Nov 14, 2013, 9:04 AM
Ext.Net 3.0
I thought I'd open discussion on the future of Ext.Net. I realize you guys may already have a road map, but I thought maybe us developers might add some feedback on where we'd like to see things go... I hope other developers will chime in as well.
Users of Web 2.X and mobile applications expect to stay on a single page if their updating contents on the same page without refresh. Thank you Ajax. Refresh is becoming, more-or-less, a thing of the past. The only time you go to a new page is when completely moving to another logical area. This means more WebForms like infrastructure. Moreover, initial client-side load-time on each page is high because of the huge runtime cost to initialize all the extjs and extnet ajax libraries on the client. I realize ext.net does provide some building blocks for a desktop-like environment, but I really cannot rely on the modularity I expect if I were building one today. For instance, if I just want to throw away and update all the contents of a panel or a user control, I have to write 2 sections of controls or 2 user controls to update contents - one that manages initial load and one that works for DirectEvents/Methods.
The reason for this comes down to the fact that attributes/properties do not all behave like api calls during DirectEvents/Methods. Some do; some don't. Even the example here... https://examples2.ext.net/#/XRender/...UpdateContent/ ... doesn't do that much if the UserControl had a TextBox where it's Page_Load sets the Text property of one of it's internal controls because the child user control is instantiated inside a DirectMethod. To set the text, I have to call SetValue because it's technically inside a DirectEvent instead of ctl.Text="x" which only works during initial load. That's what I mean by 2 paths of code. Writing api's that dynamically and manually set values of controls to dynamic data is just as tricky. Again, I have to choose the correct method and check X.IsAjaxRequest to determine how to set values in controls.
1) I'd like to see a 3.0 that is smarter. It needs to decide based on whether it's in a DirectEvent/Method or initial load and render the appropriate code and consistently for all control attributes/properties. I'd like to be able to write one set of code that updates sections on the page and instantiate user controls that work in both modes, even if they need to render new initialized controls that contain forms with attributes that use TrackResetOnLoad.
2) I also want to be able to exploit the server-side javascript construction available in Direct Events to be able to pass the constructed javascript off to separate streams so that it can be attached to listeners and MessageBox confirmations in direct events. I don't want to be limited to X-Render either.
I know that's a lot but is this a possible vision for the future?
Users of Web 2.X and mobile applications expect to stay on a single page if their updating contents on the same page without refresh. Thank you Ajax. Refresh is becoming, more-or-less, a thing of the past. The only time you go to a new page is when completely moving to another logical area. This means more WebForms like infrastructure. Moreover, initial client-side load-time on each page is high because of the huge runtime cost to initialize all the extjs and extnet ajax libraries on the client. I realize ext.net does provide some building blocks for a desktop-like environment, but I really cannot rely on the modularity I expect if I were building one today. For instance, if I just want to throw away and update all the contents of a panel or a user control, I have to write 2 sections of controls or 2 user controls to update contents - one that manages initial load and one that works for DirectEvents/Methods.
The reason for this comes down to the fact that attributes/properties do not all behave like api calls during DirectEvents/Methods. Some do; some don't. Even the example here... https://examples2.ext.net/#/XRender/...UpdateContent/ ... doesn't do that much if the UserControl had a TextBox where it's Page_Load sets the Text property of one of it's internal controls because the child user control is instantiated inside a DirectMethod. To set the text, I have to call SetValue because it's technically inside a DirectEvent instead of ctl.Text="x" which only works during initial load. That's what I mean by 2 paths of code. Writing api's that dynamically and manually set values of controls to dynamic data is just as tricky. Again, I have to choose the correct method and check X.IsAjaxRequest to determine how to set values in controls.
1) I'd like to see a 3.0 that is smarter. It needs to decide based on whether it's in a DirectEvent/Method or initial load and render the appropriate code and consistently for all control attributes/properties. I'd like to be able to write one set of code that updates sections on the page and instantiate user controls that work in both modes, even if they need to render new initialized controls that contain forms with attributes that use TrackResetOnLoad.
2) I also want to be able to exploit the server-side javascript construction available in Direct Events to be able to pass the constructed javascript off to separate streams so that it can be attached to listeners and MessageBox confirmations in direct events. I don't want to be limited to X-Render either.
I know that's a lot but is this a possible vision for the future?
Last edited by michaeld; Nov 15, 2013 at 12:17 AM.
