peterbirks: (Default)
[personal profile] peterbirks
So, following advice, and because I know I needed to do it really, I downloaded service pack 2 for my laptop computer.

First time round, it failed to install properly. So, when I turned the computer on, I was told that it was unstable, and that I should uninstall it "or contact my administrator".

Curiously, my administrator was not on call, so I took plan 1.

This caused untold havoc. Somehow, over an hour or two, I managed to piece things back together. I even succeeded in installing Service Pack 2.

But then something odd happened. Although I could get online via my wireless connection, I couldn't get online via the Local Area Connection 4. I got the message "network did not assign a network address to the computer". Helpfully, it offers you a "Repair" button. Unhelpfully, this doesn't work.

So, of course, I googled it. One person had suffered a similar problem. The responses were Greek to me. It's like algebra at school. Or electricity diagrams. There's something about networks that defeats me. Then again, I think that they defeat most computer guys too. But they adopt a man in white coat syndrome, blunder around for a few hours, and eventually get it to work somehow. But the fact is THEY ARE JUST TOO FUCKING COMPLICATED.

I tried typing ipconfig after clicking "run", but the DOS screen just appeared and instantly vanished. No one mentioned that that might happen.

I should be grateful that the wireless works, I guess. And I am sure that there are simple reasons why the wireless works but the wired connection doesn't. What I'd like is a simple means to fix it that doesn't entail wasting six hours of my life.

And you know what. A pound to a penny it's something to do with Mac Addresses going wrong.
Except it will all be my fault, 'cos it's never the computer's. Oh no.


Update: Went into Vista computer. Disabled MacAddress filtering. Wired Connection from laptop now works fine. There's a surprise. Checked MacAddress matched National SemiConductor Corp DP83815 /816 /10/100 MacPhyter PCI Adaptor (WHY DO THE NAMES ALWAYS HAVE TO BE SO FUCKING LONG?) and it does. So why should the connection suddenly stop working? Because it's all shit, put together bny a million different monkeys in a million different places, so that one small change (well, a big change, upgrading to Service Pack 2 (I've worked out why they give them long names, it's so they can then givem them shorter names that no-one understnads, like DNS, and SP2) causes untold numbers of problems elsewhere.

So, I can disable macaddress filtering, and the laptop then gets assigned a network address. But if I turn macaddress back on, and then turn my laptop off and back on again, once again it can't assign a network address, even though the macaddress filter is the right sequence of letters and numbers. So it seems to me that my only option is to go back to the way I was before (turn off all filtering shit and let someone leech my network), or never use a ired connection on the laptop, or have to go into the Belkin hom thingy and manually turn macaddress off if I ever want to do it, remembering to turn it back on at the end. Isn't life great?



What a waste of man hours security shit is. Why not just trust in people's honesty? Or shoot them if they transgress? It's really a waste of a fucking life sorting out this shit.


_____

Date: 2009-02-08 08:23 am (UTC)
From: (Anonymous)
Maybe I'm just lucky. But I've had relatively painless time with my routers. Not having stuff from previous millennium might help too. Anyway, have you tried if wpa wep etc works with SP2?

Aksu

Date: 2009-02-08 08:26 am (UTC)
From: [identity profile] peterbirks.livejournal.com
I've never been able to get WPA or WEP to work. I just don't understand it. As I say, it's just too complicated for me.

PJ

Date: 2009-02-08 08:30 am (UTC)
From: [identity profile] peterbirks.livejournal.com
Upgraded it to Service Pack Three, but this didn't fix the problem. Indeed, I was worried that the problem might have extended itself to the wireless connection, but it hadn't -- it just pretended that it had by giving me the "unable to acquire" message before it started connecting, thus showing that it had, indeed, acquired. See, the things lie to you as well! Crawled into bed at 4am, feeling that I'd wasted an evening. Depressed by it all :-) Have parked Laptop away from me where I can't see it and am exercising thought-shopping. Anyway, I want the MacAddress thing to work. I don't want to say "oh, that doesn't work, I'll try this". Call it a character flaw.

PJ

Date: 2009-02-08 09:48 am (UTC)
From: [identity profile] peterbirks.livejournal.com
Latest update. Whole thing still driving me mental. Managed to find a logging thing on the router with the message:

Feb.08.2009 09:28:13 security:307997.736 Blocked Prot=6, 89.243.195.237:54030 > 81.156.112.188:60111, S Seq=-1177753200, Ack=0 -Disallowed Destinati

Now, you might wonder, why have they misspelled the word "Port"? Because Prot is definitely not a word.

Is your answer to this question:

(a) yeah, typical dyslexia among IT workers

or

(b) what a stupid question? Everyone knows that Prot is short for Protocol.

If your answer is (a) then welcome to rest of the world, if it is (b) you are probably an IT worker. I spent about five minutes hunting around for what "Prot" meant. It's just one of those sequences of irritants that IT people don't think about, and thus they are surprised when non-IT people get annoyed...



Anyhoo, even if you know that Prot is short for "Protocol", that doesn't do you much good, because all they give you is a number (in this case "=6"). This is so that no-one who doesn't work in IT will know what it means, so they keep their jobs safe for another year. I found somewhere that this is TCP. which I vaguely recall means Transmission something protocol. So it's telling us that the protocol is a protocol. Good.

Not that this gets me anywhere. The MacAddress is fine, but the router won't let that laptop in on the wired connection. I suspect that something went wrong in that faulty upgrade to SP2. What I may do is compare the ouput from CMD ipconfig /all (yes, you have to remember to type cmd in the run box and then the ipconfig /all in the Dos box created, typing the ipconfig /all in the "run" box -- which is what you instinctively do because all of the pages "assume" that you know that you have to type the cmd command. Oh, and don't forget the space between the g and the forward slash), from the office desktop with my laptop. I have looked at both and I can see that they are radically different. What I don't know is how to fix it.

Is there a site somewhere that explains all of the "real" english meanings of computer dialect? E.G.

IPConfig - short for Internet Protocol Configuration (I'm guessing here, of course)

10/100 - short for (someone told me once, it was meaningless jargon)

CMD - short for (I have no idea)

and so on. Some historical background would be sweet too, but I know that would be too much to hope for.

PJ


SOLVED

Date: 2009-02-08 09:56 am (UTC)
From: [identity profile] peterbirks.livejournal.com
I've solved it. Clearly the Service Pack Update does something which it didn't occur to me that it could possibly do, and which I have only just noticed now because I am finally waking up. It was both the first and last place I would think of looking (by which I mean, as soon as what happened enters the realm of possibility, it's the obvious thing to look for, but it seems such a weird thing for a new service pack to do, that it didn't occur to my mind, even though I'm quite a good lateral thinker....)

PJ

IT

Date: 2009-02-08 07:59 pm (UTC)
From: (Anonymous)
<
[Error: Irreparable invalid markup ('<doesn't>') in entry. Owner must fix manually. Raw contents below.]

<<doesn't do you much good, because all they give you is a number (in this case "=6").>> Now I'll admit I havent a clue about the subject, altho' you're rant is quite gripping. Shame the number wasnt "=42"
Keith S

Date: 2009-02-08 11:04 am (UTC)
From: (Anonymous)
I'm going to take the easy option here and just blame Microsoft. Why? Because I'm lazy and noone is going to seriously defend them, but also because it is mostly their fault. Windows was just allowed to grow exponentially and multi-dimensionally with the inevitable mess that they now try to pass off as an operating system. The optimum size for a development team is probably something like 6-20 actual coders so once you're up to thousands it's not surprising that the right hand frequently doesn't know what the left hand is doing. What about the 'drivers' defence as recently advocated by Ballmer himself - it's not us, it's those evil hardware manufacturers and their shoddy drivers. Well it's true that poor drivers do cause a lot of problems but the reason so many drivers are poorly written is that Microsoft does not believe in setting standards/interfaces for Windows drivers to conform to. Indeed they have actively and vigorously tried to prevent the adoption of open standards in all areas. Their own specs are reverse-engineered to vaguely fit whatever random combination of version and service pack was in vogue last year and often actually kept secret from hardware manufacturers who arent giving them large kickbacks.

matt

Date: 2009-02-08 11:17 am (UTC)
From: [identity profile] peterbirks.livejournal.com
Hi Matt:

I've had a think about this, and I've concluded that things would be much easier if, when something failed to work (as it occasionally will inevitably do) the error message generated could be:

(a) more specific about where the error has occurred, and
(b) in plainer English.

Frequently the error messages are such generic shit (how many times do you get the "contact your administrator" garbage?) that the possible causes are too numerous to mention. In this particular case, even though I googled ALL of the error messages (from Windows, from the Belkin router), and about a dozen different solutions were offered, none of them was the right solution.

So, basically, what is needed is a better testing mechanism/diagnosis system when a flaw appears. That testing mechanism/diagnosis system NEEDS TO WORK. So, instead of the machine telling me that "network did not assign a network address to a computer", a system is needed to work out why that happened and, if you need to do something, to tell you, in plain English (not "Prot=6" shit), how to do it, rather than to "contact your administrator". I refuse to believe that this is not feasible. Computer people still in their heart of hearts, are in mainframe land rather than consumer land. It's still white coats, a special part of the building, and your own arcane geek talk. It would be nicest if things didn't go wrong, but, since they will (see Microsoft) it would be nice if you weren't assailed with geek-talk at every turn.

PJ

Re solved again

Date: 2009-02-08 11:21 am (UTC)
From: [identity profile] peterbirks.livejournal.com
I'd also point out that, even though I might proclaim the opposite, I am vaguely computer-literate, although netowrks continue to confound me. No "ordinary" computer user would have been able to solve this problem, not in a million years. And yet it seems to have been caused by a major, presumably flaw-free, Service Pack upgrade. I assume that something happened when the computer decided to switch itself off half-way through a 90-minute process (the Service Pack Three upgrade took much longer than that -- I don't know how many files were installed, but given that it took an hour and it was shooting through at about 10 files a second, it's a large number) which caused this glitch. If it's something that happens automatically, well, it's dreadful.

PJ

Re: Re solved again

Date: 2009-02-08 11:34 am (UTC)
From: [identity profile] peterbirks.livejournal.com
I see in my original post that I say that I checked that the Mac Addresses matched. In fact, they didn't. The first pair of digits of the physical address of the National SemiConductor Corp DP83815 /816 /10/100 MacPhyter PCI Adaptor had changed as a result of the Service Pack 2 upgrade from 02 to 00. I eagerly await the comment that "that isn't meant to happen".

PJ

Re: Re solved again

Date: 2009-02-09 09:19 pm (UTC)
From: [identity profile] real-aardvark.livejournal.com
Well, actually, the correct comment is "You really should have looked more closely, shouldn't you?"

And before you turn purple and explode, I make exactly this mistake, all the time. And, being an "IT worker" (please stop using this phrase. It will get you fired, like Carol Thatcher), it happens far more frequently to me. Pain in the arse, isn't it?

Fundamentally, and I'm sure Taleb has a relevant quote on this, human beings just aren't wired to use computers. Your case proves this in two ways:

(1) You (ie we) are apt to discount the possibility that, having run through an initial mental check list, we may have made a mistake. Computers have the time to do this sort of thing over and over again. We don't. Result, something that "just couldn't happen" (and which the image-processing bits of our brain tell us didn't happen) ... well ... happens. Thus the weirdo masking of 02 to 00, and the time it took you to notice it.
(2) You (ie we) take mental short-cuts with a label attached. The invention of the GUI was a catastrophe in this regard. You (ie we, ie I) associate an icon with a result, ie in this case "show me the fucking configuration." This is disastrous. You (I me my) need to associate an icon with a process, ie "I wonder what happens if I press this?" The answer to this question here is "it probably launches a DOS box." In which case, do it yourself.

Note carefully that I agree with every step you took and would have suffered in exactly the same way myself.

As for (1), well, a small minority of people who devote their lives, zen-like, to this rubbish are capable of overcoming their synapses and acting like a forensic scientist at all times. 99% of IT people do not fall into this category. I'm not sure I do. I'm damn sure users shouldn't have to.

As for (2), it largely falls four-square into the category of "Matt is right here." Windows isn't, contrary to popular "IT" opinion, a heap of shit. The fact that it works at all is pretty magical. It is, however, designed, implemented and tested by a rotating crowd of socially retarded baboons. Monkeys and Shakespeare don't work, but baboons and GUIs almost do. All large software, as Matt suggests, is like this; Operating Systems even more so. Linux appears to be going this way.

Yes, I know, not very helpful. Sorry. But I can enlighten you as to why physical devices, and network devices in particular, have such ludicrously long names. It's because they're all designed to report info through Simple Network Management Protocol (SNMP), which concatenates every single bit of potentially useful information into one dirty great string. (I simplify: see RFC 1157 for details). Except the bits that it doesn't, because they're hidden somewhere else.

In other words, it's not designed for users. You'll be delighted to know that it's not designed for "IT" people, either (except that small subset, like me, who work with SNMP). It's designed for computers.

Yup, I'm being tremendously useful again.

Re: Re solved again

Date: 2009-02-09 09:31 pm (UTC)
From: [identity profile] real-aardvark.livejournal.com
Oh, and in our heart of hearts, we are not not not not oh so not in mainframe land.

Our bosses are. The salesmen are. Generally speaking, the technical writers are. Anybody on a help-desk is. You will only find a disenchanted freak in the "Systems Administration" area -- the name rather gives it away -- who is not.

These people are not "computer people." To paraphrase Mr Vicious, "They're not good-lookin', man -- they're somethin' else."

Actual honest-to-goodness computer people like simplicity and like tools that allow somebody else (for the love of God, not me again) to figure out the solution for themselves (better still, have the problematical software auto-correct itself). As Matt says, Windows ain't built with this philosophy. The "wizards" would disgrace Saruman after the fall of Mordor.

Windows is, essentially, built to appeal to people who are "not good-lookin', man -- they're somethin' else."

And it was marketed, very effectively marketed, at precisely these people. It satisfies their perverse needs. The wonder is that normal human beings assume that it was designed with them in mind.

Re: Re solved again

Date: 2009-02-09 10:41 pm (UTC)
From: [identity profile] peterbirks.livejournal.com
Hmm, I think I see what you're getting at...


No, it was an illusion. I don't. It's all lambbda shashlik still.

Believe it or not, I had no memory of checking the MacAddress the first time round. But I must have done so, because I wrote that I did it.

And, yes, I must have done it in one of those cursory fashions (see "Ways Of Seeing" by Berger or "Ways of Fooling" by Penn and Teller) that meant I saw what I expected to see, rather than what was there.

Because, well, I kind of thought a Physical Address was a Physical Address, like a house number. Well, it IS like a house number, and Windows is clearly the local uber council, which can say that you now live at number 8 rather than at number 2. But, hell if you visited a house that had been number 2 for all the time you remember, and someone put a small "8" on the door, would you see it? I suspect not.

I assume that this is what happened in my too human brain. I honestly did not know that Physical Addresses could be changed. I thought they were immutable, which was one reason for them being so long....

Thinking about this, it's nonsense. Even though the hexadecimal number is one fuck of a big number, there are probably more physical devices in the land of computers, so that there is a finite, albeit remarkably small, chance that, if physical addresses were like fingerprints, you could get two with the same Address in the same machine. Ergo, they must be alterable. Ergo, if it can be done and by doing so you, the user, will get fucked, Windows will fucking do it.

PJ

More songs about girls and dirty mac addresses

Date: 2009-02-10 09:49 pm (UTC)
From: [identity profile] real-aardvark.livejournal.com
Um, it's both more and less complicated than that. (Surprise!) 281 billion (that's English billion) is quite a capacious range, and the basic theory behind RFC 824 is that it's all we'll ever need. I mean, it's about 40,000 per person. Even if you live in Darfur.

So (goes the theory), the IEEE issues chunks of this range to anybody who wants one. If you stare really hard at your MAC address, there are probably ten or eleven bits that shout out "I'm a Sony!" Yes, it's (in theory) burned in to the hardware at birth.

Of course, this potentially leaves a few empty ranges here and there, but it shouldn't really matter -- what the heck, Mr Darfur is limited to only 10,000 personal laptops or so. It is possible to "spoof" a MAC address when it hits the wire -- you can spoof anything when it hits the wire -- but you'd have to be insane to try. So that's all right, isn't it?

Except.

The IEEE, in its infinite wisdom, can't keep its sticky little fingers off a nice simple system like this. I'd love to blame what you saw on Windows, but I can't, really: I have to blame the IEEE. Because:

It turns out that one of those forty eight bits has a special function. Seeing as how you were diddled by a switch between 02h and 00h, you will immediately guess which bit that might be. Yes, it's our old friend

"the second least significant bit of the most significant byte." (It says here.)

What is this special function, you may well ask. And you want the answer in plain English. You have a right to have the answer in plain English, god damn it. So here it is:

"If the bit is 0, the address is universally locally administered. If it is 1, the address is locally globally administered." (It also says here.)

In other words, when your address is 02h etc, you are locally globally administered. When your address is 00h etc, you are universally locally administered. I really can't make it any plainer than that.

(No, I don't have a clue either.) This stuff only applies to local area networks, but, unfortunately, your wireless router is a local area network. SP2 probably thinks that the new setup is "more secure," for some reason I do not propose to burn grey cells over.

Yes, all that was gibberish. Accurate gibberish, but gibberish. But ponder this: since this is a universally known standard, why is it that the stupid MAC address filter on your stupid wireless router doesn't just ignore the stupid special bit, thus restricting the world to only 140 billion possible network devices?

Damned if I know. I blame the Chinese, myself.
From: [identity profile] peterbirks.livejournal.com
What's IEEE? In fact, what's RFC?

I had observed that three of the four MacAddresses already began with OO, which would indicate that this was something in addition to the "unique MacAddress". I have no idea that locally globally administered is and how it differs from universally locally administered. I note two points:

1) "globally" is roughly synonymous with "universally", unless you are on the moon or are another girl on another planet.

2) "locally" tends to be inconsistent with "globally" and with "universally"

3) In Britain, the order of adverbs is not as a rule that important. "happily quickly achieved" is an equivalent of "quickly happily achieved".


Therefore 00 is locally universally administered and 02 is locally globally administered. I am on the earth, which means that, for me, these are the same.

Except that, in computer world, they are not. And they wonder why we can't understand them.

PJ

August 2023

S M T W T F S
  12345
6789101112
13 14151617 1819
20 212223242526
27282930 31  

Most Popular Tags

Page Summary

Style Credit

Expand Cut Tags

No cut tags
Page generated Feb. 24th, 2026 03:55 pm
Powered by Dreamwidth Studios