Skip to Content
Feb 18, 2006 at 11:34 AM



I have a question related to BSP load times.

Our users have been complaining about poor load time performance during the initial start up of bsp applications.

We have a few custom BSP applications that render lists of data collected from various tables inside our backend. The selection pages for these typically include a drop down for org, and some date parameters.

None of the selection pages include a requirement for back end processing. All of controls on the selection page could be built with straight html or with htmlb logic. We used htmlb controls for our buttons and dropdowns.

We have noticed a performance lag with the initial browser load of the htmlb page over a straight html page with a similar set of selection controls.

We looked at the browser source of the htmlb. There is a fair bit of additional stuff that gets called when using the htmlb option. I suspect the server side compile, processing, subsequent render of data, download of attached javascripts and stylesheets is the reason for the delay relative to html load times.

Taking it a step further, we added all of the header links including style sheet links and javascript source links ( though not all of the control function calls or hidden fields) to an html page and compared the two load times of the this page with a similar htmlb page.

HTML was still seemed to load faster. I would expect this to be the case. Since the html approach requires less server side processing than the htmlb approach.

I have done a fair bit of development server side web development using other technologies. The general rule has been to minimize server side processing where possible to reduce server load. Is this an approach that has been adopted elsewhere when using BSP's?

What is the recommended approach to BSP development? Server side or client side?

Do most developers stick with BSP from end to end in their applications?

Any insight is appreciated.