Hi Jerry

I guess it is important that everyone be in an identifiable camp. I really don’t speak for the Linux community as you well know.

The problem you have with many of us, I can include Rogers Cadenhead, is your incredible load of side. In a corporate environment I guess self promotion and that whole ball-game is important, but Linux is not at all in that stream (although god knows many people have been doing their best to bring it to that polluted river). I guess what I’m trying to say is that what is received well in Redmond, will not go down too well with the people who hacked their OS together themselves. Mellissa is amusing no?

Upgrade to Linux...the penguins are hungry!

Ah well, not a widely read person I take it. *cough* Scholarship ?

Dr. Pournelle:

Hi

Read some Kipling, Stalky&;Co would be a good place to start. I’m sure many of your more literate readers already find this amusing.

CC

I would not have continued this but if you want to beat on it some more I’m game.

--

Upgrade to Linux...the penguins are hungry!

Chris Carson aka "the old one"

250-248-0142

Thermidor, Robespierre, and the Net

We are pretty close to that future you predicted, where anyone connected to the net can quickly get the answer to any question that has a known answer.

Thermidor didn’t ring a bell, so I went to Altavista. In amongst the recipes for Lobster Thermidor were links which mentioned Robespierre. Along with the context of your allusion, with the mention of Napoleon, I knew within seconds of deciding to find the answer that "Thermidor" was a reference to the fall of Robespierre and the rise of Napoleon.

At first, the English documents were very skimpy outlines. There were some documents which seemed like they might contain more detail, but they were in French. "Babelfish" translated one of those into more-or-less English, from which I was able to pick out most of the story.

I still didn’t entirely understand "the Ninth of Thermidor" , a.k.a. July 27. At a guess, the Revolution must have had a different calendar. A few minutes later, I found a document (in English, this time) which confirmed that guess. (And had a whole lot on the Revolution and the Terror that I had forgotten.)

Dear Jerry:

I am a long time reader of your column, I even remeber your old S-100 bus systems that you used to write your books on. Any way I want to mention a program that saved my rear today Nuts and Bolts. I came into the office early to work on our new accounting system and my workstation wouldn’t boot. I about had a cow. I finnaly got the system up to a command prompt. I couldn’t get to my ZIP drive or Network drives. I went out and bought a copy of Nuts and Bolts at the local Office Depo,. Installed it on another system here, copied the dmdos program to a floppy, and ran it. Dmdos recovered my drive and fixed the problem. I can not say enough about this program. I have been so lucky I hav used or owned a C-64, Zenith 148, Amiga 1000, Amiga 2000, Lantastic, Novell, MS-DOS2.11 to MS-DOS 6.22, Windows NT/95/98, and am now setting up LINUX servers for the accounting system here and this is the first time I have ever had a hard drive crash. I guess everyones luck runs out sooner or later.

Nice to see Byte back but I do miss the old Byte from the 80’s

David Siebert [dsiebert@eclipsecat.com]

Subject: Benefits of Microsoft's much-derided document ID numbers, and related musings

The New York Times have an article on their Web site describing how the

unique user ID number embedded in Word documents, which caused so much

brouhaha, may actually help track down the author of the Melissa email

macro worm. To read the piece on the New York Times Web site (free

What I think this brings out is that there are really two sides to the whole privacy-vs.-accountability debate. While the Clipper chip and similar initiatives have turned out to be rather blockheaded in their respective implementations, in my opinion the diehard privacy advocates are being wrongheaded when they seek complete and total anonymity for all electronic communications.

We don’t have anonymity when we purchase a house, for example, or an automobile. Why? Well, one reason is because the seller needs to know that the potential buyer is creditworthy, has a good track record in previous dealings of the same sort, etc. Same thing for purchases made face-to-face with credit cards in shops; merchants need to know who you are in order to prevent fraud.

As a matter of fact, even local telephone calls aren’t completely anonymous. Though most people don’t realize it, the local exchange carriers, or LECs, maintain complete records of all calls placed through each switch. These Message Unit Details (MUDs) are often used by law enforcement in the course of investigations into serious crimes.

We can still attempt to achieve anonymity by using pay telephones and disguising our voices, or very carefully preparing and mailing unsigned printed letters through the post when we feel it necessary. The same goes for email and other forms of electronic communications. Still, I believe the privacy advocates would make a much stronger case if they pushed not only for unrestricted use of strong encryption to protect content, but also for a universal authentication and trust infrastructure which would guarantee that if one accepts (for example) an email message with an attachment from a previously-unknown source, one would be able to tell at a glance whether or not it originated from an authenticated source. A standards-based security model of this sort would go a long way towards speeding the acceptance of electronic transactions by the public at large, one might well suppose.

My personal view is that many of these so-called privacy advocates in reality are simply using high-flown rhetoric in order to try and ensure that their less savory activities, such as software piracy, trafficking in illegal goods or services, etc. can continue unmolested in the electronic arena. I do not say this lightly; I have been involved in computer networking and related security issues since 1985, and during the course of my professional duties have made it a point to get to know some of the types of people who tend to scream the loudest about such things. Behind their noble-sounding public pronouncements, there almost always seems to lurk the wink, the nudge, which signals those in the know that while the words may be weighty, in reality the objective is simply to confound the ‘establishment’ and preserve a rules-free playground in which anything goes.

I could go on in this vein, particularly in regards to Mr. Stallman and his crowd (authors of the infamous GNU CopyLeft licensing agreement which virtually ensures that software developers will never be able to make a dime from any application they might wish to develop for LINUX; once Oracle, IBM, and the rest of the crowd who’ve jumped on the LINUX bandwagon realize just what they’ve agreed to, I think all the sound and fury over LINUX will fade away in short order), but I think you get my drift. Privacy and anonymity are necessary for the preservation of intellectual freedom in many cases, yet openness and accountability are the twin pillars of civil society.

Though it’s probably too much to hope for, I for one would certainly welcome a day in which the privacy agitators acknowledged that their detractors actually have some good points, and actually sat down with their opposite numbers and talked seriously about the concerns of both parties. It probably won’t happen in your lifetime, though, nor in mine.

"Do you know if anyone is offering verbal tech support for Linux? I wouldn’t mind paying for answers. It’s cheaper for me to work more overtime at the mill shop instead of staring at the monitor for three hours or e-mailing tech support. "

Hi Jerry (may I call you Jerry? it was rude of me not to ask before),

[snip]

[snip]

Unfortunately it would seem that you have got it correct.

Microsoft has (according to their web site, off the top of my head, I will need to look it up to tell you where it is) fixed over 400 memory leaks between sp3 and sp4 in the NT operating system.

I would use the WinNT task manager to keep track of what was happening to the memory and I discovered that the system would gradually eat more and more memory. At the end of the day I would quit all the programs, only to discover that I my memory footprint would be 10-20 MB greater then when I had started. The taskmanager could not identify where this memory went.

I came to the conclusion that regular reboots are good. Once a day for a 64

MB machine, and longer for 128MB machines

A further note on WinNT and performance

As WinNT made greater use of the swap file, I discovered that performance would suffer terribly, more than what I would have expected it to, and far more than what would happen with an OS/2 machine that I used when I over comited the memory (a totally unscientific, subjective test where I would use PhotoShop 2.5 under windows 3.1 under OS/2 with a 64MB VM on a 64MB OS/2 machine)

The people who write the diskeeper software offer the following insights

I had a workstation with 64MB of ram and a 120MB swapfile. Using diskeeper I found out that my swapfile consisted of over 40,000 fragments! Take a small guess at what this might do to the performance. of the machine when NT decides its time to thrash a little bit.

Granted, this was on a machine that had been in constant use for almost a year, but the extent of the fragmentation of the swapfile was very surprising. Diskeeper claims that it can defragment the swapfile (during a bootup defragmentation) but it was never successful on my machine. I guess it was too badly fragmented.

Whoah! What a difference that made. Yes the machine still sucked up memory, but when it thrashed it no longer sounded like a herd of stampeding tap dancers!

On a perverse note, I decided to find out how NT would react when I filled the swapfile. Remember, I did not allow this swap file to grow.

Well NT ran out of memory and warned me like I would have expected. It then hung (one process was still trying to commit memory) and overwrote NTLDR and NTDETECT.COM on the NTFS partition that the swap file was on.

It took chkdsk 2 attempts before it cleared the corruption to the NTFS partition (on another machine) and after replacing the damaged files and putting the disk back the machine would no longer boot from the NTFS partition!

I researched this problem on MS TechNet and I found the document that described the problem of why the machine cannot boot. It sounds similar to the problem that used to exist when IO.SYS and MSDOS.SYS were not at specific locations on the hard disk in earlier versions of DOS. However the difference is that under DOS you could get the files into the right place by using the Norton utilities, or by figuring out what files to move around to do it yourself.

The solution to the problem with NT and NTFS? Reinstall NT or use a boot floppy.

Needless to say, that workstation had two boot floppies (one spare) at all times to boot the machine into NT.

What was that you were asking about garbage collectors again?

I have been an avid reader of your column since I first read Byte magazine and have followed your own web site with interest. I find it fascinating to see how hard it is at times for "computer literate" persons to install, use and update commercially available software on industry standard machines.

From my own experience, however, I think that sometimes people create problems through too much fine tuning and tweaking. This may be the reason why I have a problem. My Registry (I use Windows 95 OSR2 in Italian with IE4 enhancements) has got exceedingly bloated. However much I try to clean it, it seems to get bigger. Even when I physically delete portions relating to programs that are no longer on my computer, the size of system.dat and user.dat remains the same (currently 3695 and 512kB). Question: is there a simple program available not only to clean the Registry of unneeded or unwanted items, but also to shrink it down physically in the same way as e-mail programs compress files? I tried out something like that on Nuts and Bolts a year or so ago, but various other features of that program scared me and I scrubbed it from my disk.

P.S. Thanks for the cultural note in some of your comments. I always wondered about "Bob's your uncle".

Subject: Benefits of Microsoft ID Numbers

Mr. Dobbins brings up an interesting point. Of course there are positive uses for the Microsoft ID number, just as there are positive uses of the telephone MUDs. The problem, however, is that although these things can be and often are used responsibly by the proper authorities, there is no built-in safeguard that forces them to be. He makes mention of several forms of communication that require accountability and openness, and he’s right; but he forgot one thing. In every case, there is a means of conducting the exact same transaction anonymously. I can purchase a house or an automobile with complete anonymity if I choose, by paying cash and working through someone else, a lawyer or an agent.

I can avoid the accountability of a small purchase by paying cash, instead of using a credit card. And although MUDs track which phone called a certain number, and the time and duration of the call, it does not record the conversation or prove who actually spoke. So although enough information is recorded to assist investigations into serious crimes, freedom is preserved in the details. I can avoid even that much record for something I truly need to keep private by sending a letter, or using encrypted email, or arrange for the call at a certain time and place other than my home or the recipient’s home. A local payphone or restaurant, for example.

For every form of communication or activity that is in some way restricted, there is an anonymous method of reaching the same goal. And for every form of anonymous activity, there is an illegal or immoral activity that can benefit from it. But there are also countless activities that are neither illegal nor immoral that not only benefit from anonymity but rely on it. The battered wife searching for a means of getting a hidden plea for help out. The ability to conduct a search for a better job without quitting the one you have. By restricting privacy and anonymity you hurt the valid far more than the invalid. Software piracy will not stop if you kill anonymous FTP and shut down the warez sites. Nor will any form of illegal activity suffer from "accountability" such as you propose. They were all present in our society long before the internet was available, and they will always be around.

The only way to prevent good things from being used in bad ways to eliminate all freedom, all our rights, and everything about our society worth saving. And that is a price I, for one, am not willing to pay.

Dear Dr. Pournelle,

"Gentlemen cadets from Sandhurst and Woolwich, who had only left a year ago, but who carried enormous side, were greeted with a cheerful `Hullo! What's the Shop like?' from those who had shared their studies."

And later…

"But when the real subalterns, officers and gentlemen full-blown--who had been to the ends of the earth and back again and so carried no side--came on the scene strolling about with the Head, the school divided right and left in admiring silence."

So "side" seems to imply the kind of arrogance that’s usually attributed to youthful inexperience. "You’d better get busy now while you still know everything!"

Main Entry: side

Function: noun

Etymology: obsolete English side proud, boastful

Date: 1878

chiefly British : swaggering or arrogant manner : PRETENTIOUSNESS

Apparently, someone wanted to tell you that you’re arrogant and pretentious, and used a word that hasn’t seen that usage in a hundred years to tell you. J

Bill

I’m not sure that I’m frightened by this article as much as you seem to be. Granted, any Federal program can be carried too far and become harmful to those it was designed to help, in addition to being harmful to the Federal Budget. But, I’m on the local school board, and most of the problem students we have to deal with are products of homes where the parents probably should never have been parents in the first place. Or at least failed somewhere along the way in being good parents. Too many people have children just because they can. Their own lives are so out of their control (having to live up to everybody else’s expectations, on the job, on the street, etc., and not being able to afford to do their own thing, so to speak) and subconsciously they think that they will be able to control a child, thus bringing some sense of power or authority in to their otherwise powerless existence. Parenting is the most important job anybody can have, and the majority of parents have no idea what they are doing. If you want to drive a car or fly a plane, you have to at least demonstrate a certain knowledge of how to properly perform the task. You can’t just jump in a 747 and fly it to Hong Kong. But if you want to be a parent, you don’t have to demonstrate any knowledge of anything (not even of how to perform heterosexual sex acts). You don’t have to be healthy. You don’t have to be sane. You don’t have to be financially capable of supporting the child. You don’t have to be married. Women don’t even need a man anymore (artificial insemination). This all leads to a general lack of concern by some parents as to how their children are actually developing/growing up.

Maybe first time parents should be contacted by some social service agency

to ensure that they are capable of raising a child to be a productive

Compacting Win Registry

Mr. Walker’s Registry question was answered by one of your readers some time ago. The solution is to boot into DOS without starting Windows; use the DOS version of Regedit to export the Registry to a text file (this takes a long time—plan on having lunch while it’s going on); then use Regedit again to import the text file back into the Registry. My Registry came out reduced by about one-third in size.

I confirm the previous reader’s comments that DOS Regedit works better than the Windows version. But while it reduced my Registry file, it has not made any difference in performance that I can perceive.

Subject: IDE and ATAPI

Jerry:

A couple of questions to throw out to your many knowledgeable readers:

Once upon a time, good sound cards came with ATAPI connectors, so you could have a CD-ROM AND 4 HDDs. This doesn't happen any more. Are there ATAPI adapter cards out there? Whence cometh the IDE limitation of 4 devices, and why is it that no one's come up with a better way?

I'm no novice, but I am cheap, and while I know that bigger HDDs are cheap nowadays, I hate to toss the older ones, especially, the ones that aren't really that old. I try shuffling them down to lower-end machines for dedicated purposes in the house, but my wife is beginning to tire of furniture that hums, and really, it's tough to justify a PC in a kitchen already over-flowing with tupperware. So how do I get 8 IDE drives in a box, without going SCSI? Linux? other hardware or software? Is the limitation software or hardware driven?

Thanks!

Dear Jerry,

Just read of your continuing problems with Eagle One. This message will be of little or no help to you.

I use main boards from ASUS (ASUSTeK Computer, Inc.) and previously from M-Technology (now part of SOYO). The main reason that is they allow the IRQs to be set in the BIOS by the installer. It would be even better if they allowed the I/O and memory addresses to be preset, as well.

They also have an Auto setting to allow the "normal" situation:

resources that periodically rearrange themselves to create havoc.

Might Eagle One be cured by the "tried and true method"? i.e. reload from scratch. This is, in my oponion, the only way to effectively clean up the registry/DLL mess that ensues as a WinXX system ages. Unfortunately, unloading applications, drivers, and protocols does not generally remove the registry settings. And the "Uninstallers" aren’t much better.

John G. Ruff.

Perhaps I should’ve taken the analogy a bit further. Yes, one can purchase a home, land, an automobile or whatever anonymously, but -someone- is held accountable for its use. That’s why deeds are public, why one must register a vehicle and purchase insurance for it, etc. Since so much of the livelihood of so many people now extends into the virtual economy of information and ideas, does it not make sense to extend those same sorts of accountability, where appropriate, to the Net? Mr. Beland seems not to have considered these points.

I also have observed that people who tend do things such as purchase houses, cars, and so forth in such a way as to try and obscure ownership are invariably doing so in support of some shady enterprise or another. And, no, this isn’t the simplistic ‘if you’ve nothing to hide, why do you want privacy’ argument - it’s more nuanced than that. Give me a bit of credit, eh?

And it’s not about a drive to ‘eliminate all freedom, all our rights, and everything about our society worth saving’, either. This is precisely the sort of inflammatory setting up of straw men which leads both sides in this rather important debate astray from the issues at hand.

This silly Melissa macro worm, for example, has cost real people a great deal of the most precious commodity any of us will ever have - time. Why did it run rampant, as we already have the technology available to prevent such things from occurring? Namely, because the people who develop technology tend to be so ad hoc, so go-go-go in their collective mentality, that they fail to consider the risks inherent in what they implement, and by so doing fail to make allowances for both malice and incompetence.

Eric Raymond’s ‘Web of Trust’ model is certainly workable; we have the basic technology to implement it today. Many of my clients pay me to do precisely that sort of thing, using digital signatures, public-key servers, etc. Unfortunately, electronic commerce and related activities of a legalistic bent on the Internet will remain ‘amateur night’ unless and until the brightest people on both sides of the privacy/accountability divide can sit down, work out procedural compromises, then come up with not only standards but actual code that can be used to build products which end-users may access.

Sophomoric prounouncements on how it’s better to let all the evil run rampant because we might conceivably slight someone’s inalienable evil somehow are also beside the point. By that logic, we should legalize child pornography, animal abuse, etc.

And I haven’t proposed -eliminating- anonymity, either - rather, simply -adding- a more secure and accountable system side-by-side, while leaving people the choice to use whichever they choose.

See what I mean about how the ‘privacy advocates’ can’t seem to stay focused? Mr. Beland is arguing right past me; he didn’t even bother to read what I wrote. And so I return to my original point, namely, that the solution to this problem is a long ways off precisely because of the intellectual sloppiness and overt emotionalism on at least one side of the debate.

It’s too bad, really.

Subject: Outlook:thank you

Dear Jerry

I’ve been reading View40 with your Outlook problems. As a result I was able to solve a problem that has been bugging me since the middle of January: I cocked up the whole machine back then and lost the contents of my inbox and other folders. I resingned myself to their loss as nothing I tried got the mising stuff back. Last weekend I was messing about and deleted Windows Messaging (I thought it was obselete) only to lose the inbox contents again and also the right click list had gone as well as the contents of the Actions drop down menu. I got those back but still no inbox contents.

Now I have read your column I have the lot back: renaming outlook.pst to new.pst then opening Outlook and pointing to an old outlook.pst which I found with Start/Find got me back to my first crash and then importing successive versions of outlook.pst and finally new.pst finished the job.

Once again: Thank you

The poor kid was excited by the film, took a pringles can, put a contruction paper nosecone on it, stuffed it full of newspaper, a tube of liquid model airplane paint, 3 matches, and a STRING for a wick, put three bandades on the outside as thermal indicators, since they would turn brown before a "burnthrough" so that the kid and his buddy could run for it, and he took THAT to school expecting to shoot it off in an empty lot near school afterwards. This they have classified as a firearm under the definition "Rocket with propellent greater than 4 ounces" and have given him a supension till the end of school without right of appeal.

Jerry,

I looked forward to your column in Byte for years. Glad to see it is now online.

I had similar problems with PnP. They were compounded by the fact that I was running a three operating systems on a multi-boot system. The problems were solved by turning PnP off in the system BIOS configuration. Currently I have PnP turned off for all IRQs except those known to be in use by PNP. If I add a PNP card I enable one more IRQ.

The most memorable time (pun unintentional) I had serious system memory problems was when a trial version of CleanSweep was installed. It added about 2 MB of memory to the size of every process running on my system. The moral of the story is: If you don’t need it all the time, don’t run it all the time. My startup folder is empty and the only program that I run all the time is a very up-to-date virus detector. I even begrudge running that but it is, unfortunately, essential.

In regard to Robert Maxwell's questions about IDE/ATAPI:

The first and most important thing that must remembered about this interface is that the creators cared about only two things, cost and short-term effectiveness. When this was being thought out CD-ROM drives were not a factor. Tape drives were just as likely to run off the floppy controller as IDE or SCSI. Two drive seemed like plenty. Since having a CD-ROM drive on the same bus as your hard drive would usually have a horrible impact on performance it made sense for the sound card that was the center of a multimedia upgrade kit to do the job of running the CD reader. Especially since most early drives weren't IDE but a proprietary interface instead. Some readers might recall the generation of Sound Blaster that had no less than three different CD-ROM interfaces.

When CD finally became a required item for most systems it became common for motherboards to have the now familiar dual IDE controller arrangement. Until recently this was completely adequate for the great majority of PCs, and SCSI was always available if IDE couldn't get the job done, provided you were willing to pay the toll. In addition to supporting many more devices per IRQ claimed, SCSI is also much better at dealing with big drives. Ever since drives larger than 500 MB have become cheap the hassles of IDE become ever greater.

One of the things I truly dislike about IDE is the brain dead cable design. Now, a proper IDE cable should little ridges on it to assure it can only go in the right way but it is apparently a few cents cheaper to leave those off since so few of the cables I encounter have them. (They often lack the ground marker as well.) This makes it very easy to plug in a cable upside down and get a non-booting system for your trouble. To any engineers out there: No matter how many pins your interface uses, NEVER put it on a symmetrically shaped plug and socket. If it cannot be correctly connected without being seen it's a bad design.

Then there's the horror of IDE cable length (or the lack) and the need to bend the cables in weird ways to get two drives connected. I'm quite anxious for USB and IEEE 1394 to make ribbon cables a thing of the past. For those of us who make our living sticking our hands inside these boxes full of metal teeth it won't for at least another three years.

Until then, if you're looking at lots of little drives and SCSI is inconvenient or too expensive, I have two words of advice: charitable donation. IDE drives of immense capacity are just too damn cheap to make it worth wondering how to get more small drives active. You can buy additional IDE controllers at your local Fry's or equivalent, and that will often overcome the limitations of an older BIOS as well as enabling more drives but why bother? Make your local school or library a better place and use the tax break to buy bigger drives at less than half the price per gig of the old ones.

Main body subjects Sysinternals, Outlook suggestions, your detractors 11f

From: asdfasdf

Dr Pournelle:

First, a tip of the hat to your correspondent who recommended sysinternals.com and its products. I have one CADD unit which unaccountably and abruptly went temperamental and slowed down a while back. The more basic maintenance routines didn't alleviate its problem. Pointing Sysinternal's Filemon at the machine immediately found a McAfee ini file gone berserk: while the hard disk light had remained out during periods of apparent inactivity, in actual fact the processor had been heavily occupied in irrelevant activities which drastically reduced time available for other work. Uninstalling / reinstalling the offending program resolved the problem.

Second, with regard to suggestions to pass on to Microsoft about its Outlook program, two items come to mind: I'm running Outlook 98 on a Win95B platform dialing into a company provided access and occasionally traffic in Outlook email with large attachments. The connection is not "fast" (and it wasn't so long ago that I had to pay long distance charges to access it), and hence I am only too aware that:

1. When I send an email, it is first transmitted from the "Outbox" folder, and then it is re-recorded back into the "Sent Items" folder on-line with appropriate header information. If the sending process for an outgoing 6 Mb package requires 30 minutes, I must wait another 30 minutes online for the package I originated to be read back into the "Sent Items" folder from the server. I may have missed a checkbox somewhere in Options and I (hope and) am more than willing to be corrected, but I haven't found a way around this as yet. I could probably elect not to synchronize the "Sent Items" folder, but I want that transmittal record. What I am angling at is having the server send back only the header information of the sent message which my PC would then append to my original message (still somewhere on my hard disk), which in turn would then be transferred (internal to my PC) to the "Sent Items" folder. And thus reduce on-line time.

2. I am on various distributions for progress photos (rarely compressed) which I can't refuse (office procedure) but don't often need to view, plus I get the inevitable and unwanted jokes forwarded to me, some of them heavy with graphics (again rarely compressed). As I believe now available with IE5, I would like the opportunity with Outlook to review before downloading what email the company server has stockpiled for me and delete or postpone downloading those files I judge unneeded at the time. In fact, it would be nice to be able to download the text of email with large attachments--without the attachments and with the option of later downloading those attachments I deem needed. That might allow me some insight into what I would otherwise simply trash for lack of time.

In addition, I agree with your lament about the lack of documentation for Outlook. I would guess that it extends into the sysops' domain judging from the blithe dismissals from mine about the still incredibly long waits I must suffer thru for "synchronizing" and "processing mail rules" even after I have substantially reduced the (formerly seemingly eternal) waits on-line after invoking the various tweaks I have read about.

In closing:

One, I'm comfortable with the organization of your website.

Two, with regard to such drivel as the "side" comment and the resulting communications--you're letting your detractors sidetrack you from the productive efforts and dialogues for which I and, I would wager, most of your regular readers come to your website. I understand that incoming criticism should be evaluated and discussed in the hope that most of it is constructive. The snidely commented criticism from the self-styled intellectual wasn't in that vein and needn't have been even aired. Don't let such stuff waste your time--and mine: time's a commodity becoming ever more precious and (inevitably) we're all running out of it much too quickly. Your credentials are solid in my book and I'm grateful that you have developed this alternative to replace your old Byte column. In fact, this current milieu is better than your old column.

Jerry:

It appears there is a link between the Melissa Virus and the MS HWID flap.

the FBI traced the Melissa Virus by looking at the HWID in the original Word doc that hosted the virus, and then finding instances of that HWID in Word documents on some cracker / virus coder web site!

While I don’t like the "Big Brother is Watching You" aspects of having a unique identifier in any document I create, I’m glad the FBI was able to trace down this jerk.

Thanks,

Subject: IDE Hard drive woes: MBR &; EMBR

Jerry,

Just read about your recent troubles with invalid partitions and hidden logical drives. Something similar happened to me last week; I was handed a drive marked ‘bad’, that could be FDISK’ed and formatted fine, but would spit up ‘Invalid Partition Error’ on boot.

On the whole master/slave thing, we had a similar problem to yours with a Western Digital drive that only wanted to be a secondary slave, not a secondary master. It’s not been worth the time to find a work-around; just thought you’d like to know that You Are Not Alone. ;-)

Julian Morley [magnus@gyral.com]

To: rmaxwell@pei.wabco-rail.com Reply

I myself run two EIDE hard drives (a Fujitsu 2.1GB and a Conner 1.2GB, partitioned in toto as 9 logical drives C-K) plus a Teac CDC68E CD-ROM internal 6-changer, all three hooked to a Tekram DC690CD caching PCI EIDE/ATAPI controller. Tekram no longer makes this controller, but seek and ye may find.

You can install two DC690CD’s onto a single mainboard and this way run up to 8 EIDE devices, says the Tekram’s (rather poorly machine-translated) manual, but I never tried it. You *should* be able to have a CD-ROM AND 4 HDDs. Alas, though, if you install a second DC690CD (trying to get 1 or 2 CD-ROMs and 8 HDDs), you cannot hook a CD-ROM reader to either, because the second DC690CD needs the port address which would otherwise be used by the ATAPI port. And I suspect that if you try hooking 4 HDD’s to the DC690CD, you need a bit of luck as to the make and model of drives.

How fast does it run? Well, I still use WFWG 3.11; I considered Win95 to be nothing but WFWG 3.11 plus the Helix Hurricane and Cloaking utilities, slapped under a stupider-OS/2-clone interface, and I refused to accept the extra layer of button-clicking necessary to start my programs. Yes, true, I am a CLI diehard—though, my wife’s machine does run Win95, because she needs Office 97. I had to spend many man-hours tweaking the BIOS and system.ini to get the Tekram and drives to run acceptably fast. Also, I once tried hooking a WD Caviar 3.1GB drive to this controller, and never could get it to run error-free. If anyone wants the details on tweaking system.ini for the DC690CD controller, I’ll charge just a nominal fee—say, one DC690CD in working order (many used ones have nonvisible defects which totally destroy them --and I’d like to pick up 2 or 3 good ones. I bought mine for $12.50 in January of ‘97; the store had about 3 cubic yards of them, unsellable because nobody could get theirs to run under Windows 95 with the Caviar disks which said store used in all its computers), or a package of Helix Multimedia Cloaking, or a 4MB 30-pin memory module for the controller. I have 4 of these memory modules, 16MB total, the max, on mine.

Using old DOS Norton 6.0 Sysinfo’s disk benchmark, being run within a DOS window just now, here are the results. The DC690CD’s caching action increases the transfer rate for the first three runs. Then the cycle starts over. My machine is a homebuilt 166MMX Pentium which I am overclocking at 250 (83MHz system bus, 41.5MHz PCI bus):

The third run’s anomalously high Disk Index on Drive 2 is an artifact of running under Windows, wherein the shown CPU Index oscillates from 54.8 to wildly high values around 1,000.

The Conner drive’s own Disk Manager software shows sustained transfer rates of just over 20,000 KB/Sec once cacheing action kicks in.

Here is another test—and I would like to know how today’s Ultra ATA drives using DMA, and also SCSI drives with a cacheing controller such as the DPT Smartcache 4, fare in similar tests. The test is to xcopy /s/e (I, of course, use DOS 6.22 XCOPY, not the XCOPY32 from Win 9x) my entire \WINDOWS directory, with all subdirectories --55,214,253 total bytes in 986 files—to a freshly created \WINIMAGE subdirectory on the other physical hard drive. Using the following batchfile:

I get the following result in %ramdrive%\time.txt:

Current time is 12:14:16.89p Enter new time:

Current time is 12:15:31.31p Enter new time:

...for an elapsed time of 74.42 seconds. Dividing this interval into 55,214,253 total bytes, works out to a real-world transfer rate of 741,927.61 Bytes/Sec.

I am converting to Linux—I suffered a Back Orifice attack on August 15^th, 1998, interesting because I was using WFWG 3.11 _not_ Win 95-- and my wife had an NT4 SP3 network at work and so I know the security problems thereof, and thus my motto is TAASTAASMOS (--as a secure Microsoft OS). I consider Microsoft to be Opersysta Non Grata from now on forever, on my machine. I can hardly wait to try and tweak this controller under Linux.

I totally agree with Eric Pobirs’ comments WRT IDE’s brain dead (ribbon) cable design. Oh, the twistings and turnings and backward connections thereof. But if you want to charitably donate your old "small" (say 10GB or less—actually, I could also use a few small (120-400MB) ones too) IDE drive, why, let me know and I’ll send you my mailing address.

Happy Easter,

Philip