Week of 1/4/99
Friday, July 05, 2002
A (mostly) daily
journal of the trials, tribulations, and random observations of Robert
Bruce Thompson, a writer of computer books.
January 4, 1999
No Book of the Week this week. I read quite a few books last week,
including another Turtledove, but nothing worth buying.
* * * * *
I noticed this weekend that PGP (Pretty Good Privacy), Phil Zimmerman's
encryption program, is now available for Outlook 98, so I spent some time
playing around with it. You can use PGP to digitally sign and/or encrypt
email messages you send. PGP works using a combination of public and
private keys. Your private key is secret--known only to you. Your public
key is just that, public.
You use your private key to sign and/or encrypt an outgoing message. If
you encrypt a message to a particular recipient using his public key, only
he can read it, and he must use his own private key to do so. If you sign
a message but do not encrypt it, anyone can read the message, but a valid
signature guarantees that the message is from you and that the contents
have not been altered. Overall, a pretty useful set of functions, and one
that Microsoft should have built into Outlook 98.
Instead, they chose to use a competing security/encryption technology
that requires you to get a certificate from Verisign, which you can do for
about ten bucks a year for the lowest level of certificate. The advantage
to this method is that an actual certificate authority exists that can
attest to the world that you're you, based on the certificate they issue
you. The level of assurance varies according to the type of certificate.
At the $10/year level, it offers some minimal assurance, because Verisign
basically issues the certificate upon request. Higher (and more expensive)
levels of certificate actually involve Verisign doing some investigation
to verify that you're you before they issue the certificate.
The PGP mechanism doesn't depend on a central authority. Anyone can
register keys with a public key server. In theory, this means that I could
masquerade as someone else, simply by creating a key set for that person's
name and email address and registering it with a key server. In practice,
PGP gets around that problem by allowing you to sign other people's keys
and allowing them to sign yours. As you accumulate others' signatures,
they are also attesting to the world at large that you're you. As you sign
others' keys, you're attesting to the world at large that they are who
they say they are, based on your personal knowledge. All of this works
pretty well, and without the need for a central authority or an annual
PGP is now marketed commercially by Network Associates/McAfee, but you
can download an uncrippled free version for personal use only from the NAI
web site at:
The file name is PGPfreeware602DH.exe, and it's about 5.8 MB. You have
to go through the usual garbage to prove you're you and that you live in
North America. Once you've got it, you run the executable to unzip its
contents to a scratch directory. Then run Setup.exe, which installs PGP
Overall, the product looks pretty good. I ran into a problem when I
installed it, though. I'd installed an version 5.0 back when I was using
Eudora, and then version 5.5 when I was using Outlook 97. I still had my
public and private keyrings, but I had no idea what my pass phrase was
from back then, and without the pass phrase I couldn't use PGP with that
set of keys. I usually file information like this in my Reference folder,
but I couldn't locate it. Eventually, I concluded that it would be easier
just to create a new set of keys.
That was when the second problem arose. I installed PGP on my main
workstation, kerby, which accesses the Internet via the WinGate
proxy server. Unfortunately, PGP (at least in the freeware version) has no
obvious way to accommodate a proxy server. It wants direct access to the
Internet, both to publish your own keys and to retrieve others' keys. I
was able to get around this to some extent by installing PGP on sherlock,
which does have a direct Internet connection, and publishing my own keys
from there. I was then able to import keys to kerby from sherlock,
which kind of works, but makes things very clumsy. In short, unless I'm
missing something, PGP works fine on a PC that has a direct Internet
connection, but is unusable for all practical purposes on one that
accesses the Internet via proxy.
But even with the limitations of access by proxy, I was able to play
around with PGP on kerby. It seems to work properly and is well
integrated with Outlook 98. I used it to encrypt and sign some test
messages, and it worked as expected. One minor downside. I noticed that
with PGP installed, Outlook 98 slows down somewhat. Normally, to change
from folder to folder within Outlook 98, you just click the folder name
and see the messages that the new folder contains displayed immediately.
With PGP installed, there is a short lag before the message list in the
new folder is displayed. During that lag, Outlook displays a status bar
dialog that lasts a fraction of a second. It's not a major delay, but it's
rather disconcerting to see that status bar each time one changes folders.
So, I removed PGP from my main workstation, at least for now. But if
you have a direct Internet connection, and if you want to be able to sign
and/or encrypt your email messages, PGP is definitely worth looking into.
* * * * *
And this from Tom Syroid:
Our site is up, members.home.net/the.syroids
-- the only target date I've met in several weeks. The cost in hours was
high, but affordable; my return was several I-didn't-know-that's, a
multitude of Mmmmm's, and a few Ah-ha's. Valuable stuff. I know you're
plate's full, but if you're out and about browsing, have a peek. I could
use some insight from an experienced author/publisher, like yourself,
who sees beyond the flashy graphics.
The book review progress well. I'm a day
behind where I wanted to be, but you'll have the first chapter by
tomorrow. Truly excellent material.
I spent half an hour or so last night jumping around your
site and reading your content. You've obviously done a lot of work on it,
and it looks good. Just wait until you find out how much work it takes to
keep it up. :)
* * * * *
Gary M. Berg did send me one of those strange email messages he
mentioned last week--the ones that open an instance of Internet Explorer
without asking. When I put the cursor over it and Outlook did the
autopreview, it did indeed fire up IE, which started to pop cookie
warnings up right and left. I denied all of them, and finally ended up
getting the IE window closed by right clicking it on the task bar and
choosing close. In the mean time, it had started playing an audio file. I
finally deleted the message without ever opening it and shut down NT to
kill everything. Once I brought it back up, went in and cleared out all my
cookie files, used LiveUpdate to update both the Norton Utilities for
Windows NT program and its virus scanning files, and did a virus scan. No
damage apparently was done, but that's definitely a mail message I
wouldn't want to see.
* * * * *
And this from Gary M. Berg:
> Well, that one you sent me was
Sounds like your experience was almost
identical to mine, step by step. I got the music, had to right-click on
the task bar, etc. And my reaction was like yours, to do a full virus
> No damage apparently was done, but
that's definitely a mail message I wouldn't want to see.
I agree; I've gotten about 3-4 of them over
the last couple of months. So now I wish I knew what to make of the
message. I'm going to send along the info I had copied into the body of
my message but it shouldn't cause you any grief. I even made sure to
force this message to be plain text, so it shouldn't do anything weird.
This was what I think the body of the message contained:
I found that this was all of the internet
header info that I could get Outlook 98 to tell me about:
Received: from 184.108.40.206 by
mail.chemineer.com with SMTP
(Microsoft Exchange Internet Mail
Service Version 5.0.1458.49)
id VZQ2MKPR; Mon, 28 Dec 1998
Received: from mx.infod.com
([10.63.11.22]) by www.hn.cninfo.net
(Netscape Mail Server v2.02) with
SMTP id AAA25654;
Mon, 28 Dec 1998 18:41:58 +0800
Received: from 220.127.116.11 by
netaddress.com with SMTP; 28 Dec
04:18:18 GMT 1998
nullReceived: from unknown (HELO
by mx2 with SMTP; 28 Dec 1998
From: firstname.lastname@example.org <mailto:email@example.com>
To: firstname.lastname@example.org <mailto:email@example.com>
I couldn't really tell from that why it
might have ended up in my mailbox at work; I am set up to receive mail
sent to things like firstname.lastname@example.org, email@example.com, and maybe a
couple of more. Any explanations would be helpful. The attach.htm that
was attached to the message contained:
Well, I spent a couple of minutes looking through the
headers and attachment text, and I have no idea what's going on. I don't
much like it, though. Perhaps one of my readers can tell us more.
* * * * *
And it's back to work on the book. Barbara is now officially
self-employed, and will be helping me on it. She still has her office torn
to pieces, however, and will probably spend much of the day getting it
into shape so that she can get to work.
* * * * *
And this from Dan Bowman:
Good day sir,
The link shown in Monday's area (daybook,
week of Jan 4) for Tom Syroid's home page has a "ttgnet.com"
embedded in it yielding a 404. Just an FYI. I found you though
Pournelle's site and enjoy your commentary and computer musings.
Thanks. I'd actually found that myself a little while ago,
but hadn't gotten around to republishing yet with the fix. I'm not sure
why FrontPage did it that way. I just cut Tom's message and pasted it per
my usual, but there was no "http://"
in the pasted string. At any rate, it's fixed now. Thanks for the kind
January 5, 1999
My indoor-outdoor thermometer tells me that it's 14º F (-10º C) this
morning, which is pretty cold for around here. With the breeze, it must be
well below zero Fahrenheit. Even the dogs didn't want to spend much time
outside this morning.
* * * * *
And I see in the morning paper another outrage against the taxpayers.
The local social services department apparently accidentally wrote a check
for about $28,000 to a woman who was instead supposed to have received a
$50 welfare check for a utility subsidy or something similar. She promptly
took the check to a car dealer and used it to buy a used car for $16,000.
The car dealer wrote her a check for the $12,000 difference which she
promptly blew on furniture, clothes, and trips. All of this apparently
took place nearly a year ago. A judge ordered her to give up the car in
September, which she hasn't yet done. Apparently, the best they can do is
enter a notation in her record that if she ever comes into another
windfall she must repay the money. It looks like no one will have to pay
the money back and no one will go to jail for this theft.
Now, it seems to me that everyone involved here is culpable. The woman
certainly must have known that she wasn't entitled to this much money. Her
actions--spending the money as quickly as possible--indicate that she knew
that there had been a mistake. The car dealer must have realized that a
$28,000 welfare check was unusual, to say the least. All they had to do
was call the social services department to verify that that check was
valid, which they apparently didn't do. And then there's the social
services department itself. What kind of control mechanisms would allow a
$28,000 check to pass unnoticed? Who was responsible? The paper doesn't
say, and makes no mention of anyone being disciplined for this, let alone
losing his job.
There are at least two reasonable actions that could be taken to
recover the taxpayers' money, and apparently aren't being taken. First,
the social services department should be attempting to recover the full
$28,000 from the car dealer. That company knew or should have known that
that check was questionable. After all, how many businesses would take a
$28,000 third-party check in payment for a $16,000 item and then write a
$12,000 check for the difference? Second, they should make this woman
ineligible for continuing welfare payments, on at least two grounds.
First, by writing her this $28,000 check, they've essentially pre-paid any
welfare benefits to which she would have been entitled. Second, her
$28,000 income should certainly have made her ineligible for welfare in
But it looks like nothing much is going to happen. The woman is
supposed to turn in the car, if she ever gets around to doing it. She'll
continue to receive her welfare checks. The car dealer won't be penalized.
And it looks as though the social services department will write this all
off as a harmless mistake that anyone could have made. Geez..
January 6, 1999
And another really cold night last night. I forgot to reset the
thermometer so that it would record the minimum temperature, but it was
supposed to get down into the lower teens. I blame it all on global
* * * * *
I got quite a bit done on my chapter yesterday, about 2,000 words or
so. That's especially good given what I was writing about--chipsets and
memory timing. I was able to get so much done for several reasons: first,
I strictly limited myself to an hour's worth of answering mail and working
on this site. Much as I enjoy receiving mail from everyone, reading it
(and particularly, answering it) take a lot of time. Second, Barbara had
pretty much finished up her deep cleaning. I can't concentrate well when
she has that 25 horsepower vacuum cleaner going, even when I have the door
to my office closed.
Barbara had to leave in late morning to take her dad to the hospital
for tests. It appears that everything is okay, although we won't know for
sure until the lab results come back. Her dad is in remarkably good shape
for 76 years old. He does all his own yard work, and walks 18 holes
regularly. Still, at 76, perhaps the time has come for him to take things
a bit easier.
When she left to take him to the hospital, she found her cell phone
battery was dead and asked to borrow one of mine. That brings up an issue
that I'm going to have to do something about. We each have a Motorola flip
phone, and they're both three or four years old. Hers still has the
original Motorola battery, and I have a couple of aftermarket batteries
for mine. All of our batteries have gotten to the point now where they
don't hold much of a charge. They're all Ni-Cd units, and although some
claim that the "memory effect" doesn't exist for Ni-Cd
batteries, my experience is that it does. All of these batteries have been
properly cared for--periodic deep discharges followed by full charges, no
partial charging, etc.--and yet they are pretty much unusable now.
So I guess I need to buy some replacement batteries. The problem is
that I'm not sure I want to spend a lot on batteries because I'm not sure
how much longer we'll use the phones. The cellular deal I have is pretty
good for around here. Forty bucks a month for both phones, each with its
own number and a third number that rings both of them, and 150 minutes per
month of airtime. But we've never used as much as one hour of airtime in
the years that we've had this plan, and with Barbara working at home now
I'm not even sure we need both phones. I'm thinking we might just get one
DCS/PCS phone on a $20/month deal and share it.
But in the meantime, I need to get a battery or two. We were just at
the cell phone place a couple of weeks ago to get our phones reprogrammed
for the new area code. They only had one flip phone battery on display,
and that was not an original Motorola battery. They wanted $35 for it,
which seemed a bit expensive, particularly since it didn't have any specs
on it for amp hours. We only have the original slow chargers, which I
suspect limits me to replacing these old batteries with Ni-Cd ones.
What I'd really prefer is either Ni-metal hydride or Lithium ion, but
those batteries are quite expensive and require special chargers.I don't
want to spend that much on batteries and chargers if I'm likely to move to
a different technology soon. So I guess what I need to do is find a good
source of Ni-Cd batteries and pick out something with a decent amp-hour
rating. Any suggestions will be gratefully accepted.
* * * * *
And here's another minor aggravation about Outlook 98, a product which
I generally really like. There's a folder icon on the toolbar that lets
you move a message from the Inbox to another folder. The drop-down menu
this icon displays has the ten most recently used folders, with the most
recently used at the top. There's a final entry in that menu that lets you
display the entire folder hierarchy when you need to move a message to a
folder that's not on the list. That's all well and good, but it has two
First, the list is apparently hardwired to list a maximum of ten
folders. If you have ten folders already showing and move a message to a
folder that's not on the list, that folder moves to the top, and old
number ten drops off the list. Why only ten? I'm running Outlook at
1024X768, and I have room for many more folders in the list. Why can't
they make the number of folders to be listed a settable option?
Second, and more important, the list changes dynamically as you move
messages to folders. For example, when I move a message from my editor at
O'Reilly to the O'Reilly folder, that folder moves to the top of the list.
I'm sure that the programmers thought that this would be a helpful
feature, and at first glance it appears to be a good idea. The problem is
this: I never know where the folder I want will be in that list, so I have
to scan the list each time to find the folder I want.
What they should have done, at least as an option, is allow you to
decide yourself which folders you want listed and in what order. That
would be more in accord with how people actually work. When I need to move
a message to the O'Reilly folder, I know that it's about 2/3 of the way
down my folder list, and I can select it without having to read the entire
list to find it. It's the same place every time. Failing that, even an
alphabetized list would have been better than the dynamic list they use
* * * * *
And Barbara is getting started today as my assistant. I've sat her down
with a stack of PC hardware books so that she can get a gestalt of PC
hardware. By the end of the week, she should know more about PC hardware
than most people. Then I'm going to have her put together a shopping list
of the components needed to build a PC for herself. I'll check the list,
and then send her off to buy them. When she gets home, I'll have her build
her own PC. I'll let her make her own mistakes and learn by the process.
In a month, she'll be able to help me do chapters. Because she'll still be
a relative PC hardware novice, she'll also be in a good position to help
me see that viewpoint.
January 7, 1999
And my search for replacement batteries for our cell phones has been
successful. I got mail from many readers suggesting many 1-800 and
web-based battery specialist resellers. The most commonly suggested one
and so I started with them. They have a big banner on their home page
quoting John C. Dvorak:
"1-800-Batteries has a slew of power related gear and seemingly
every battery imaginable for laptops, cell phones, camcorders, you name
it. The prices look unbeatable!" -- John C. Dvorak, PC Magazine
As it turns out, Dvorak is full of it, as usual. They have every
battery imaginable, all right. But their prices aren't particularly good,
and they appear to be selling after market batteries rather than original
Motorola replacement parts. Or, if they are original Motorola, I couldn't
tell from looking at the part numbers. But the prices were what really got
me. For example, their extended talk battery, part number 19069, described
as "Battery Pack, NiCad, 6V, 1000mAH, 90 min" was $69.00. Their
standard battery, part number 18995, described as "Battery Pack,
NiCad, 6V, 800mAH, 120 min" was $29.00. Those prices didn't strike me
as particularly cheap.
So I decided to go check out the Motorola
site just as a point of comparison. After all, we all know that buying
stuff direct from the manufacturer is the most expensive way to go, right?
Wrong. The battery that came with the phone is an SNN4019 700 mAh NiCd
battery rated for 90 minutes of talk time and 15 hours of standby time. I
couldn't find that exact battery on the Motorola site, perhaps partly
because they've changed their model numbering scheme. I did find a
standard Motorola battery, number 99561, that cost $17.65. The extended
talk NiCd, number SNN4057 (or 99314 with the new numbering scheme), costs
$22.94. This is an 1100 mAh NiCd battery rated for 150 minutes of talk
time and 24 hours of standby time.
I checked a bunch of web sites that sell cell phone batteries, and most
of them wanted $40 for the standard battery and $60 for the XT battery.
Motorola is selling them direct for less than half that. I looked at every
battery place I could find on the web, and didn't find anyone selling real
Motorola batteries for anything near what Motorola sells them for direct.
Even the no-name aftermarket batteries were usually more for the
equivalent specs. So, after making absolutely sure that this was the
correct battery, I went ahead and ordered two 99314 high capacity NiCd
batteries directly from the web site.
Their web ordering mechanism works much as you'd expect if you've ever
ordered anything from Amazon. After I provided my name and shipping
information, the site prompted me for my credit card number, which I
entered. At this point, I knew that I was committing to two batteries at
$22.94 each, for a total of $45.88, but I had no idea what shipping was
going to be. When I completed the order, I found out the good news and bad
news. The bad news is that they take three or four days to prepare the
order before shipping it, and that they do charge state sales tax. The
good news is that shipping via FedEx next day air is included in the
price! So I ended up with two high-capacity OEM Motorola ni-cd batteries
for a grand total of $48.64.
If you have a Motorola cell phone or other portable communication
device and you need to buy accessories for it, look to the Motorola web
site first. You'll get actual Motorola parts, and they'll be cheaper there
than the OEM or aftermarket stuff you can buy elsewhere. Highly
January 8, 1999
Got mail from David Rogelberg, my agent, last night, telling me that he
and O'Reilly had finally hammered out terms for the book that Jerry
Pournelle and I are co-authoring. I hope we can get started on it in the
next week or so. I'm really looking forward to working with Jerry. In
particular, he'll be doing a lot to rewrite, expand, and improve first
draft material I've written, so it'll be interesting to see what my stuff
looks like after it's been "fixed" by a top-notch author.
* * * * *
And FedEx showed up this morning with my batteries from Motorola.
They're exactly what I wanted. Apparently, Motorola tells people that they
take up to three days to ship just in case they're delayed. In fact, I
ordered the things at 3:23 p.m. Wednesday afternoon and they showed up
here at 9:18 Friday morning. They're noticeably heavier than the standard
batteries that came with the phones. That's okay, though. I don't mind an
extra couple of ounces to give me 2.5 hours of talk time and 24 hours of
standby. I just put them on the charger to give them their initial 14 hour
* * * * *
And this from Chuck Waggoner:
I agree with your assessment of Outlook.
Actually, in my opinion, it would be more logical for them to drop the
most recently used folder to the BOTTOM of the 'move' list, as having
just used it, it's far less likely to be used next.
Overall, I can't believe Microsoft gives so
many options, but yet somehow excludes the very ones I would like to
change, enforcing their own choices instead, which obviously have not
been derived from practical experience. It's really maddening that, in
their various text viewers and word processors, I can't even specify
whether--much less how many--text lines repeat on a page up or down
command. And in Outlook, I would like to specify exact clock times of
when Outlook checks my mail, so that IE 'subscription' downloads and
email checks don't occur at the same time (because if they ever do,
whichever one is done first will hang up on the other). Outlook provides
only an interval from Outlook start-up. The only way I've found to deal
with this, is to have Outlook started daily by the Task Manager.
Combined with the proper interval, that can keep Outlook from logging on
at the top of the hour, which I have reserved for the subscription
You might find it easier, as I have, to
simply keep the Outlook folder list (tree) open on the left of the
screen à la Windows Explorer. I have created half-a-dozen folders for
more frequent and important contacts under the 'Inbox', which I keep
expanded all the time. The rest of the mail folders I've placed under
'Contacts', which I only expand when needed. Then I just drag mail to
the proper folder. I don't know whether, in final analysis, it saves any
mouse clicks over hitting the 'move to folder icon', but it works for
I understand that Outlook 2000 has considerable
improvements, but I haven't looked at it yet. Actually, Microsoft is
usually better than most third-party software companies at providing
customization options. If you into a Microsoft product, like FrontPage 98,
that is sadly lacking in customization options, it's generally a safe bet
that it's a product that Microsoft bought rather than developing in-house.
If I didn't know better, I'd think Outlook 98 belonged in that category.
I can't comment on the periodic mail check issue, because
although I have my copy of Outlook set to do mail checks every five
minutes or so, I don't let Outlook dial. Instead, I tell it that I'm
connected to a network with Internet access (via WinGate). If the Internet
connection happens to be down when Outlook wants to do a scheduled mail
check, WinGate dials the phone automatically.
And you're right about keeping the folder list up. I've
done that since day one, and often drag messages to the appropriate folder
rather than using the menu to move them.
* * * * *
And FrontPage 98 just exhibited another strange bug. Bo Leuf mentioned
to me that my current Daynotes page format used one big table with two
columns and seven rows to contain the seven days of the week. Something I
wasn't aware of was that this slows down page loading because the entire
page must be downloaded and rendered before the browser can display
anything. That's what's responsible for the blank page being displayed for
anything up to a minute or so when you refresh this page. Apparently, you
can get around that, at least with some browsers, by making each day its
own separate two column by one row table. That way, the Monday table can
download and render while the Tuesday through Sunday pages continue to
download in the background.
I guess that would help infrequent visitors, because they could read
the top of the page while the bottom continues to download and render. It
won't help frequent visitors much, because they've already read the
earlier entries, and will have to wait for the later ones to render, just
as they always would. I guess it's marginally better to look at a page
with some text on it, even if you've already read it, than to sit staring
for a minute or so at a blank page.
At any rate, I always have the next week's page set up in advance.
Every time I start a new Daynotes weekly page, I first copy that week to
next week, ensuring that I have a blank template available for the
following week. So, I already had the 1/11/99 page set up. Instead of
risking messing it up, I copied it to junk.html and did my experimenting
on that page. I got the individual tables set up for each day, and added a
"top" link to the bottom of each of the left columns. I saved
the page, and it appeared to work just fine locally.
At this point, although I hadn't published it yet, I decided that the
new page format would probably work just fine. So the next step was to
delete the old 0111RTDN.html and rename junk.html to 0111RTDN.html. I
deleted the old file without any problems and renamed the new version.
That's when something weird happened. I changed the file name in FP
Explorer and hit Enter to save the change. At that point, FP98 popped up a
dialog that told me that the current file referenced two other files, and
asked if I wanted to fix those links. I told it Yes, which was a mistake.
FrontPage 98 immediately created two new files, one named
NewPage(2).html (or something similar) and one named EmailRobert.html (or
similar). I wasn't sure why it created those files, so I double clicked
the first one from FP Explorer to see what it contained. FP Editor
generated an error message saying that it couldn't retrieve the file. Same
thing on the second one. The two files were listed in FP Explorer, but
apparently didn't exist on disk. Okay, back to FP Explorer to look a
little closer at the file properties. Both new files showed a date of
12/31/69 at 7:00. I deleted those two new strange files, and everything
appears to be working normally now. You can look at the new file format
(although there's nothing in it yet) by clicking here.
There's an old saying something like, "Fool me once, shame on you.
Fool me twice, shame on me. Fool me thrice, it must be a Microsoft
product." This is the third time that FrontPage has told me that
there were problems and asked if I wanted it to fix them. Each time, I
told it Yes, and each time it's screwed things up. The moral here is: If
FrontPage 98 asks you if you want it to fix something, tell it NO.
January 9, 1999
DHL showed up yesterday afternoon with four Maxtor DiamondMax Plus 2500
10 GB hard disks. I'm really looking forward to getting these installed
and working with the Promise FastTrak IDE RAID controller. We're waiting
on some other eval units from Intel that we'll use to build a low-end
system around, including an Intel RC440BX integrated system board and a
Celeron-A processor. I think I'll temporarily install the Promise FastTrak
card and the Maxtor drives in that new box, because I want to wring them
out under Windows 98. Once I've done that, I'll probably move the card and
drives into bastet for a long-term evaluation. That box currently
dual-boots Windows NT Server 4.0 and Windows 98. I was thinking about
converting it to a resource server, but I may just use it as a test-bed
I used Maxtor drives years ago, and was always favorably impressed with
them. My first one was either a 40 MB or an 80 MB IDE, which should tell
you just how long ago that was. Maxtor has a major presence in retail
operations--Best Buy and similar places--but is not heavily represented in
the kinds of places that I buy hardware. Those places carry mostly Seagate
and Western Digital, so that's what I"ve been buying mostly for the
last few years.
But I've always liked Maxtor. They make solid drives. Over the years, I
must have bought at least a couple hundred systems with Maxtor drives in
them for myself, my companies and my clients. I can't remember even one of
those drives dying. That's better than my experience has been with either
Seagate or Western Digital. So when I decided to request some eval units,
I thought it was about time that I paid some attention to Maxtor. And from
the specs on these Maxtor drives, they should be real barn-burners. We'll
* * * * *
And I ran into another little FrontPage 98 surprise last night. This
one started when I was sitting looking at my home page and noticed that it
said January 8, 1999 at the top of the page, but
Copyright © 1998 by Triad Technology Group, Inc. All Rights
Reserved. at the bottom. I figured that should be easy enough
to fix. All I needed to do was change it to read Copyright ©
1998, 1999 by Triad Technology Group, Inc. All Rights Reserved.
FrontPage uses a concept called Shared Borders. There are four Shared
Borders--top, bottom, left, and right. The content of each of these four
Shared Borders is global to the web site. That is, anything you enter in,
say, the bottom shared border automatically appears at the bottom of every
page in the site that has the bottom Shared Border enabled. My copyright
notice is in the bottom shared border, so changing it in any page
simultaneously changes it for all pages. I made the change.
It seemed to work exactly as expected. It took half a minute or more to
save the page because FrontPage 98 was updating all pages rather than just
the one I made the change in. So far, so good. The first surprise was when
I closed FrontPage Editor, which redisplays FrontPage Explorer. I expected
all of the files to have their date/time stamps updated to reflect the
change. They didn't. Only the page I'd actually changed showed the later
date and time.
Then I got to thinking. Jerry Pournelle added a copyright notice to his
site using the bottom Shared Border. I hit his site and, sure enough, his
pages still claimed only the 1998 copyright date. So I emailed Jerry to
let him know he needed to update his copyright notice. While I was
visiting his site, I went to check out new mail. I found a message
describing software that automatically checks all of the links on a web
site. The sender of that message had included a link to a document that
listed all of the broken links on Jerry's site. Among them were several
broken links listed as http://www.ttgnet.com/index.htm.
The message even specifically mentioned that link as an example of one
that should end in .html instead of .htm. Now, I know that I entered that
URL myself by using a pick list within FrontPage. There's no way I could
have made a typo, because I didn't type in the URL manually.
That's really strange, I thought, because I was certain that the name
of the page was indeed index.htm. After all, I call the thing up in
FrontPage every day and then save it to reset the date. I entered http://www.ttgnet.com
in my browser. My site came up fine. So I explicitly entered http://www.ttgnet.com/index.htm.
The server returned a Page Not Found error! Okay, maybe it's my
imagination and the page really is named index.html. I explicitly entered http://www.ttgnet.com/index.html
into my browser, and the site loaded.
At this point, I was beginning to think I was losing it. Perhaps heavy
doses of Ginkoba were in order. Just to make sure I wasn't crazy, I fired
up FrontPage. Sure enough, there is a file named index.htm, and there most
emphatically is not a file named index.html. Well, I thought, perhaps the
FrontPage Explorer Files View is screwed up for some reason. So I closed
FrontPage and used Windows NT Explorer to look at the directory that
contains my local copy of my web site.
When the directory list displayed, I click the Modified column header
to sort the most recently changed files to the top. Oh-oh. The file
index.htm was missing entirely! Or so I thought, until I looked more
closely at the date/time stamps. Every single *.htm and *.html file in
that folder was showing a date/tiime stamp of 1/8/99 at 6:43 p.m., exactly
the time when I updated the Shared Border.
At that point, I thought for a moment that I had it figured out.
FrontPage Explorer updates the file list each time you open it, but does
not keep that list dynamically updated. For example, if I have FP Explorer
open and use another program to create an image file in the local web
directory, that file doesn't show up in FP Explorer until I've closed and
then re-opened it. I figured that must have been what happened. FrontPage
Explorer wouldn't show the updated date/time stamps for the files until I
re-opened it. Well, it was closed by then, so I went ahead and fired up FP
Explorer again. Nope. It still shows the original date/time stamps. I
haven't published since I changed the Shared Border, but my guess is that
I'll have to republish every HTML page because of that single change to a
shared global setting. We'll see when I publish this in a few minutes.
This is all particularly aggravating because I have never
entered the extension .htm or .html when saving a file. FrontPage 98 does
that for me. And apparently it is not consistent in how it does it. Most
of the time, it uses the .html extension, but sometimes it apparently just
decides to use .htm instead. I have probably 150 individual HTML pages.
All but about maybe a dozen have .html extensions, but those dozen use
.htm. I'm thinking about changing the .htm files manually to use the .html
extension, but every time I do something like that, FP offers to
"fix" the problems that result. If I tell it to fix the problem,
it bollixes up everything terribly.
Sometimes I like Microsoft better than other times.
* * * * *
And this from Tom Syroid:
There's an old saying something
like, "Fool me once, shame on you. Fool me twice, shame on me. Fool
me thrice, it must be a Microsoft product." This is the third time
that FrontPage has told me that there were problems and asked if I
wanted it to fix them. Each time, I told it Yes, and each time it's
screwed things up. The moral here is: If
FrontPage 98 asks you if you want it to fix something, tell it NO.
advice, my friend.
2000 is typically Microsoft -- and I immediately developed a love/hate
relationship with it. It's easy to use, but hides it's power in
unexpected places and innoculous dialogue boxes. There is an important
distinction in our experiences, however (product version aside)-- you
are an experienced user of the product. I am both novice in FrontPage,
and novice in what exactly I'd like to do with it. This adds a new and
somewhat adventurous dimension to my decision tree. Am I asking the
program to do something impossible? Is there a better (read less
frustrating) approach FrontPage is trying to hide from my novice-ness?
What exactly is it I want to accomplish here, and am I looking under the
wrong rock or in the wrong field? You get my drift... Doom on the reader
who hasn't the skills to pick up on new products quickly, and 'Feh' to
Microsoft's legendary useability improvements.
pertinant: Bo and I have had some long and insightful discussions this
week over basic web design, browser gottchas, and general formatting
cautions. I'm still somewhere between 'lost and found' but still chewing
on much of the material. I like your new Daynotes format. It is similar
to changes I'm mulling over, and for similar reasons. My idea used 2
side-by-side tables per day, though (one for navigation, one for text
body) and I wonder aloud what advantages there are--if any-- to your use
of one table with columns. I'm trying for KISS above all and wonder if
the added format of a column does not increase complexity beyond
necessity. Your view on this is appreciated...
question: How do you deal with the inevitable formatting problems of
cutting material from reader's mail and pasting them to your pages?
Again, ask Bo of my struggles. Such a simple and common task is very not
so the way I've been approaching it (why just going CUT in one document
and PASTE in the other--isn't that they way it's supposed to work??). I
have a whole new appreciation for Pournelle's struggles of a few months
2000 is somewhat more polished, and has added some good useablity
features, but overall remains much as it was in '98 for small
frustrations. E.G., the move menu has the same quirk you mention. With
all the input we beta testers gave them over OLK98, I'm amazed that
OLK2000 is so irritatingly similar to OLK97. I plan to get a preliminary
review up on the whole Office 2000 package RSN. As you may have gathered
from my Insights this week, productivity is complete gone from my field
of vision, and just treading water has become the order of the day.
Well, let's not forget
that FrontPage allows us to do an awful lot that we couldn't otherwise do
without expending a lot more effort. I, for one, have no desire to learn
anything about HTML. All I want to do is publish pages, and FrontPage does
a good job of making that pretty easy. All in all, I wouldn't trade it for
anything else I've seen or heard about.
As far as your layout, go
with the tables. They're not a significant complication, and they do make
things easier on readers. Anyone who's still using a browser that doesn't
support tables properly deserves whatever he gets.
As far as cutting and
pasting mail, I don't have much problem
with formatting mail for publication. People who, like you, use HTML
format make it very easy. I just cut the message and paste it in. The only
minor problem occurs because my default in Outlook is to use Arial and
blue, while my default in FrontPage is "default" and black.
That's easy enough to fix, though. I just paste it in as Arial/Blue and
then click on the HTML tab at the bottom left of the FrontPage Editor
window. Each paragraph begins with a font tag and a color tag. I just
delete those and the text returns to default settings. It takes only a few
seconds per message. For mail that arrives as plain text, the job
is even easier. I just highlight the text in Outlook, copy it, and paste
it into FP Editor. It gets pasted as default font and color. Once all
that's done, I just use select the whole message in FP Editor, click the
Indent button to indent it, and use the drop-down font list to change it
to Courier New.
As far as the new
generation stuff, I'll probably end up
giving FP2K a try, and OL2K as well. I just don't much care for their
"drop dead after 50 uses" policy. We'll see.
January 10, 1999
Today is our day to work around the house. Barbara's getting started on
cleaning house, and I need to get the laundry done.
* * * * *
And this from Dave Farquhar:
First I wanted to thank you for your advice
on the computer lab I'm working on. As Jerry said, the Linux-vs.-Windows
argument is a very complex one. Since it's not entirely my decision, the
three of us working on the project will need to sit down and talk it
out. The flood of advice that's come in has been unbelievable.
As for the Maxtor DiamondMax Plus drives,
they're fantastic. I can't say enough about them. They're a little more
expensive than Seagate and Western Digital drives of similar capacity,
but they're also much faster. It's really a shame that people go pay
$600 for a Pentium II-450 CPU, then put an el-cheapo hard drive in it
that kills performance. In my experience, getting the fastest hard drive
you can within reason gives you far more bang for the buck. And, well,
if you're looking to build a 400 MHz system, you can probably afford to
spring the extra $30 it'll take to put a 7200-rpm hard drive like a
DiamondMax Plus in it. (I'm still unconvinced that SCSI is worth the
high premium except in servers, but I'm completely sold on buying the
fastest EIDE/UDMA drive available, regardless of cost.)
The DiamondMax Plus is the first Maxtor
drive I've owned -- I've always bought Quantum drives and had excellent
results -- but I can vouch for Maxtor reliability in my work experience.
I've probably seen equal numbers of Maxtors and Western Digitals, and
while I've had a few Western Digitals fail, the Maxtors and Quantums
have kept going. Not that Western Digitals are junk, but the Maxtors and
Quantums do seem better-built.
Yes, I saw the follow up on Jerry's site. Good luck with
your computer lab project, and thanks for confirming my impression of the
The computer lab Dave refers to is one that he is helping to set up
with donated equipment for an inner-city church to give disadvantaged
children access to PCs for homework and general computer literacy. It
appears that they may set up workstations running Linux and Corel
WordPerfect, but Dave is not an experienced Linux person. I suggested to
Dave as an alternative that he solicit corporations for donations of
Wintel boxes with MS Office already installed. To meet Y2K compliance
goals, many corporations are essentially discarding large numbers of
relatively recent PCs that would be suitable for Dave's computer lab.
Anyone who has done something similar and is willing to offer help or
advice can email Dave at firstname.lastname@example.org
* * * * *
And this from Channing:
I cannot get through to the Promise website,
hasn't been up for the past 2 days, anyway, does the card support W98
out of the box or on their website. First time visiting your site, neat
site. Thanks in advance.
When I first read your message yesterday, I had the same
problem getting to the Promise
web site that you did, so I decided to delay my response until I could
find something out for you. I looked on the FastTrak box. I just received
this a few days ago, and was hoping it would list Windows 98 support. It
lists DOS, Windows 3.x, Windows 95, Windows NT, NetWare 3.x/4.x, and OS/2
Warp, but not Windows 98 (or Linux). This morning, I tried the Promise
site again, and it's now responding. It lists the same operating systems
though. Although Windows 98 isn't listed, I'd bet that it's either
supported natively (it's not really much different from Windows 95) or
that Promise will have a Windows 98 driver available for download.
Ultimately, I suppose the only way to be sure is to try it under Windows
98, and I plan to do just that.
Coming Soon (I hope)
Here are some things that are currently on my to-do list.
I may start some of them this coming month. It may be a while before I
start on some of the others, either because I don't yet have everything I
need, because interdependencies make it necessary to do other things
first, or simply because other work takes priority. But I'll get to all of