Valid Points
There has been some great point / counter-point going on in my aggregator today. I am not sure if the authors know it or now so I am going to bring it all together here.
A Valid Point
One of the first posts I read this morning was My Ruby Tipping Point on Igor Stoyanov’s blog. He was holding up the ActiveRecord validation methods in Ruby as the proper way to do validation in an application. the argument as I understand it is that the validation goes in the domain objects (Model) because that is where the data is manipulated.
In a post later in the day, Martin Fowler talked about Contextual Validation. I wonder if Igor’s post wasn’t somehow a provocation for Martin’s post (after all, they do work at the same place). Mr. Fowler cautioned against treating validation as a boolean. Right in line with yesterday’s Humane Interface post he suggests that some data may be valid for saving, but not necessarily valid for acting upon. This seems to be easy to implement in Ruby by setting the available validation methods. Both posts had valid points.
The Big Debate
Then there was the talk about Big today; big as in applications and scalability. It all started with a post on Om Malik’s blog entitled Even in Web2.0 Scale & Size Matter. Om’s post was a smack in the face to the idea of small startups solving today’s problem and growing organically as needed. However alot of people seemed to miss his point about planning and assume that he was talking about starting with a massively capable system of infrastructure.
Om tried to speak to his critics as he updated his post several times during the day. James Robertson pointed to a post by Dare Obasanjo that reminds us we cannot tell the future. You can have a plan for scaling and optimization but you cannot foresee tomorrow’s problems.
Both of these discussions are things that I might weigh in on soon, but I am enjoying listening too. These are very relevant to the work I have been doing on StatZen, especially as I consider writing the interface with Ruby on Rails. Do you know how to scale Rails? I have never done it before. Which is kind of the problem with what Om says. If you have been through a scaling process before then you can prevent yourself from making the same mistakes twice, but your plan is only as good as your experience and your luck.
