from developing Google-fu. Competitors answers should be sufficient. If they aren't, make a separate doc and have it be easily Googleable from a hint. This keeps them from developing a dependency on that "Answer" button that would show up on the end.

2. The big problem with currently existing tools for this is that they haven't come up with a gameified way to give -structure- to a program. Which is a problem because for a real program structure is the very first thing you do.

Post reply on HN