HOME

Current View

VIEW Home
MAIL Home
BOOK Reviews
 
©
This week:
Monday
Tuesday
Wednesday
Thursday
Friday
Saturday
Sunday
 


 
Top

CHAOS MANOR MAIL

A SELECTION

MAIL October 11 - 17, 1999

REFRESH/RELOAD EARLY AND OFTEN!

CLICK ON THE BLIMP TO SEND MAIL TO ME

The current page will always have the name currentmail.html and may be bookmarked. For previous weeks, go to the MAIL HOME PAGE.

Fair warning: some of those previous weeks can take a minute plus to download. After Mail10, though, they're tamed down a bit.

IF YOU SEND MAIL it may be published; if you want it private SAY SO AT THE TOP of the mail. I try to respect confidences, but there is only me, and this is Chaos Manor. If you want a mail address other than the one from which you sent the mail to appear, PUT THAT AT THE END OF THE LETTER as a signature.

PLEASE DO NOT USE DEEP INDENTATION INCLUDING LAYERS OF BLOCK QUOTES IN MAIL. TABS in mail will also do deep indentations. Use with care or not at all.

I try to answer mail, but mostly I can't get to all of it. I read it all, although not always the instant it comes in. I do have books to write too...  I am reminded of H. P. Lovecraft who slowly starved to death while answering fan mail. 

If you want to send mail that will be published, you don't have to use the formatting instructions you will find when you click here but it will make my life simpler, and your chances of being published better..

This week:
Monday -- Tuesday -- Wednesday -- Thursday -- Friday -- Saturday -- Sunday

Current Mail

HIGHLIGHTS:

• The ghost in the machine
  • Scrubbing has its perils too.
• Bad memory
• Ethernet speeds; ECC
• Remote control
• File sharing trick
• A debate on my Intellectual Capital column
• Should we subsidize internet access?
  • On Universal Access
• Query about Dual Pentium Socket 7
• Gone into the dogs. Or vice versa.
•  
•  
•  

©
Monday
Tuesday
Wednesday
Thursday
Friday
Saturday
Sunday
 
Top

Monday October 11, 1999

There will be a great deal on this, some included more because it stimulates comments from me than for its content, but this is informative. First see the original statement of the problem if you haven't already. As usual I read and post subscriber mail first...

Dear Dr. Pournelle:

Regarding "Squirrel" which seems to hang on like Jason in the slasher movies; assuming that you retained the same hard drive without reformatting it, Windows retains the network settings in the registry (or somewhere else) regardless of whether you have a NIC or network connection. You may not see it when you don't have a NIC, but it's still there. I've personally experienced it.

Are you changing the system ID in the network setup utility in Control Panel? What does it say in the "Identification" tab when you do <start><settings><control panel><network><identification>? Does it say "squirrel" or "seattle"? If the wrong entry is there, change it.

The other [small] possibility is that you have "squirrel" still in your hosts files on the PCs. Hosts has entries for hostnames and IP addresses for systems on the network. It should be in the Windows directory. When you install Windows, it automatically creates a "hosts.sam" file. This serves as a template for your actual hosts files, and is copied to "hosts". It tells the network how to find systems by their names if you aren't using DNS. Now that I think of it, there's another similar file, called "lmhosts". That relates directly to the SMB settings. Try checking for it in the same place, and make sure that there's no reference to "squirrel" in it.

Hope this helps.

Good Luck

Chris Morton Rocky River, OH

Naturally I have changed the "Identification" and the description in that ID as well; and I used regedit to search for the word "Squirrel" in Seattle's registry. Nothing there. And I have removed the NIC (Network Interface Card) from Seattle, shut him down, brought him up without the card, removed all traces of network software including Microsoft Network Client so that the Network category vanished entirely, shut down, restarted, shut down -- reinstalled the NIC lo! it knew that this machine was named Seattle. Everything went fine. No trace of Squirrels. But the ghost remains, all my systems can see a mcahine named Squirrel which doesn't exist.

One problem was that Seattle (the original Squirrel) was so fouled up that it needed a pretty complete brain scrub; I removed all cards, removed all devices, and brought it back up. Because "I do all these silly things so you won't have to..." I did NOT reinstall Windows 98, which I probably would have done had I not been curious. Seattle now runs fine, fast, crisp, without all the dragging gubbage that caused Dobbins and I to gnash our teeth and wasted most of his time here. Had that not been needed we'd have had a look at the servers.

It's pretty clear to me now that the name "Squirrel" and the system description are tied somehow to the NIC ID in the server, so when the server see that NIC it assumes that Squirrel is on line; but it's also smart enough to see that Seattle is also on line. Interesting, and I think that's a flaw in the networking system worth learning more about.

I was also pleased to see that Mr. Dobbins, who knows a lot more about Microsoft NT networks than I do and about as much as anyone I know, hadn't encountered this problem before, so at least when I do something stupid I do it uniquely...

Jerry,

Let me preface this by saying "I'm not even 'close' to being any kind of expert with Windows networking and NT" but..

I think the problem you're having with squirrel showing up in network neighborhood has nothing to do with the 'squirrel' hardware, it's an "NT" thing.

I think I read awhile back that you set up a Windows NT 'domain' the last time you re-did your network. As I understand it, when you do that the 'domain controller' learns about systems on your network and remenbers them by name, associated with the MAC address of the ethernet card.

When you click on network neighborhood on a Windows machine, in a Microsoft 'network' with a domain controller, the 'local' machine doesn't build the network machine list internally. It gets it from the domain controller. In your case, the domain controller is still remembering squirrel based on the MAC address of the ethernet card that was in it.

We run into this and associated problems, often, when we try to re-name a machine or upgrade an ethernet card in an existing machine.

There is a 'way' in the Domain controller to delete the old machine. I don't know exactly what it is and don't have access to the local domain controller to play with it, but I'm sure that the domain controller is where you have to delete squirrel. After it's deleted there (and possibly the controller must be re-booted) then you can re-use the ethernet card in another machine.

Hope this helps. If you aren't able to figure out how to delete squirrel from the domain controller, let me know and I'll consult with our NT 'guru' for a set of instructions.

regards John

John Rice coredump@enteract.com http://www.enteract.com/~coredump 

The Internet - Somebody's LAB experiment gone horribly wrong.

Yes. I am convinced that I need to get into the Domain Controller and delete that name also. I am not sure I know how to do it...

Bob Thompson has this to say (sorry I am not editing mail so you will get some redundancies, but I haven't time, and I do find this interesting.)

> I have removed the Ethernet Card from Squirrel, removed all traces of ALL network software, powered down, brought up the machine with no network card. But as soon as I put a network card back in Network Neighborhood sees Squirrel on all the machines. Of course you can't access it. But it's there.

My guess is that you're running WINS and that it's mapped the Windows Computer Name (NetBIOS computer name) of Squirrel to that address. You can check WINS Manager and delete any mappings for Squirrel. It may also be something weird going on with the Master Browser. A couple of suggestions:

1. If you're using the same IP address for Seattle that you did for Squirrel, try changing it. That should solve the problem if it's related to computer name -> IP address mapping. If you're using DHCP to assign IP addresses, you can do a quick test just by going into Network Properties on Seattle and changing from using a Server assigned IP address to a manually specified one. Make sure that any address you assign manually is not already in use. If Mr. Dobbins set up the DHCP Server for you, I'm sure he excluded a range of addresses from DHCP Server. Use one of those excluded addresses, which you can find by firing up DHCP Manager on the DHCP server.

2. If you're using the same Ethernet card, try swapping in another one. That should solve the problem if it's related to computer name -> MAC address mapping.

Robert Bruce Thompson thompson@ttgnet.com http://www.ttgnet.com

The IP addresses are assigned by the network system; clients get a lease on certain numbers for some period of time that I don't recall. The only ones assigned manually are the servers and the printer; the rest are dynamic.

But I think the problem will be solved by doing what you say in your second paragraph: remove the NIC, and kill all references to it, and replace it with a different one. I'll reserve the current NIC for when Squirrel is revived (as he will be when I get a new MSI 6147 board, details on THAT problem in the column and I have promised not to duplicate the column here on the web site). I suspect that changing NIC will kill off the problem, and it will be interesting to see if another machine, using that card, will then be recognized as "Squirrel".

>>But Squirrel lives on in Network Neighborhood, and cannot be eradicated. The machine no longer exists, although his hard drive lives on in Seattle; but Squirrel will not die.

Jerry,

This problem (in Windows 9x anyway) is because of the way that Windows caches the names of the computers. The computers will announce themselves to the network and the computer that has been elected at the master browser will keep a list of the machines on the network (to speed up the display of the network neighbourhood).

When a machine disappears from the network the master browser will, after a period of time (66 minutes?) drop the name from the list. Did you wait that long before deciding that Squirrel won't die? If you have, then perhaps you are experiencing a new windows oddity.

The other solution is to shutdown every machine on the network and then restart them all. (ack!)

- Paul pdwalker@quagmyre.com

Waited a lot longer than any 66 minutes. And shutting down every machine isn't something I am anxious to do. (Shutting down various work stations does NOTHING; I can reset him until the cows come home and Parsifal still sees Squirrel out there. So does Princess, who has been reset with Seattle both on line and off...)

Jerry,

You said, "I have removed the Ethernet Card from Squirrel, removed all traces of ALL network software, powered down, brought up the machine with no network card. But as soon as I put a network card back in Network Neighborhood sees Squirrel on all the machines. Of course you can't access it. But it's there."

Is the network just seeing the NIC's identifier? Or does it still happen with other NIC's than lived in Squirrel? I don't understand the parameters...

Troy Loney

I think the NIC serial number is registered to the name Squirrel in the server; the server also sees that NIC as Seattle, and it's smart enough to process demands for Seattle resources and files, without being smart enough to kill Squirrel... So at some point I'll simply replace that NIC and see if a new one still keeps Squirrel alive as a ghost (he ain't even a zombie because he doesn't DO anything but exist as a name...)

  Scrubbing your machine can have its perils too, as Ron Morse observes:

>>The moral of this story, other than "I do all these silly things so you won't have to..." is don't try to save time: if you >are starting up new machine, scrub it down and start over, or start with a new hard drive...

This has its own perils. This weekend one of the drives on the desktop machine started developing bad sectors...discovered because Norton's Disk Doctor kept hitting one of them and locking the machine tight. Apparently a disk doctor file was corrupted...so I got a new IBM 18G IDE drive, saved all the important stuff I could recover from the old drive and decided now was as good a time as any to reinstall W98SE.

Every previous Windows install I've ever done on this machine has been 100 per cent problem free. This time I had a number of resource related conflicts (mostly related to the SoundBlaster Live) and it took all day to get it right. The drivers (which had happily installed before) simply wouldn't load, claiming insufficient space on target drive where there was damn sure sufficient space. One of my games (Terminal Reality's FLY!) won't install again despite the fact it worked well yesterday. Alas.

There's an inverse application of Moore's Law at work...every configuration change will take twice as long, cost twice as much and leave the machine working less well than before.

Now to try and figure out what's making Galahad unhappy. It thinks tojerry/baen is out of string space. I thought Baen collected string.

Regards

Ron Morse

I hope you are wrong, but I expect you're right. I have found SoundBlaster Live! to work well with new stuff, but choke up on legacy system games. Not always, but sometimes. Good luck. I fear your inverse to Moore's Law may not be a joke.

©
Monday
Tuesday
Wednesday
Thursday
Friday
Saturday
Sunday
 
Top

Tuesday, October 12, 1999

My son and I built a computer (for him) three months ago. We did everything right except one thing. It has occasionally behaved erratically, but last week ceased to boot. It kept trying to restore the damaged registry. (Sound familiar?) It had to be the memory.

You recommended Crucial memory. I ordered it (128K, PC100). My son installed it, then we both spent three days trying to make it crash. No luck. Only when we tried to run seven (7) programs simultaneously did it start to drag, but it didn't drop.

I'll never buy cheap memory again!

Thanks for putting light on the subject of memory; few of us realize the difference that quality memory can make. (Though my own memory has been failing with age [since about age twelve, I think], my computer memory doesn't have to.) You have done us all a great service.

Duane Toole

Thanks. You're fortunate: flaky memory can disguise itself as something else far too often.

Mr. Pournelle, you mentioned in your latest Byte column that the 100 Mbit/s ethernet do not give you ten times the performance of 10 Mbit/s ethernet. That's not entirely correct. You can achieve that performance, though a number of factors can effectively prevent you from saturating all the bandwidth available.

The most straight-forward limiting factor is, of course, plain hardware. One might mistakenly assume that it shouldn't be hard for the computer to saturate a 100 Mbit/s ethernet, given that hard disks transfers are higher by many orders of magnitude. Alas, it is much more difficult to send data through the network, because of the way network protocols work.

This can be easily tested. Benchmark the network on a computer and then bechmark it on a faster computer of similar configuration.

The impact the hardware plays is heavily dependent of other factors, though. First, how the Operating System and the Application work, and how their strategies interact. As the competition for the internet server market grew, for instance, many operating systems incorporated what is known as the "sendfile api". It's an operating system call specialized in transfering files from the disk through a network connection. The more streamlined the call gets, greater the improvement for applications such as ftp servers, at the cost of supporting more complex file transfer protocols.

Second, the network card itself can have a lot of impact, depending on how brain-damaged it is (or isn't). Some card manufacturers seem to be intent on producing cards that can send a packet extremenly fast through the network, only to lose all the performance gained in setting up for the next transfer.

And, of course, the network driver is also of great importance. It must take advantage of whatever the operating system and the network card offer, or otherwise it will be stealing processing time from the main cpu.

So, if you are not satisfied with the performance of the network, you might want to play around with either the hardware or the software, or maybe both.

As a side note, you keep reminding your readers of the importance of good memory. I'd go further and recommend ECC-capable motherboards (I don't like "mainboards" either :) and ECC memory. Some suggest that this is only for "servers". How come anyone can suggest that reliability is luxury, I cannot understand.

-- Daniel C. Sobral (8-DCS) dcs@newsguy.com dcs@freebsd.org

"I always feel generous when I'm in the inner circle of a conspiracy to subvert the world order and, with a small group of allies, just defeated an alien invasion. Maybe I should value myself a little more?"

Right on both counts. Actually, though, net speed isn't the limiting factor on most of the things I do. And there's only so much room in a column...

Thanks

You wrote : I could only fix that from Squirrel's keyboard; there's no way to do it remotely.

I say: Wrong

Try VNC http://www.uk.research.att.com/vnc/

VNC stands for Virtual Network Computing. It is, in essence, a remote display system which allows you to view a computing 'desktop' environment not only on the machine where it is running, but from anywhere on the Internet and from a wide variety of machine architectures. You can also control the remote machine using a mouse and a keyboard.

It is very small (less the 1 meg server &; client) , free, legitimate and usefull.

What else would you want from a program ?

Well, yes; it never occurred to me that anyone would be confused. Of course there are remote operating programs, of which the best known and probably most useful is Traveling Software's LapLink. All of these have to be installed on the machine you are to control. I thought it clear that I meant a normal Windows 98 or NT installation without special remote control software. Apologies for the confusion. I'll write about Traveling's LapLink another time.

Jerry

I have just read your column concerning file sharing in windows. In particular, your problem about a read-only share name. There is a poorly documented feature that can make sharing a little more secure. Any share name that ends with a $ sign is not shown on client Network Neighborhood dialogs or any other program that may show share names (such as the command line command NET VIEW "computername"). However, if you know the name you can share the folder by defining it explicitly. NT 4.0 etc. makes this easier by automatically defining C$, D$ etc. My point is, you can define a $ type sharename as Full Access and read only access for normal share names. Thus, only those that know the hidden share name can get full access.

Charles Godwin

Thanks. I didn't know that.

Hello Your latest Byte essay mentioned CD-RW. I have an external SCSI Plextor CD-R and use Adaptec's software. I am wondering if I can use CD-RW simply as another A or Zip type drive? That is to say if I wanted to save something I could just "save as.." to an appropriate folder on the disc. Or do I still have to go through the loops with an Adaptec type interface?

 Thanks, Peter. pldurand@home.com

Windows 98 and Windows 2000 do not at this time recognize CD-RW drives as "normal" drives; thus you must run some kind of software i fyou want to use them as "just another drive." Adaptec's Direct CD does that. I am told there are other software packages that will do it also; but I am most familiar with Direct CD.

The very latest version of Adaptec Direct CD is said to work with Windows 2000; I put it that way because while my information is reliable, I have not done it yet. It certainly works with Windows 98, and a Plexstor PlexWriter with Direct CD does look like "just another drive"; you can drag and drop to read and write to it, etc.  I have one in my latest machine. Works fine. I have had some problems with Sony IDE CD-RW under any software whatever.

©
Monday
Tuesday
Wednesday
Thursday
Friday
Saturday
Sunday
 
Top

Wednesday October 13, 1999

The following came in response to my Intellectual Capital column on "The Vanishing Manufacturing Jobs." 

Subject: Re: America's Vanishing Manufacturing Class

I greatly enjoyed your recent article, "America's Vanishing Manufacturing Class". But a couple of points of disagreement:

1. I disagree that the elimination of manufacturing jobs is due primarily to automation. It is due primarily to the heavily subsidized export of jobs to very low wage countries. The only jobs that are being automated are those that require more than human precision - any job that can be done by 50 cent an hour labor, will be and is. Consider the capital costs of automating the sewing of a shirt, factor in the risks associated with the possibility that the automation might not work properly, or that the need for the factory will end before you have recaptured your initial investment. Consider that you can get the same job done for 50 cents an hour by hand, with no start-up delays and no risk and no possibility of spending a billion dollars on a white elephant, and it's clear why automation is in fact not very widespread.

Note also: none of this is due to 'free trade'. The very low wages found in third world countries are counterbalanced by the instability and lack of the rule of law to be found in such places. It is government-subsidized political risk insurance, loan guarantees, bailouts etc. that have caused the spurt of investment in the third world, NOT low tariffs. Replacing 10 dollar-an-hour Americans with dollar-an-hour Mexicans sounds good, unless of course your entire billion dollar factory is burned to the ground or expropriated by some bandit government. The CEO of a big multinational would sooner go to a board meeting naked than invest in a place like Vietnam or India without cover from the US taxpayer. Whatever else you might say, don't you DARE talk about 'free trade' in anything other than the abstract or speculative sense.

At the end of the 19th century England began exporting much of its key industries to places like Germany and the United States. The decay in England's industrial base was blamed on a failure of the educational system. But if the tools and supplies that make up an industry are shipped overseas, how can education cause people sitting in an empty building produce something? All the blather about education today seems like a similar bit of misdirection to me...

2. I disagree that 'knowledge' is now the key to personal prosperity. Consider that in India, talented programmers make only about $2/hr - that's if they can find work as a programmer (The average wage of people who are trained to program is likely much less). All by itself India could produce enough programmers and engineers to perform all such tasks in the entire world, with enough of a surplus to hold wages in these fields to sub-poverty levels, and all without increasing its literacy rate over 50%. And India is just one third-world country among many. In the long run supply and demand applies to programmers and engineers just as it does to machinists and steamfitters, trust me.

Thank you for your consideration,

- Tim Gawne -

I found it highly interesting, and responded:

You know, I think I agree with you, but let me in my defense point out that I never said automation was the main cause of job exporting or job loss. Automation raises it's own problems, but since the actual manufacture and production stays home, so do the goods made; I don't think I said automation was a direct cause of job loss, and if I did I didn't mean it. What I meant was that automation raises the minimum standard skills requirement for long term employment, and that places extra demands on the education system.

Moore's Law requires better education without regard to exporting jobs.

But your point about political and economic subsidies to overseas nations being highly important to job export is very much on target, and I don't think I have given sufficient thought to that. THANK YOU.

I have asked this elsewhere and perhaps I should do a column about it: has anyone done an actual economic analysis of free trade (tariff = zero) given political externalities such as the cost of anomie and neighborhood breakups, political instability of job uncertainty, and the economic necessity of supporting in ever increasing opulence those who do lose their jobs to job export? I see no reason why zero is the optimum number in all cases; in Operations Research which was once my professional field we tried to make systems insensitive to assumptions, but if the answer were always a single number no matter the inputs, we suspected our models were deficient. Why should zero always be the right answer?

Note that I prefer tariff=zero to "free trade" which as you point out is usually not what happens. Perhaps among civilized nations. Perhaps.

Thank you for a thoughtful letter.

Jerry Pournelle

I also asked permission to post the original letter. The response was:

Thanks for responding to my comments. I have always been a fan (columns and sci-fi books) and it's flattering to get a response. Of course you can post it at your site.

As far as zero tariffs always being the 'right' answer, I think the real problem is not so much assumptions as motivations. When the same people who claim that zero tariffs are some sort of divinely inspired absolute law turn right around and demand subsidies and bailouts and free insurance, it's hard for me to believe that all the fuss about zero tariffs is anything other than propaganda. So while I think your idea about doing an analysis of the effects of true 'free' trade is a good one, I question that it would be of much use (actually John Maynard Keynes had some interesting thoughts about trade, there is a brief section in his general theory book). I respectfully suggest that the only real analysis of the effects of the status-quo on trade that count are whether the people at the top are making money now, and the answer to that one is yes.

Subjectively it seems to me that zero tariffs would only really work between nations that play by roughly the same rules (civilized or not). Otherwise the differences between laws, customs, wage rates, social stability etc. would eventually cause so much friction that something would have to give. But that's just an opinion, I have a hard time thinking about how to quantify it.

As regards automation raising minimum skill requirements, I'm not sure about that one. If there is little automation then surely this doesn't apply? But can't automation also reduce skill requirements? - a borderline innumerate at McDonald's can make change faster and more accurately than Albert Einstein, thanks to the newer automated registers. If much of current automation requires high skill levels to operate, couldn't that be due more to the ready availability of trained labor rather than the other way around? Consider the extremely high maintenance costs of Microsoft Windows. If big corporations really couldn't find enough skilled people at affordable wages to tweak all the glitches out of their systems, wouldn't market pressures to make an office computer along the lines of the Sony Playstation (A sophisticated real-time multi-tasking operating system used by millions of kids, you turn it on and it runs, period.) become overwhelming? (If farmers in southern California couldn't get enough migrant workers, might they develop less labor-intensive techniques, as was done in the Midwest?) OK, separating cause and effect is tough, but I am unconvinced that even if automation does become more widespread, that it will by itself require more skills - although depending on the relative supply and cheapness of skilled labor the SPECIFIC automation that actually is developed MIGHT require more skilled labor.

A horse is a more sophisticated automaton than any current machine, and lots of people without Ph.D. in robotics work them just fine.

Granted, though, a specialized billion dollar industrial facility will probably never be 'dumbed down', because there are not going to be enough of them around to make it pay, but I wonder how much of the labor force this will ever effect? Not sure about that one...

Thanks again for your consideration, and when can we expect a new book?

- Tim Gawne -

All of which is worth thinking about in some detail, and I haven't got a reply yet; one of these days I will have, and this entire discussion will warrant a page of it own over in "debates".

My point on automation and skills is that silly and repetitive jobs will be automated out of existence: I once had a major company hire mental deficients to solder mini-plugs; the work was important, but very boring, so the turnover in the department was extremely high. The answer was to stop hiring smart people to do that job, and get retarded people just smart enough to do it properly: they liked the work, which was challenging to them, it was important (this was a defense industry) and thus they were making a contribution to the nation, and the pay was very good. Now of course that job has long since been automated out of existence. But Dr. Gawne's point is well made.

©
Monday
Tuesday
Wednesday
Thursday
Friday
Saturday
Sunday
 
Top

Thursday October 14, 1999

This was sent to both me and Bob Thompson:

Dear Bob / Jerry,

I recently came across this article at http://www.usatoday.com/life/cyber/tech/ctg382.htm on "America's Digital Divide."

It is the most concise, yet thorough, intelligent, and even-handed article I have yet seen on the subject. As you well know (particularly Bob), >56K is simply NOT available at any price (excluding a dedicated T1) to many, perhaps most, of us. As at the inception of telephony, rural and lower-income-per-square-mile areas are being shoved aside by both cable and Telcos. I do not fault their management people. They do not run a charity. They serve their shareholders well.

HOWEVER: Each provider has a monopoly in its area of coverage. I have NO choice of whose cable or copper I connect to. Bandwidth unavailability is the fault of our Government (as usual) for not making it crystal-clear to bandwidth providers that "You WILL provide broadband access to all", as they told AT&;T and the Telcos "You WILL provide telephone access" not so long ago. Somehow, they all managed to muddle through - even turned a tidy profit while doing so. Sure it cost me a few cents a month more, but it is well worth it. I can now call Aunt Minnie in Bumsteer, Arkansas and hear her as well as a call across the street. It can be done if heads are knocked together firmly enough.

I really think such Broadband Equal Access legislation / regulation is LONG overdue. Certainly the FCC &; FTC are impotent on this issue. Thus far they have done nothing but waffle and wimp out; their only effect being to exacerbate &; confuse an already bad situation.

What can we, as underserved consumers, do to expedite broadband access?

Regards,

JHR --

culam@neteze.com [J.H. Ricketson in San Pablo]

Before I could post it, I got this as well:

Well, as I've said in the past, I'm against universal access, which is really just another subsidy. I don't see why I should have to pay more for my telephone lines just so that someone in West Podunk can pay $20 a month for a phone line that actually costs the phone company $200 per month to provide. Let that guy pay $200/month for his phone line, do without, or move somewhere that the phone company can provide a line at a more reasonable price. In the same way, I pay higher home insurance rates to subsidize home insurance for people who choose to live at the beach. It isn't fair, it isn't right, and I resent the government forcing me to pay to subsidize someone else. The same holds true for high-speed data connections. I shouldn't be forced to pay more for high-speed data access simply so that someone else can get that service at a subsidized rate, and that's all that "universal access" really means. When it comes to government intervention in a free market, be careful what you ask for. You might get it.

Robert Bruce Thompson thompson@ttgnet.com http://www.ttgnet.com

My own views are a bit more complex. The original Framers gave the Federal Government power to build Post Offices and Post Road, put directly into the Constitution, and Benjamin Franklin was the first Postmaster General, which was the fifth senior Cabinet Post and remained in the Cabinet until fairly recent times.  The only ones senior to the Postmaster General were State, Treasury, War, and Attorney General; giving some notion of the importance that Washington and the first Congress put on the Post Office. From the first Congress on, postage rates have been uniform, which is to say, city dwellers subsidized rural post routes for letters. Parcel Post was different, because it wasn't considered all that important to the stability of the country that everyone be able to send packages and Christmas presents at low rates, but letters were in those time the main means of communication, and communication among all citizens was considered important.

Newspapers were from the early times on given subsidized post rates, for the same reason. This isn't some modern liberal notion; this was agreement among Hamilton, Washington, Jefferson, Franklin, Adams, Madison -- the people who conceived this new nation and whose early actions determined how it would go. They thought it vital that citizens be able to communicate with each other and with their government, and that this include ALL citizens including those in fairly remote places.

One may now argue that this is already assured, or that it's irrelevant to the modern world; or that we have uniform postage rates now, and they're pretty comparable to what they were 50 years ago -- a 3 cent stamp was 3/5 of a nickel candy bar when I was a kid, and that candy bar is smaller for a quarter or more. The question is, in this age of rapid communications, is postage enough, or is electronic communication important enough to warrant the kind of equalization rates we have always applied to postage? The strict libertarian answer is that subsidies are evil. Since I am nothing like a strict libertarian, I don't have an automatic answer.

I do think, though, that things are likely to take care of themselves, as more and more people take advantage of remote access to move to remote locations. As economic clout relocates to places like Cedar Rapids and Mount Pleasant and Colfax and other small towns and cities, the demand for wide band communications will go up; while Moore's Law makes it all cheaper and easier to do. So my guess is that it's a matter of concern and interest, but the best thing the government can do at the moment is probably nothing, since it's not possible to predict precisely what the best form of broad band communication will be.

No one predicted that microwave channels along the old railroad rights of way would lead to rivals to the bit AT&;T long distance monopolies; and attempts to regulate haven't met with great success. I suspect that in this instance technology is far ahead of government's abilities, and the best thing to do here is to let things take care of themselves, which they are likely to do.

All governments can do is create bureaucracies. Sometimes bureaucracies are the right answer to a problem. But in all cases, the first purpose of that bureaucracy will be to pay itself increasingly higher salaries and to increase its size. Always. After that come the purpose the bureaucracy was set up to serve. Like the principle that a nation is better off if the citizens can and do communicate with each other, the principle that the purpose of government is to hire and pay government workers is eternal...

©
Monday
Tuesday
Wednesday
Thursday
Friday
Saturday
Sunday
 
Top

Friday October 15, 1999

Jim Koontz <jkoontz@simplyweb.net>

I have a dual Pentium Pro board with the megabit Pentium Pro chips, but being down here at the beach I don’t even know who made it, much less any others. I know there used to be such boards for Socket 7. We had a Diamond Flower at one time, too. I haven’t done much socket 7 work in months. Perhaps someone will know. Thanks for the kind words.

This was sent to me in email. I don’t vouch for any of it.

Anne V - 10:01am Sep 13, 1999 PDT (# 1403 of 1405) There is not enough almond milk in the world to solve an event of that kind.

And I have no comments…

Steve Setzer who invites any reader to come by the house and see it in action

Well, I’d love to see it in action! But I make no doubt that there are many solutions to the limited bandwidth problem. Thanks.

©
Monday
Tuesday
Wednesday
Thursday
Friday
Saturday
Sunday
 
Top

Saturday

©
Monday
Tuesday
Wednesday
Thursday
Friday
Saturday
Sunday
 
Top

Sunday October 17, 1999

Most of this mail belongs in next week, but being down here with a terrible connection this will have to do:

Best regards,Abraham SI

Subject:
ZIP drives
Date:
Mon, 18 Oct 1999 14:54:32 -0500
From:
"Spragens, Henry" <Henry.Spragens@compaq.com>
To:
"'jerryp@jerrypournelle.com'" <jerryp@jerrypournelle.com>




Jerry,

Steve Gibson's TIP proved a valuable tool a while back when I was having
mysterious ZIP drive problems. I have IDE ZIPs on the Compaq Deskpros and
SCSI ZIPs on the Macs at home, and interchange disks among them all the
time, mostly in Windows FAT format. I started noticing that the Performa
6400 with the internal SCSI ZIP was having trouble reading disks from the
DeskPro 6000. Then all the machines started having problems, and failures
started to snowball. Re-formatting disks produced uncertain results. I was
starting to think all my drives were bad, or that a bad disk had damaged
multiple drives, recalling a disastrous episode with Syquests at work a few
years ago.

TIP is slow, but it, SpinRite, and FWB Hard Disk Toolkit (the Mac equivalent
of SpinRite) allowed me to track down the problem to the 6400's internal
drive, which would damage a disk's format just by reading it. As you found,
one of the IDE drives couldn't reformat the disks correctly. Since SCSI
internal drives aren't made any more, I got an external and hooked it to
that machine and all was well, except for the ugly hole in the front panel.
The funny thing is, the bad drive works fine when it is removed from the
computer. If I reinstall it, it goes bad again after a few days. I suspect
some sort of frame flex and binding of the head carriage or something
similar, though I can't confirm any problem on the bench.

At the time, I was getting pretty leery of Iomega's products, but since that
one problem was resolved, I haven't had any further errors. I regard Jaz
disks as inherently risky, since they fly the heads on removable hard disk
media. ZIPs ought to be more forgiving of the odd dust particle. The main
problem with depending on them is that they are single sourced, and if
Iomega has a bad batch or gets into financial trouble, you're stranded.
Alas, we don't seem to have a good industry standard replacement for the
floppy yet. Ethernet and servers aren't really substitutes for removable
media.

Regards,
Henry Spragens

---

Subject:
PaperBytes
Date:
Mon, 18 Oct 1999 20:11:23 -0400
From:
rgrunsk@ibm.net
To:
jerryp@jerrypournelle.com




Dr. Pournelle;

The software that Mr. Si is looking for is for the old Byte "PaperByte"
system. This was a bar code system that was to be an alternative to
cassette and paper tape for mass distribution. Cheap floppy drives
killed it (I Imagine.) I have several of the old Byte PaperByte books
(Basex, SuperWumpus, K2FDOS etc, all from 1979.) I might also have, but
it will require some time to look for, the original PaperByte Loader
book. I remember seeing it but I can't remember if I saw it last month,
last year or in the last decade. If I find it I will inform Mr. Si
privately.

In the meantime, he might try a library that has back issues of Byte.
The time period that he should be interested in is 1978, 1979.

Yes Jerry, I have been reading your column in Byte since it started and
it was the main reason I continued to subscribe to it long after Byte
became a manager's magazine.


Rolf Grunsky

--
-----------------------------------------------------------
rgrunsk@ibm.net Remember, what you see coming
at you is coming from you.
"Jungle" Jack Flanders
-----------------------------------------------------------

©
Chaos Manor home

Entire contents copyright 1999 by Jerry E. Pournelle. All rights reserved.
Comments and discussion welcome.