Fork me on GitHub
#ldnclj
<
2015-09-11
>
Pablo Fernandez08:09:23

It got too long, so I broke the blog post in two parts.

jamiei08:09:47

Nice one @pupeno , Not so bad on the word count, were you counting the code in that?

Pablo Fernandez08:09:07

jamiei: yeah, the code is counted

Pablo Fernandez08:09:32

I did a bit of editing as well.

oliy08:09:07

@pupeno very interesting and nice work to document all the steps needed to make it work. it is quite long, but it's not immediately obvious what could be cut. maybe the "javascript, on the server" para could be cut down a bit

thomas08:09:11

@pupeno looks very good. thanks for the write up.

agile_geek10:09:39

@pupeno: blog looks great. Haven’t got time to read it during the day but I’m going to avidly consume on the 3 hour train journey home tonight. See you at ProCloDo on 22nd if not before. 😃

otfrom11:09:17

Please consult your doctor before taking ProCloDo

agile_geek12:09:17

@otfrom: side effects include serious Pizza addiction?

xlevus14:09:36

how do people choose where to start a project. so many choices of webservers. Two different react libraries, a bunch of scaffolds, including one massive one

xlevus14:09:45

so daunting

frankiesardo14:09:16

I know, that's a pretty common problem

frankiesardo14:09:35

My suggestion would be: go with the mainstream option for the first time, learn pros and cons

frankiesardo14:09:54

that would be ring+compojure for web server, om for react frontend, leiningen as a build tool

frankiesardo14:09:15

These project have plenty of documentation + bigger community + lots of examples on the internet

frankiesardo14:09:09

And they usually have a solution for pretty much every use case, even if not the pretties solution

frankiesardo14:09:03

Then as you get more confident with a technology you can start to experiment different approaches that make what you want to do more idiomatic e.g. pedestal web servers with server sent events, reagent for quick & easy react prototyping, boot for flexible build scripts ..

xlevus14:09:47

Yeah, I had been looking at luminus. but there's just so much stuff in there, I have no idea what it does

frankiesardo14:09:35

A good project has a balanced quantity of novelty and boredom. Slowly introduce new stuff to keep it exciting but too many new things at once and you will never make progress

Pablo Fernandez14:09:42

xlevus: I’m very happy with luminus but apparently I have a good ability to not be bothered by my own ignorance. Eventually I got to know and understand all of it, at the beginning it wasn’t a problem. Finding where I needed to modify code to do what I wanted wasn’t that hard. Luminus documentation that introduces the framework to you is actually quite good.

Pablo Fernandez15:09:20

For me, personally, it’s easier to remove component X because I don’t need it, than figuring out I need to add component X and how to add it… specially when it’s something you are not thinking about or something you don’t even know it exists, like better error reporting, csfr, sane header generation, etc.

Pablo Fernandez15:09:26

About other components and libraries, what I do, is ask around, look at the pulse in github, to make sure they don’t look abandoned, read their documentation and if two things are equivalent (like bidi and silk), throw a coin. A couple of times, my reason to use the a library was that the author hangs out in this channels.