Posted on: 2018-09-07 14:49:59.000
When I started to write Rockets, I had a crazy dream. Then the dream came true, and I didn't know what to do.
The dream of newLISP on Rockets was to produce something unique and interesting, and share it with the world. I wanted to show that you could make a lightweight web framework that could make functional and fast websites with a minimal amount of code. And I wanted to do it with LISP, because I had fallen in love with that language.
I had some experience making websites with the Dragonfly framework, including my own personal site. That framework wasn't being updated any more, though, as the author had moved on to other things. So using Dragonfly as a starting point, I wrote my own framework. It was an intellectual exercise for me, and when it was finished I didn't have much else I wanted to do with it. It was out in the world, and other people could use it for free and add to it or extend it as they wished.
That was five years ago.
Since then, I got a new job where I used Rockets to launch one major internal tool, and prototype others. But the company I worked for was all-in on Python, and so I was strongly encouraged to learn and use that language. There's nothing wrong with Python-- it's a clean and flexible all-purpose language with strong library support and a large community. There is a lot of comfort and safety in Python.
In that time, I've read essay after essay from people who once used LISP to do amazing things and who have all moved on. Paul Graham, after failing to convert the world to his own new Lisp dialect, launched the successful Y Combinator startup incubator. Andy Gavin, who by his own admission was a "diehard LISP advocate" for twenty years, coded Crash Bandicoot and four Jak & Daxter games in his own Game Oriented Object LISP language. Now he codes in Ruby and says that "Libraries and implementation matter a lot. Momentum too." And the author of the Dragonfly framework, after bemoaning that not enough people picked up his object-oriented framework for newLISP, left the language entirely.
And here I was, coding in Python at work and thinking of LISP fondly, like one does of a first love. The years went by and memories got hazier. Was I just going to be the next person in a long line of coders who abandoned LISP?
Surely all these smart people had a point. Why go it alone with LISP when you can use something popular like Python and enjoy tons of libraries and support? Why struggle with something that's so different and unpopular and a bit weird?
But then again, why did all these smart people spend so many years (and even decades) doing exactly that?
In my spare time, I am a science fiction author. I have a new book coming out next year and I needed a website to help market it. I thought about setting up a site on Squarespace (yes, I listen to a lot of podcasts!) It's fast and easy, you can customize it to look like you want and... it would be just like all the other author book websites that nobody reads. I wanted something very different and unique. Something nobody else would have.
And then I thought about Rockets.
Now, I had already partially converted my personal site from Dragonfly to Rockets, but this was just using the new headers and login system, and doing some find-replace on a few functions (you can read about how to do that here:
http://newlisponrockets.com/rockets-item.lsp?p=85). Much of the site was a bunch of custom code that just worked for my personal site, and wasn't the same as the site you get from the newLISP on Rockets GitHub repository. So Rockets did some things and my personal blog did some other things, but neither of them did what I wanted from a new website to promote my book. And my personal website was looking a bit old-fashioned anyway, and it was running on an ancient t1.micro instance on AWS that was expensive and barely supported and needed to be moved over to a new instance years ago.
That's when it hit me. I needed a new version of Rockets. One where I could recreate my personal blog and also have a marketing site for my new book, using the same framework. With a simple Apache configuration, Rockets already supports multiple sites-- I have several different sites running on that one AWS instance (see here for how that works:
http://newlisponrockets.com/rockets-item.lsp?p=99). But I want a single codebase that can make different-looking sites depending on what you want.
newLISP on Rockets has a few advantages other than uniqueness. It is fast--wickedly fast. And it's simple-- the whole design principle was to make the simplest possible framework for making dynamic websites. What it's missing are features. I had thought that other people would take the basic Rockets framework and add their own features, but I think a lot more people would rather have a simple website with all the basic features ready to go.
To get there will take some time and effort on my part. But I have a good motivation to do so-- my own new website-- and I'm writing out a plan for me to get there. I'm going to make a new thread to report on the work being done. At the end, I hope to launch Rockets 2.0-- a basic but full-featured blog, ready to go, out of the box, batteries included.
I'm excited about the journey!