Thread

James Clarke

Software craftsman, singer and lover of clean code

Oct 22, 2016

As a programmer what single discovery has given you the greatest boost in productivity?

Learning about Version Control Systems, 2 years back, has been the great productivity boost for me. I liked the idea of cloning a new branch for every new feature and have never looked back since.

So as a programmer what single discovery has given you the greatest boost in productivity?

#programming

Responses(16)

Josh Montgomery

web developer

May 14, 2018

Besides version control (for whatever reason, the past few jobs I've had did not use versioning until I got there), I would say the fact that I don't know everything about web dev, and that, not only is it ok to ask for help, but when you do find the right community, finding out the answers to my problems become so much easier :)

Ignacio Chavez

Front end software engineer

Oct 28, 2016

For me is the combination of VIM + Snippets. Keeping your hand movements inside the keyboard has been a good productivity boost. And as for the snippets, basically I spend little time updating/adding them, but it pays of by being able to reuse pieces of code that are usually common. For example, a Backbone extend or React Classes, function filters, maps, ways to find a property inside of an array of objects, etc.

Lorefnon

Open Web Enthusiast

Oct 23, 2016

That there exist programmable programming languages. The concept of meta-programming.

The discovery of languages like ruby (and more recently Lisp and F Sharp) has significantly changed my perspective on what I used to consider expressive code.

The notion that there is nothing really sacrosanct about the constructs built into the language and you can modify core language features themselves using the language itself, and emulate constructs that are as succinct and expressive as core language features takes the idea of "programming as an art" to a whole new level.

Interesting... how did this make you more productive, concretely? Do you often extend the languages you write with?

Emil Moe

Senior Data Engineer

Oct 23, 2016

Without guarentee for any money in the future I quitted my daytime job and focus 100% on my own company.

Casey Cole

javascript

Oct 23, 2016

when working with javascript if you want it done quickly just write vanilla

Mario Giambanco

Director of User Experience Development

Oct 23, 2016

https://www.youtube.com/watch?v=boj75h3urLU

Just go with it. Just "slide". Stop caring about what others are doing; how their doing; why their doing it. Just do you.

zealsham

Self Taught Software Engineer(STSE)

Oct 23, 2016

Algorithms and data structures, once i discovered the power of algorithms and data structures it changed me totally. i never saw the use of all of it till i opened my mind

TheSheriff

Co-Founder, Founder, Entrepreneur & Problem Solver

Oct 22, 2016

A more curious answer:
Sites and music similar to focus@will

Luis Orduz

Software Engineer

Oct 22, 2016

A while ago, it was docker. It really makes it easy to isolate and bootstrap one's architecture.

More recently, it was continuous integration/deployment/delivery. I used to be skeptical about it since I couldn't see myself trusting the results but once I set up push-to-deploy, at least, I haven't looked back.

And, pretty much a month ago, I decided to stop resisting the cloud, which I also found to be a huge boost. It turns out that, while I didn't hate managing architecture (in production), I didn't particularly enjoy it either and I'd much rather write code. And thanks to app engine (the cloud provider I chose), and some help from Gitlab CI, I can do that almost exclusively now.

The most time every engineer (and team in total) is spending on reading a code not writing. Optimizie for reading. Become a good writer. The best documentation is a well written code with comments, short simple functions, no abstraction.
Instead of wasting time on reading about which design pattern is best and applying over 9000 patterns every time only because you know them, thinking what architecture you need to choose, benchmarking microperfomance - just do it. Just write a simple code which will solve a simple problem. If you just need a single constant or a plain function, define it and not write a whole Interface, Class with some patterns. Later you can change it.
Invest time in self-development. Keep reading, learning, experimenting. (Tech < 50%)
Learn your tools very well (OS, IDE and everything else you use every day), all the shortcuts and tips.
Learn basic math, calculus, algorithms, data structures, software engineering principles and patterns

Can't say which one would be "greatest". Probably the greatest would be - being flexible and combining those "discoveries" together. But I definitely can say that productivity has no connection to programming, it works in every aspect of life.

So, learn the life.

sssss

Student, Geek, Open-Sourcerer

Oct 22, 2016

Taking care of my body. Eating healthy, sport, meditation, going to sleep at a normal hour, waking up early. Doing this has resulted in having a lot more energy and motivation to work and study (i am a student). I know a lot of people say this but it really does make a big difference. More than any tooling has ever done for me.

@fzeagru I would like to point out the importance of this . The way a good healthy routine affects you is miles more than any small tooling can do . It takes a while to establish as muscle memory but once it does , it pumps you to work in a focused manner and with a constant stream of active concentration spans

stuff ;)

Oct 22, 2016

I don't know, for me it was working on something i found inspiring ... versioning and all the other things are cool and maybe helped boosting it. But in the end the most productive things were focus and a honest interest/curiosity regarding the task/application.

Oh and the thing that really helps me is .... the internet .... :D

Selwyn Rogers

Web Developer

Oct 22, 2016

LESS, Grunt

And mostly: Stack Overflow

I have used and written some very productive DSLs.

While the idea of extending language may seem out of ordinary if you are not coming from ruby/lisp etc. but writing DSLs makes a lot of everyday programming grunt work easier and sometimes even enjoyable. Let me see if I can elaborate a few "concrete" use cases without bogging down with implementation details.

Perhaps the first DSL I wrote was to build data entry wizards. This was an obligatory routine work for some society in my college. But I was able to write a high level ruby DSL with basic instructions and could delegate much of the work to freshers, with limited ruby exposure who could get productive right away.

The code would look something like this:

wizard do

    step do
        before do
            # ... fetch some data.
        end

        inputs do |f|
            f.input :name
            f.input :email
            f.input :department
            # ...
        end

        validate :name, presence: true
        validate :email, presence: true, format: :email
        # ...

        after do |submission|
            # save submission
        end
    end

    step do
        # ...
    end

end

Implementation wise it was a light wrapper over formtastic backed by a simple Rails application which would persist the data in a sqlite database. But though it was implemented in ruby, it could be used by people with very little ruby familiarity - and this was a lot simpler and flexible to work with than GUI based form builders. I would love to be proven wrong by an Excel/VBA enthusiast.

In a job in a ruby shop, when building administrative dashboards, activeadmin's user friendly DSL had proven to be very helpful when building internal crud interfaces without having to bother with styling, UI etc.

Later I extended the wizard builder code when I needed to build some questionnairs for marketing. An advantage of using a high level DSL was that I could have the same specification to generate user facing forms as well as integrate with our internal activeadmin based dashboard - both having very different appearances.

Similarly, we had a use-case where using SQL was much more convenient than abstracting through an ORM. So I wrote a simple DSL to manage SQL templates.

template "find_users" do

    accept do |input|
        input.string :name
        input.string :age
    end

    query do
        "select * from users where age = :name"
    end

    map do |result|
        # Post processing to transform result array to 
        # a well behaved ruby object 
    end

end

Yes, it was not database neutral. And I am pretty sure that somewhere there is a commercial solution that does something like this in a lot more powerful way, but the thing is it was written for a specific use case, insulated users from common sql interpolation errors, was written in a single evening and got the job done.

Similarly our deployment scripts used Capistrano. In pre-docker era, when capistrano 3 had recently been released and good cloud service integrations were yet to be released, we had extended the Capistrano DSL with AWS ruby SDK to support EC2 deployments and integration with ELB and Cloudfront. This worked for a long while - without having to invest a lot of effort in a complex configuration management systems, and also without having to manage a lot of messy shell scripts. I am never going back to maven.

The common theme here is that a lot of cases where many people resort to XML, YAML or other configuration languages, you can handle more expressively through a DSL. This also a big advantage that you can just put a debugger somewhere and use pry to walk through the configuration.

This is applicable for templating languages like HAML as well - if some div does not show the correct value just put a debugger inside the div and inspect the local environment. Try that with velocity.

Search Hashnode

As a programmer what single discovery has given you the greatest boost in productivity?

Responses(16)

Recent threads