that “Serverless” name
Back in 2001-2003, I did some cutting-edge work involving rich browser functionality in a web application. What began with creative use of DHTML techniques turned into Web-based spreadsheets with hundreds of interactive, live-updating rows and thousands of cells mapped to an object model in what would today be called a reactive style. It was implemented using a hidden Java applet driving a Model-View-Controller design.
The best thing about it was that because it was Java on the server, and Java on the client, the model-layer objects were portable. No need to re-implement any business logic in Javascript. The only scripting involved was in using the browser’s rendering engine as a graphics context for the code running in the applet. I named it SHwing. (Get it?)
Shwing was cool, but very much before its time. I thought the name was clever, and might have gotten some traction, but the whole thing was proprietary to the company where I built it.
My good friend Durran and I wanted to somehow make the tech available to the wider world, but the executives were real old-school assholes and didn’t even pretend to understand why that might be a good idea. In fact, after leaving the company, when we tried to open source a clean-room reimplementation, they promptly threatened us with a nasty cease & desist letter from their big time lawyers.
I was very depressed about the whole episode. One of the hardest things was that even though I was proud of my innovations in the field and wanted other people to get excited about it too, I had trouble describing it without resorting to the long description detailed in the paragraphs above.
It didn’t have a proper name.
So imagine my perpetual amusement, years after the grieving was over, the style that we had invented did get a name and mainstream acceptance.
First some James guy coined the term AJAX to describe using the server to update the user interface dynamically between page requests. Then the terms RIA (Rich Internet Applications) and SPA (Single Page Applications) came into use to describe browser-based applications that dispensed with server rendering altogether.
Since then, well, it’s been one framework after another promising browser-based nirvana. All because clear labeling of what had been a fuzzy concept crystallized support and provided a foundation for a huge international community to rally behind what has become a significant leap forward in web application development.
While that’s all ancient history now, I think it pays to learn from history, on account of that whole repeating itself stuff. I think the same kind of thing may be happening again right now.
Microservices is definitely a thing already, and has been for a couple of years, but the term Serverless crystallizes a concept related to cloud-computing that we have only been able to hint and dance around before.
With that in mind, in December 2015 I started working on documenting what people call “serverless” application architecture style.
I wrote up my early experiments in a widely-read blog post and kicked off a new book project titled “Serverless: Patterns of Microservices Application Design Using Microservices (Amazon Web Services Edition)”.
The book is ambitious and I’ve been working on it non-stop during my entire holiday time off. The manuscript is already up to a couple hundred pages of content detailing how to build distributed systems using microservices deployed on Amazon Web Services.
I’ve only really had one issue in the last month, and it involves naming.
There are only two hard things in Computer Science: cache invalidation and naming things. — Phil Karlton
When I announced the book project, several commenters pointed me to JAWS, a growing open-source project intended to make application development with Amazon AWS’s Lambda and API Gateway products easier. I loved it immediately. Lots of energy and great talent behind the project.
The name JAWS, which I presume was meant to represent Just AWS. It had some issues. For one, it conflicted with a longstanding screen reader for blind people. I didn’t think it was that big a deal, but apparently the authors thought differently. Less than two weeks into my work on the book, they bought serverless.com and renamed the framework itself Serverless.
I was so annoyed when I found out. It’s like if back in 2005, upon being beset with complaints from people in train technology, DHH had renamed Ruby on Rails to simply “MVC.”
Within a couple of days I was able to rationalize away my annoyance by telling myself that any validation of the Serverless concept and name is probably a good thing. And I have to admit that the Serverless framework is probably the best way to experiment with API Gateway (APIG) + Lambda functions
While it would cost me some rewriting effort, it would still make me very happy if they rename the framework again.
