If you're an electrical or computer engineer specializing in networks you should have enough experience to know that a single bent or corroded pin, or slightly non uniformly applied piece of solder can ruin you day.
If you're on site somewhere, especially somewhere remote, it's hard to know just how things will get messed up. What works in a lab is very different than after you've shipped it off some place and tried to get it to behave there.
Before you deploy a network you obviously test it in your own lab under exactly the same humidity, temperature, radiation exposure, altitude and personnel as for on site right? To what tolerance? You also test all of your backup equipment by having samples you store in exactly the way they're going to be stored at a live test sight, so you know what the probability is of something happening to them during storage?
Now we know single bit flip in an ethernet packet is just the sort of low hanging fruit of problems that we have network engineers for right? So I'm guessing you developed your own mathematically perfect CRC that you have published and that we should all use, to solve the 'low hanging fruit' of single bit flip errors? Just like a thread on nut and bolt right - you can take your perfect errorless network hardware, put in an aircraft, fly it to a remote island 12 time zones away you know you, with absolute certainty, that it will work 100% of the time? You should get a PhD and write articles about your techniques, the rest of us could really benefit from that.
Maybe you're not on the software side of things, but more hardware, say telephone twisted pair. Now as you know, the reason we twist pairs of wires is to prevent a signal on one wire from inducing a field on another. So I'm guessing you have some piece of equipment that can verify that all the twisted pair sets of wires you use are optimally twisted? What's it called?
Ok I'll stop being a snide asshole, unless I find out you're one of my former students.
You're right, that yes, good engineering is supposed to predict problems in advance and plan for them. You do as many tests as you can, and hope that you've figured out what problems will arise. Unfortunately, it doesn't always work perfectly, there's always some random error involved, that you have to cope with on the fly. On the ground I would say 'try another bolt', up in space, when you've got a dude in a space suit simple solutions become very expensive, time consuming and very risky. I used to do something very similar to network engineering as an on site guy, and problems that take 5 minutes to solve in the lab can take hours in the field. And think about the problem they're having they removed an old unit, and in doing so a bolt shaved. They don't, apparently, have spare bolts easily accessible for this. Now they have a tool that can apply more force to the bolt, but that could break the bolt, so rather than trying it (and it might work, and everyone is happy, and no news story gets posted on
Notice also how they seemed to have some idea what to do when there were shavings from the bolt - they tried to blow away the pieces with nitrogen - someone planned enough to figure carrying a can of nitrogen might be useful, but I suspect that's a tricky problem with gloves on where you can risk puncturing the glove.
Trying to work in space, and to a lesser degree underwater, is very much an exercise in trying to not make things worse - even if you think you have a solution to this problem you're better to not screw it up and wreck hundreds of millions of dollars in equipment or a bolt that probably several hundred if not several thousand dollars to even get there (a single 100g bolt would cost anywhere between 400 dollars and 4000 depending on what launched it there).
debra messing ayaan hirsi ali rachel uchitel new years eve party ideas strait of hormuz mars needs moms gary johnson
No comments:
Post a Comment
Note: Only a member of this blog may post a comment.