You forgot that there is software that does all the real work. We only have to use the software correctly.
I miss steam power. I had a small steam engine when I was a kid. It was pretty neat. But it was fueled with a solid sort of Sterno. When I ran out of that my uncle put some wood in it and turned most of it black from the smoke. Yuck! lol
Rick C.
Not people who have learned Forth.
Rick C.
If threats can be tracked, why can't they be blocked? The internet
I still have my Mamod steamroller (powered by a meths burner) in its rather faded box.
Haven't fired it up in decades.
Interesting observation of Java and C/C++.
The people that originated and defined Java repeatedly referred to lessons learned in many other languages. (e.g. see Gosling's 1996 white paper).
The people that define C/C++ repeatedly refer to papers about C/C++, as if it was the only thing that existed.
Such myopia is revealing and worrying.
There are no wrong solutions to an engineering problem, just solutions which don't meet the requirements. All things being equal it's probably best to look at the simple ones first - better is the enemy of good and premature optimization is the root of all evil.
At least then you have a baseline for how badly they fall short of the mark.
"we must destroy it, in order to save it" is an example of "better is the enemy of good." If you're okay with less than perfect destroying what you already have is usually unnecessary anyway, you can often get an 80% performance boost by changing 20% of "it."
Oh, I have no illusions that anyone will fix the internet security mess in my lifetime. With a zillion IoT devices and international espionage, things will only get worse. IBM and Intel and Microsoft messed things up too badly to fix any time soon.
I have some good ideas of how to make things better, but nobody here wants to discuss them, or suggest any alternates. This "design" group is mostly emotionally hostile to new ideas. Only a few people here are capable of brainstorming, or of serious electronic design.
So, call me an idiot, or suggest a better os and internet architecture, and we'll see which group you are in.
Also the whole "throw it away and start again" is not what I'd describe as a very "conservative" philosophy of engineering, either.
Not hostile to new ideas but "scrap it all and start over" is incidentally not what I'd call a very "conservative" philosophy on engineering!
ed
The problem is that pretty much everybody here knows more about the real wo rld that John Larkin, so his ideas about how the real world might change - an incipient ice age or the like - get treated with the contempt they deser ve.
Fred Bloggs - for instance - is quite insistent that the world is going to change, but his idea conflict with those of Anthony Watts, and John Larkin chooses to believe what Anthony Watts tells him (which is to say what the K och brothers pay Anthony Watts to tell to the gullible).
That does seem to be the mindset that confuses socialism with communism, an d fails to notice that the problem with communism was that it wasn't democr atic.
Joseph McCarthy concocted a narrative in 1950, and some people still take i t seriously.
John Larkin's idea of a "good idea" is one that he finds emotionally satisfying. Intellectual satisfaction does seem to be beyond his scope.
Your "new ideas" tend to be old mistakes that you have rediscovered. The hostility is to somebody who is wasting bandwidth because he hasn't done his homework.
And John Larkin isn't one of them.
John Larkin is an idiot. You don't have to be all that clever to be aware that there are lot of operating systems around but you do have to know quite a bit to be able to creditably claim that one of them might be better than the bunch we currently use.
The current internet architecture is a stunningly effective distributed system. It's complicated.
I know very little about it, but I do seem to know more than John Larkin.
Wooden pegs go back centuries, and shingled roofs were known in the stone age (no guessing required, stone-tool-using natives made canoes and totem poles and shingled roofs here before the first explorers showed up). I hang my bathrobe on a wooden peg, and have shingles on the roof.
I'm 'willing to consider' but not planning for change. I have the shingles, and the wooden pegs, NOW, doing excellent service. Obsolescence is not a generally true principle.
Negotiation between network designers, operators, users means multiple different ideas about what is 'nasty' will always be around. Negotiation between commercial software sellers, software designers, hardware designers, and users is even more fragmentiing. If the current situation is going to improve, tell me HOW.
Us old guys have all met someone who thinks that the soup is improved any time he pisses in it.
I don't see any real ideas, just complaining. Saying "someone should do something about the Internet" is not brainstorming.
Rick C.
And you think other people haven't already thought of and considered those?
It has been an active research topic for at least a quarter of a century, with the results you see.
Publish your ideas. I'll lay good money they have already been found wanting.
Just because you have some skill and experience in electronics, means zilch about your skill in software and networks.
Brainstorming is trivial, particularly when ignorant of the theoretical, practical and commercial constraints.
Implementation is "difficult".
See the work done on /interplanetary/ internet. They take security seriously.
I wrote three preemptive RTOSs and a few compilers; I independently invented and deployed Token Ring, maybe before IBM did. I'm not just a hardware jock, although I find hardware design to be more fun than coding. I built and managed timeshare systems that were a lot more secure than most of what's in use now. We sold time to some high schools who were serious rivals, and they spent time trying to hack one another. DEC did memory management right. So did Motorola. IBM picked the wrong CPU and the wrong OS source.
What's silly is than nobody here, except me so far, will even consider new ideas about internet security. Or new ideas about most anything.
Are my ideas stupid? Suggest something better. Playing around with ideas is how design happens.
I guess people who do have ideas haven't time for usenet.
Many of the problems with computer security were solved, but the world often succumbs to bigness over bestness. You see that effect in a lot of industries, where a giant, bloated, inefficient enterprise dominates, and makes billions, mainly because they are big.
My original suggestion, that the internet transport layer could apply security for everyone, is hardly a suggestion to scrap anything. I'd pay a ISP more if they scrubbed hazards at the network level. We already pay a DNS provider to not serve up suspected IP addresses. Some free ones now do that too.
Nobody else has yet suggested anything. Just complaining about anyone having ideas.
Nobody here, obviously.
Three RTOS, several compilers, lots of modem and protocol designs, original work in networking, maybe a million lines of code. But experience doesn't make ideas good or bad; ideas stand on their own.
In many, maybe most fields of study, there is sufficient group-think that experience eventually becomes a liability; nobody dares to question accepted theories. This group is mostly now people who don't do design and are reflexively hostile to ideas.
Brainstorming is trivial? Do you always work alone? And get all your designs from textbooks and appnotes?
Sure. That's why it's fun.
John Larkin stated
1) There exist NO 100% secure system, and never will. 2) Lots of work has been done to make the internet more secure.What ideas? have you even ever written a TCP stack?
My mistake.
I've never liked the 8088 (etc) family, always preferring others.
Having said that, the cache problems are a side effect of the speed-at-all-costs benchmarketing compulsions. They are an artefact of the implementation, and are not fundamental to the 8088 (etc!) architecture.
Better ideas that what? I'm not aware you have posted any concrete ideas here.
Beyond that, of course "extraordinary claims require extraordinary proof".
It isn't quite that simple, but yes, industrial/commercial pressures have been too dominant.
It is interesting to consider what, if anything, would roll that back.
I never mentioned the halting problem.
Multicore, with a modern architecture and proper hardware protections, is the obvious path towards secure computing.
c dumps too much danger into a shared memory space, and x86 doesn't provide enough tools to make it safe. What tools it provides aren't much used.
Buffer overrun exploits should be flat impossible.
Have something to add? Share your thoughts — no account required.
Ask the community — no account required