RE: Re: date validate

midrange-l-request@xxxxxxxxxxxx wrote:

>   6. Re: date validate (Bill Meecham)
>
>I agree.  To lessen network traffic as much validation as possible should be
>done in the browser so that each event is checked.  There are at least a
>couple of ways to do that.

I was almost certain that this would be brought up sooner or later.


>Validation can be done on the server but I'd rather have it done as data is
>entered than validating every input on the form after submit.

I was happily agreeing on all sides and making mental comparisons to age-old 
debates about validation in DSPF DDS specs vs. business logic in programming 
and similar kinds of threads.

But web interfaces, I'm not clear on...

First, it seems to me that most uses wouldn't be high-volume data entry. A page 
is navigated to, a few clicks on pull-downs, a few values entered in entry 
fields, and <Submit> denotes end of input.

Network transmission of the web page itself to the browser plus the return 
values for the input form will happen on every transaction regardless of an 
error on input. (If no error validation in script in the browser, then the 
return values are sent when <Submit> is clicked. If error validation script is 
done in the browser and no errors are detected, the return values are still 
sent when <Submit> is clicked. And if there are errors, the return values are 
still sent when <Submit> is clicked after errors are corrected. And assuming no 
cancellation, which can happen whether errors are made or not.)

The _significant_ difference in network traffic seems to be at one of two 
points -- (1) the web page is sent to the browser and includes script for 
validation or (2) the server sends back an error indication and the user must 
resubmit after correction.

For the question of network traffic, it seems to come down to whether or not to 
send a larger page to the browser to include validation script instead of 
resubmitting after error notification from the server.

If errors are unlikely, then there's no good (network traffic) reason to send 
extra script to the browser; it's wasted on every access that doesn't have an 
error. And, perhaps, if errors are likely, maybe the pages need some redesign.

I know there are many, many reasons for browser-side validation and for various 
scripts. I'm mostly curious if the consensus is that network traffic reduction 
from error handling is one of them. I've seen some pretty complex scripts.

(I am NOT a web designer. Just looking for fundamental concepts of server 
access.)

Tom Liotta