Muffins and Networking

This morning I made corn muffins from scratch, and I tried to figure out a networking problem. The muffins turned out well, especially with strawberry jelly, and the network problem is still intractable.

The problem is that our laptop and our desktop have lost the ability to communicate. I think that its because something is blocking communication, but its hard to tell, because the communication tool doesn't say 'I was blocked', let alone 'I was blocked by xxxxxx', but rather 'Unable to access...." and "You might not have authority....". No clear statement of problem, no clear path for resolution. Now as it happens I do like problems of this type, but what I really mean when I say that is that I really like when problems of this type are solved, and I've learned something in the process. I enjoy that. But its not actually all that much fun when the problem is still there. How do we get into these predicaments?

Because the tool is more powerful than the controls; it can autodrive itself most of the time, but when it can't, then you either live with it or you open the hood and start flipping switches and drawing conclusions. If you can; the conclusions aren't always obvious, and sometimes they're contradictory. (Are you thinking of Nomal Accidents yet? ) Put another way, communication between PCs needs five or ten or fifteen things to work, and they don't all have to work, just most of them, because they're redundant, and they don't have all yes/no values, some have if this is yes then if that one is no, fine, but if this one is no then don't check that one but go look over here.... and it goes five or six layers deep. There is no clear path of 'when this happens, look here, and here'. So it makes problem recognition, let alone resolution, an iterative process. And sometimes, the iterations lead you into the weeds. I swear I've passed that tree before !

So, since I know little about networking, I spent about an hour, trawling the net, looking for phrases that describe the problem I'm having, and cutting and pasting the text I found in various chat rooms and the like into a document. Now I'm going through the document, extracting the common suggestions, and noting the ones that contradict. (This is what the Acme AI will do, once its publically available for prices of less than a zillion dollars: Find the Literature. Read the Literature. Draw Conclusions. Summarize. Recommend. ) For example, most of the suggestions that mention 'workgroup' say that all of the computers need to be connected to a common workgroup, but one suggestion said no, thats not necessary. Well, is that person wrong? Maybe. Or maybe they're right, but for a specific instance or a specific computer or a specific problem. Prior to doing this research, I wouldn't have realized that. Another example: most suggestions say that the 'node type' for the computer should be Hybrid or Mixed or Peer to Peer, but not Unknown; however, some say that Peer to Peer is bad, don't use it, and some say that Unknown is actually okay. So, is this something to look at first, to look at eventually, or to ignore? Don't know, but it goes into a category of Things To Think About.

That's why they call it learning. And when its all done, and this jewel is working again (with luck, because of something I do, and how I hope that in the process of getting there I don''t break something else), I'll have a clear path : 'when this problem occurs, check this and this', and I'll have forgotten the other things. I'll summarize it, mayube draw some general conclusions. Which is how you remember the important stuff, but I think its also how you forget things that don't matter this time, but might, the next time. And I don't know how you fix that. Perhaps thats part of the biology of learning, like when the planaria do it? They hardwire the important stuff, chuck the rest.

Tracking down a problem is fun, but its not wildly fun.