I had Netscape on my 486DX2-66 with a 33.6 modem. Win 95, along with ICQ, mIRC, some NNTP reader I can’t recall… You get the picture.
Everyone I’ve told this to thinks I must have been out of my mind. But for a period of time that I recall as months I had some sort of phenomenon where Netscape would stop loading a web page (could take 10s of seconds, you know) unless you MOVED THE MOUSE. Continuously. The animated “N” on would freeze and if you didn’t move the mouse the page would just be blank, or partially loaded. Move the mouse, it resumes. Stop moving the mouse, it stops. I used to have to move my mouse in figure-eights, cajoling the machine to not give up and keep downloading.
You’ll think I’m crazy, too. But when I share this story I keep hoping someone, somewhere had the same experience. And maybe, someone who knows what was going on will chime in on some obscure IRQ conflict in Windows along with some optimization used by Netscape in one iteration caused this bug for a brief moment in time.
Ahh…. I was there my friend. Similar setup, 486 DX4 100, USRobotics modem. I had the IRQ conflict. Me and my friend figured out how to change the channels by reading the mainboard‘s manual. I had to change some jumpers around. It was my first modem and I had never connected to the internet before.
On linux /dev/random will use inputs such as mouse movement to generate random data. If a program needs random data for something such as encryption it will seemingly hang whilst it generates enough. This isn’t good on servers without an active user so you configure it to use /dev/urandom instead. Perhaps windows had similar back in the day.
Ok, going to scream into the abyss here…
I had Netscape on my 486DX2-66 with a 33.6 modem. Win 95, along with ICQ, mIRC, some NNTP reader I can’t recall… You get the picture.
Everyone I’ve told this to thinks I must have been out of my mind. But for a period of time that I recall as months I had some sort of phenomenon where Netscape would stop loading a web page (could take 10s of seconds, you know) unless you MOVED THE MOUSE. Continuously. The animated “N” on would freeze and if you didn’t move the mouse the page would just be blank, or partially loaded. Move the mouse, it resumes. Stop moving the mouse, it stops. I used to have to move my mouse in figure-eights, cajoling the machine to not give up and keep downloading.
You’ll think I’m crazy, too. But when I share this story I keep hoping someone, somewhere had the same experience. And maybe, someone who knows what was going on will chime in on some obscure IRQ conflict in Windows along with some optimization used by Netscape in one iteration caused this bug for a brief moment in time.
Ahh…. I was there my friend. Similar setup, 486 DX4 100, USRobotics modem. I had the IRQ conflict. Me and my friend figured out how to change the channels by reading the mainboard‘s manual. I had to change some jumpers around. It was my first modem and I had never connected to the internet before.
Were you by chance running a proxy, even on localhost? Here’s a good description of that issue: https://news.ycombinator.com/item?id=29539106
That thread also mentions the Windows 95 requirement for randomization on mouse movement. A page you visited regularly may have been using this.
On linux /dev/random will use inputs such as mouse movement to generate random data. If a program needs random data for something such as encryption it will seemingly hang whilst it generates enough. This isn’t good on servers without an active user so you configure it to use /dev/urandom instead. Perhaps windows had similar back in the day.