winscape 2315

Very likely :-). I've been trying to use Unix terms but I think I'm getting the subsbreastutions all f***ed up.

Yes. PDP-10s called it exec mode.

Right. And these are usually instructions that "talk" to devices, controllers, and CPUs; IOW, all the hardware. On PDP-10s, these were CONO, CONI, BLKI, BLKO, DATAO, DATAI, CONSZ, and CONSO which deal with un-setting hardware conditions, block transfers at the iron level. These would affect bit settings in the APR, the PI, etc. Now I'm way out of my expertise because I've never written nor executed one of these instructions.

Yes. However, if this code has to issue machine instructions that can only be done at the EXEC (or kernal) level, then you have to allow any random user mode program to have intimate insertions into the power of the kernal. This was always a big no-no in our philosophy.

TOPS-10 usually invented a privilege that could be checked and was given by the sysadmin. This was another human process and a pretty effective firewall. "What a sysadmin giveth, the sysadmin can taketh away." Gentleman's timesharing worked quite well; peer pressure kept most people from making havoc.

But you're not out of your depth yet. You're not thinking about ISP nor computing services businesses. It is your business to provide computing services to your customers. What they do, such as their data and code, is NOT your business at all. You, as the owner of the gear and service providers, have no idea what is a "good" node nor "bad" node. Think about compebreastors using your computing service. Each considers the other "bad".

But you see, the biz has to go this way. Hand distributions of bits can no longer be done the way we used to do it--on tape and off-line. When the global economies and banking systems depend on rapid fixes done yesterday, there is no place for a 2-year window. That is how long it took us to start a major monitor developement to a customer receiving it in the mail. This doesn't include how long it takes a customer to use the shipped bits.

That is what is happening now because people have to find the drivers that have been fixed. Consider the games that need a bug fix to drive the video cards. Watch all the problems these gamers have. Notice how expert they have to become in order to just play a simple computer game. Extrapolate this learning and installation processes to real live businesses.

The gamer biz is a realistic sniff of what has to be solved and implemented emoticon makes guesstimate within the next 2-3 years. This pattern has been going on for the last 10 years.

Now place yourself as the owner of a business that provided computing services to businesses that use this data and human users who have to provide this data to do whatever. You, the owner cannot look at their bits or you will lose all your subscribers. But you say, "But I won't use their data for nefarious means." You cannot guarantee that one of your employees won't turn into a crook. Nor can you guarantee that one of your employees isn't nosy and enjoys snooping. Snooping seems to be a basic human trait. Nor can you guarantee that the business that buys you out will never use the data your customers have.

My trusted source cannot be your trusted source if you are using a compebreastor.

Good Grief!!! No customers should trust their OS vendor. Ours never did.

Yup. DBMSers. My office was next to a guy who could think well both ways. I was always in awe of the guy. He could switch hats with no troubles. JMF was never a compiler thinker. That doesn't mean he couldn't do the work, but he'ld have to be forced to since it wasn't his natural ability.

People who think this way tend to write task-based OSes, not timesharing OSes. For example, Cutler was task-based thinking and never could comprehend timesharing style thinking.

Right. Now you're getting the hang of thinking the way we had to in a development environment.

Yup. If you notice, a newsgroup posting can do a similar thing.

Nope. Not revision, isolation. Consider software that is used to manage the world's ICBM installations. That is useful code but it sure can't be on the internet wires. However, the national and international banking does have to be on comm wires. When copper becomes too slow, these networks will have to be wireless, which means any radio broadcast could f*** me data base up.

BAH