> IE5, I think. At least as far back as IE4 IIRC. I'm not sure about earlier, as while I was around for them being in circulation I never used 1/2/3 extensivel

> IE5, I think. At least as far back as IE4 IIRC. I'm not sure about earlier, as while I was around for them being in circulation I never used 1/2/3 extensively. > Once connected, it started progressing even without receiving any packet And could climb up to about 80% before the first packet was received. It was an intentional con to pacify the user and it worked: I've know people convinced that IE was faster than other browsers because when they access a script that connects quickly but takes ages to send out the first byte of response IE initially looks like it is doing something sooner than [insert competitor here] that truthfully showed no progress until data started arriving. It meant that even if MS wanted to change the behaviour to be more truthful in later versions they would have an issue: the same people would make a noise about the new version being slower than the old. It adds a layer of interpretation to some user problem reports too: “it nearly completed, but then just displayed an error” could mean one of a wider range of things than otherwise.