Optimize web container controls

Container controls offer tremendous value to developers. Their ability to pre-package groups of controls, especially labels and fields, promotes consistency throughout an application’s UI. We’ve integrated them extensively across our web application because they significantly expedite the development process. The feature where controls can be nested within other controls is especially commendable.

However, there’s a significant challenge we’ve observed: performance degradation when multiple container controls are employed on a single page. Using more than three or four on a page results in noticeable lag. In several instances, we’ve had to revert our design, breaking down the container back to its basic elements to achieve an acceptable load time.

I acknowledge and appreciate Xojo’s mission to ensure uniform functionality across different platforms. While development consistency is important, from our experience, the efficient performance of complex container controls in Web2 is currently lacking and thus is not consistent real world cross platform development. (Essentially, you can develop the same way, but it won’t work when you deploy it)

To address this, I’ve initiated feature request #73570 . I kindly urge Xojo Inc. and Geoff to consider dedicating resources towards optimizing the performance of container controls. Such an improvement would not only benefit our team but also countless other developers who leverage these controls for their web applications. Please vote it up.

Geoff is too busy spending “extensive” time on AI.


While I wouldn’t expect Geoff to be writing code for the IDE or frameworks etc a lot of time spent poking at AI definitely suggests a lack of focus on the things that are impeding users ability to use Xojo effectively

Like in Jays case where getting someone to focus on WebContainers would be a win for everyone using Web projects
Or similar issues in desktop, iOS etc

Getting an AI to generate better code for Xojo is actually something anyone could do - just train one with known good code*.

But fixing frameworks etc ISNT something anyone else can do - only Xojo can fix those things
I wish they would

EDIT : added italics - exactly how do we know that code is “good” or not ?
Who is the arbiter ?
If the AI is trained from Xojo examples then will it produce “example level” code; which isn’t always great quality. Thats not an examples point. Its meant to be an example; to demonstrate a point and so skips good habits like error checking etc

So somewhere xojo needs to get good code to train this AI


Wow, they actually include it in the 2023 r3 milestone.

Can you ask for the control arrays? :laughing:

Include in milestone doesnt mean it gets FIXED in milestone
But it is better than nothing at all