var a as int = 1
var b as int = 100
var dd as int

writeln(@servertime())
while a <= b
{
formfieldvalue("subform", "Field1", a, "Hello World")
a = a+b
dd=@formnewrecord("Subform")
}
writeln(@servertime())
 

Steve, if "a = a + b" and b starts as 100, does this ever loop? Also, are you setting one field and then creating a whole new record with each iteration (if it does ever loop).

If it is not intended to loop, and isn't looping, and it is taking 26 minutes to pass less than 500 bytes through the net, there is something very wrong with the net or how you are set up.

I assume you are using a table view subform. Are you also on a 100 Megabit ethernet lan?

If your lan is 100 Megabit (i.e.: 100,000,000 bits per seconds) and your wan is 475 Kilobit (i.e.: 475,000 bits per second) then the scale of difference would be 100,000,000 / 475,000, which equals 210.5 (or so). So your lan should be about 210 times faster than your wan. If it takes 10 seconds on your lan, it should take 2105 seconds on your wan. 2105 divided by 60 is 35 minutes.

If the subform is used primarily for data entry, I would advise using a form view. Form view, having to only deal with one record at a time, requires much less data be pumped between client and server. It is also much easier for data entry.

In 1.1.n, Sesame needs to keep each row of the table in synch with the result set on the server all of the time, meaning that mere movement from line to line requires a number of commands go back and forth. These are very small commands, only a few bytes each, but there are a lot of them, and with a typical MTU setting of 1500-ish, a network will send at least 1.5K bytes back and forth each time it transmits at all.

Normally this is not noticed, because it is still much faster than a human typing. But, because you are automating the creation, data entry, and navigation of table rows, it becomes apparent.

To mitigate this, in Sesame 2.0, the table is allowed to stay out  of synch with the result sets on the server until the parent record is committed. This eliminates a lot of traffic and creates larger less frequent blocks (much more efficient). This can be seen in the times listed above.