- From: Christopher Rolfe <christopher.rolfe198@gmail.com>
- Date: Sat, 4 Apr 2015 01:13:44 +0100
- To: Bobby Mozumder <mozumder@futureclaw.com>, whatwg <whatwg@lists.whatwg.org>
Hello, I've been watching this discussion and as a web developer, both backend and front end, I felt that it was probably worth adding my two cents. The concept of a built in SPA builder, which is essentially what this proposal is, is an interesting one. Personally I've used frameworks such as Angular and Ember before and I've found them reasonably straight forward to work with. They all vary so much, they are all so opinionated and to try and determine which opinion is correct would be a very tough decision. That's not really the point I want to go into though. Looking through the examples on github I find myself quickly confused even looking at the basic examples. We seem to be declaring models in the head and then using them throughout the body in a complex manner. The issue I have with this is we start to have a page that looks a lot like some weird form of XML, and as a developer who missed most of the XML generation I've only ever deal with JSON. Looking at the example I find it very confusing and from the looks of it hard to keep track of. If I as a programmer am struggling already how can we expect our Tumblr pro's to keep up with it when they create something really complex. Another issue to consider is this. The idea that MVC suits everyone has already been brought up, but what of this, the idea that someone tries to do something that you haven't considered? You haven't really mentioned much about how Javascript fits into this new system, this proposal isn't going to be the end of JavaScript. You've said that callbacks will be accessible for certain actions but it still seems a little raw. For example, is it possible to update values within a model from JavaScript, you've said you'll expand on that section later. Another potential consideration is SEO. How is the url affected when you click a link, you give an example of when an mref link is clicked more data is pulled in, how is the url affected? Would you use something like the HTML push state to change the url when mref links are clicked? Is there a way to define this etc. My last point I guess echoes what a few people have already said. What's the point of this, who is this targeted at and why. Is the average Joe really going to take their time to learn a complex form of XML/HTML if they aren't going to learn JavaScript. I think not. As I said I build websites myself and the majority that I build are not Single Page Applications, I'd wonder how many out there are and really do we need to change HTML to suit what I suspect would be a minority of sites. Good luck and I hope you can understand the concerns I've raised. *Christopher Rolfe*
Received on Saturday, 4 April 2015 00:14:30 UTC