-
☆ A M B ☆
- 3,141 Posts
From an addon development perspective that will be horrible. It will mean that anyone making CMPs will need to know ExtJS 3 (complicated enough), ExtJS 4, jQuery etc etc, and make a different package for everyone to work.
I’m all up for the default one using ExtJS (3 or 4, not both) and another less JS dependent one.. but no more! lol
Without knowing a lot about how the Manager is programmed,
I am assuming that the Manager will be totally rewritten in
the new library (no dependency on ExtJS).
Good point about the extras though, there would have to be
different versions... BUT... COMPETITION...
the manager that is easier to develop for would be the one that
will have most of the extras!!!
Website: www.mercologia.com
MODX Revo Tutorials: www.modxperience.com
MODX Professional Partner
From an addon development perspective i would simply ignore alternative manager interfaces and stay with the shipped one.
Quote from: splittingred at Jun 09, 2011, 04:32 PM
To be honest, the one with the most extras would most likely be the one that the Core Team supports.
Absolutely. That said, if you develop a better Manager that performs all the functions of the existing one (good luck with that), the Core Team might adopt it.
Quote from: Moondawgy at Jun 09, 2011, 02:13 AM
Being a MODx user since the start and creating over 100 sites in it - I can categorically say I will never use Revo until someone much cleverer than me creates a non ExtJS admin.
Clean HTML 5 with easy to edit CSS with a bit of Jquery and AJAX as the backend please.
I love MODx and the community, never, ever get me wrong about that. But trying to create a ’desktop’ like admin is overkill.
There I said it - I would even donate quite a lot IF you could find someone who could take the time to do this.
I’m very happy with the Extjs-driven manager in Revo. The core-team did and does outstanding work with it.
I think it would take years to develope a new manager from scratch for Revo.
But perhaps it would be somehow possible to port the evolution-manager to Revo and add Revo-specific features in a bit less time than years.
-
☆ A M B ☆
- 2,475 Posts
I’m still baffled by the dependencies here -- of course the manager will ship with some JS and some CSS, but I don’t understand why this should affect the CMPs in the way that it has.
I might sound like a dead horse here, but let me offer this comparison gleaned after writing a book about WordPress plugin development.
In WordPress, any plugin can register the Javascript/CSS files it needs and also any dependencies it has. So if the plugin uses a jQuery extension for example, it executes a couple functions that says that it needs to load up the jquery_extenstion.js file and that this file is dependent on the main jquery.js file. I can build a jQuery WordPress plugin and reference the included libraries, or I can build a JS Ext plugin and reference my own libraries that I include with my plugin. In other words, you shouldn’t need to write multiple versions of ANY of your CMPs: write it once, include the files you need, and be done with it.
I got nothing but respect for the MODx team -- you guys have done some really amazing stuff and I owe you many thanks, so please understand that the following comments are intended to be constructive and with all due respect.
The best metaphor I can come up with how the manager should work in is the old Volkswagon Beetle. The name "Volkswagon" translates to something like "The People’s Car" and it was designed to be simple to repair... Anyone with a screwdriver and a wrench could fix the thing. I’m not much impressed by the WordPress architecture, but holy smokes, it’s so simple to get in under the hood to alter stuff. This JS Ext/CMP thing right now reminds me more of working on a jet engine. I know MODx wants attract developers, but the manager is not much of a draw right now for junior or mid-level devs with their wrenches and screwdrivers because they simply don’t have the necessary tools.
The other problem here is that the manager still remains woefully heavy... in principle an Ajax-driven application should be heavy to load the first time, but the sub-requests made by Ajax should execute quickly. We’ve seen some improvement here, but coming to Revo from Evo or WordPress is still a shocker, and forget about it if you’re on a slow connection. I’ve had to move some Revo sites off of Bluehost shared hosting simply because Revo eats up too many resources (PHP memory primarily). So the reality of Ajax frameworks is somewhat different than how they are advertised: I’ve yet to see one that’s been able to outperform a bone-headed simple interface in a normal everyday-usage scenario. There’s a balance in there, I’m sure, but I feel the "correct" approach for a manager might be quite a bit skewed towards requesting regular light-weight pages than towards the heavy Ajax approach -- in layman’s terms, the Evo manager or the WordPress manager are much more usable.
-
☆ A M B ☆
- 872 Posts
@Everett
The manager’s heavy on the client, ajax or not, even if it was not built with ExtJS, the PHP memory would be the approximatively the same no ?
That being said, i agree with you, and even if i love Ext (I’ve been working with it before it came to MODx), the WP admin eat the MODx one.
For example, the process to create a custom TV is a bit cryptic compared to adding a custom field in the WP admin on each page without having to code one line of code.
Shaun pointed out that they did come with an hybrid solution where the manager was less dependant on Ext, and that’s what caused lots of issues in previous revo releases (2.0.x)
As they always says, if there is a community driven alternative they will probably welcome it...
I would sacrifice a little bit of functionalities for more easiness to deal with the manager.
MODX Staff
.