|
Daynotes
Journal
Week of 7 April 2008
Latest
Update: Saturday, 12 April 2008 08:33 -0500 |
09:56
- Fixing Firefox.
One
of the things that's always annoyed me about Firefox is its poor cookie
management. They've even unimproved it in later releases. It used to be
there was a checkbox on the cookie dialog screen. You could check that
box and any cookies you deleted would be automatically added to the
list of sites from which Firefox should never accept cookies. In a
later release, they removed that checkbox, leaving typing in the cookie
name the only way to add that cookie to the do-not-accept list. (Well,
if you know how, you can add a couple of items in about:config that
restores the auto-add functionality, but it's no longer doable from the
main GUI configuration screen.)
Periodically,
I go in and look at my cookies list, deleting the obnoxious ones. I did
set up about:config to auto-add the bad sites to the banned list, but
even so it's a pain to maintain that list. So, the other day I was
pruning my cookie list when I got to wondering. There's a checkbox in
the cookie dialog that is marked active by default called Accept
cookies from sites. I'd always left that checkbox marked, because the
alternative seemed to be to reject all cookies. But what if I cleared
that checkbox and added the cookies I did want to accept to the
exception list?
It seemed worth a try, so I cleared the
checkbox, halfway expecting Firefox to wipe out my good cookies. It
didn't. So I went into the exception list screen, which at that point
contained a list of bad cookies, all marked Block. I deleted all
cookies in that list and started adding my good cookie sites, marking
each of them Allow.
Sure enough, whitelisting works. Firefox now
automatically rejects cookies from any site that's not on the exception
list. Most sites work just fine without cookies enabled, and most of
the sites that I want to be able to use cookies are already in my
allowed list. If I come across another site that requires cookies to
work properly and I want to use that site, I'll simply add it to the
exception whitelist. For some of those, I'll mark Allow for the site.
For others, such as the New York Times, I'll mark Allow for Session for
the site, which should cause Firefox to delete cookies from that site
as soon as I close the browser.
What we have here is a classic
example of a badly-chosen default setting combined with a poorly-named
checkbox. Firefox should default to not allowing cookies, with a note
in the dialog that tells the user to add allowed sites manually.
Granted, that's a bit more work for the user, but it's a one-time
effort and the benefits are large. With Firefox configured this way,
privacy is greatly enhanced, and there may be security benefits as
well.
With cookies tightly controlled and AdBlock Plus with the
FilterSet.g updater and Flashblock installed, browsing is a lot more
peaceful, a lot more private, and a lot more secure than using Firefox
in its default configuration.
Next up, I'm thinking about
blocking TCP/IP connections from all but first-world countries. There's
no reason I need to get to Russia, China, Eastern Europe, Africa, and
other second-, third-, and fourth-world countries, and I don't care if
they can't get to me. In fact, I don't want them to. The
overwhelming majority of packets from those countries and regions are
spam-, malware-, and phishing-related.
Right now, I'm using a
D-Link WAP as my border router. I haven't looked, but IIRC it doesn't
offer a lot of options for blocking IP addresses and inbound traffic,
but I'll check it out. If the D-Link is insufficient, I'll simply use a
Linux box to control inbound and outbound IP addresses.
08:06
-
A bit of excitement last night. We'd hoped to get the TOC and index for
the home chem lab book completed over the weekend, but it wasn't until
after dinner last night that the TOC was finished. As I read through
it, I was horrified to find what turned out to be a major problem.
For
Chapter 14 (gas chemistry), the TOC listed only lab sessions 14.1,
14.2, and 14.3. Lab sessions 14.4 and 14.5 weren't listed. At first, I thought it was no big deal. They could
simply add those lab sessions to the TOC and everything would be fine.
But then I looked at the final layout of chapter 14 and found that
those two lab sessions were also missing from the actual chapter.
Uh-oh. Somehow, that chapter had been truncated during layout, and none
of us had noticed.
There was no way we could just drop those two
lab sessions, but adding them back in would require reflowing and
repaginating the whole book from chapter 14 on, not to mention re-doing
the index. So that's what I, my editors, and the production folks spent
last night doing. The corrected chapter 14 with the two lab sessions
now added back in arrived at about 10:15 p.m. I did a quick proof on
that galley and emailed everyone to say that it looked fine to me. Then
the book designer and indexer went back to work to reflow and
repaginate the remaining chapters and update the index for everything
from lab session 14.4 through the end of the book.
The designer
sent us all email to say that it was all her fault and she felt
terrible. Everyone else involved sent her email to tell her not to beat
herself up. We all saw the truncated chapter, and we all signed off on
it without noticing the missing material either. Every one of us had a
lot of balls in the air, and every one of us had made catches that
everyone else had missed. This time it just happened that none of us
caught the problem until it was almost too late.
As Brian Jepson
commented, better late than never. In the past, books have gone to
press with similar problems that were caught only after the book had
been printed. At that point, the only option would have been to scrap
the printed books and start over. So, as of now, there's no harm done.
The book is as it is supposed to be, and will shortly be going to the
printer. I understand the new pub date is 28 April rather than 15
April, but that had been changed before we found this problem. We'll
have printed copies of the book available in time for Maker Faire,
which is what really matters.
08:25
-
The book is finally finished and off to the printers.
I don't look at my web stats often, because I don't really care, but I
took a brief look at them yesterday. The one thing I noticed
immediately was the user agent section. Firefox dominates, even on
Windows systems. The other interesting thing in that section is the
number of hits from different operating systems. Windows XP was in
first place, but Linux was in second place with about 20% of XP's
numbers. Vista came in a distant third, barely nudging out Windows
2000. After nearly 18 months on the market, Vista's numbers are truly
pathetic. And OS X isn't even an asterisk.
I realize that my
readership skews toward Linux, but even so these numbers indicate that
the real OS war, at least among my readers, is between a fading Windows XP
and a gaining Linux.
10:27
- In all the tens of thousands of tennis balls I've served, I never thought about doing this, from Nick Scipio's site:
Obviously, this is the toss. The cat is on its way up, not
down, the racket is moving back into the "backscratch" position,
and her shoulder turn toward the net has not yet begun. It's also
obvious that this will be a very weak serve. The cat probably won't
even make it all the way to the net. From the position of her hand and
the cat's tail, this is a "rainbow" toss. Rather than moving straight
up and down, the cat is also moving right to left, and probably back
toward the server.
And it's already too far back. By the time
she rotates and brings the racket through the cat, the cat will be
nearly over her head and much too far to the left, which means she'll
be hitting up on it instead of out and slightly down. She'll end up
with a high trajectory, a lot of topspin on the cat, and no pace on the
serve.
She needs to modify her cat toss by positioning the cat
farther to her right and lifting the cat straight up, so that if not
struck it would fall about 6 to 8 inches (15 to 20 cm) inside the
baseline, directly ahead of her racket as it comes through the cat.
That gives her a flat serve with maximum pace on the cat.
09:00
-
Paul and Mary took us out to dinner last night to celebrate finishing
the book. Also, I think, they wanted a chance to examine me closely for
(further) signs of insanity, since we hadn't gotten together since my
announcement that I'd be flying out to California the end of this
month. Frequent flyers both, they had some good advice to help me avoid
being arrested by the TSA.
Right now, I'm doing everything
possible to avoid starting work on our state and federal tax returns.
They're due next Tuesday, so I'll probably get started tomorrow or
Saturday. It's not like I don't have plenty of other things on my to-do
list.
The lab equipment and chemicals for the demonstrations
I'll be doing are now on order. If I have time, I'll do trial runs
before I leave. Otherwise, I'll do trial runs on site before Maker
Faire opens to the public. Actually, I'll probably do that anyway,
since it'd be good for the O'Reilly folks who'll be around to have seen
the demonstrations before I do them in public.
Barbara and Mary
both mentioned that they want to see the demonstrations, and suggested
that I get someone with a camcorder to record them as I do them. We
planned to do that anyway, and post the videos to YouTube.
O'Reilly
also wants me to do a couple of talks, one about the death of chemistry
sets and the other about backyard astronomy. Those'll run 20 minutes or
so each and be given in a small auditorium that seats a few hundred
people. I'll probably just do a bulleted list of talking points for
each of those and speak extemporaneously. But I do need to get my
thoughts organized, so that's one more thing on my to-do list.
This is actually going to be fun. All except the traveling part.
08:23
- Things are looking pretty dim for Microsoft when even the pro-Microsoft Gartner Group starts talking about Windows collapsing.
FTA:
April
10, 2008 (Computerworld) Calling the situation "untenable" and
describing Windows as "collapsing," a pair of Gartner analysts
yesterday said Microsoft Corp. must make radical changes to its
operating system or risk becoming a has-been.
A couple weeks ago, Jerry Pournelle, who has historically been pro-Microsoft, commented in his column
that Windows is doomed if Microsoft didn't take the steps he had
outlined. He commented further that Microsoft wouldn't take those
steps. I pointed out to him that if one accepts the truth of his two
statements, the logical conclusion is that Windows is doomed.
Of
course, I've been saying that for a long time, and nothing I see leads
me to change my opinion. Vista is catastrophically bad and selling in
embarrassingly small numbers. Windows 7, when and if it ever arrives,
is almost certain to be even worse than Vista.
Microsoft
obviously realizes their future in software sales is dim, because
they're doing everything they can to get out of software and into other
businesses. Their hostile takeover of Yahoo, if it actually happens,
will go down as the most ridiculously overpriced buyout since
Time-Warner bought AOL. That's not going to work. A combined
Microsoft-Yahoo still can't compete with Google.
If I were
Ballmer, I'd scrap Windows entirely. Do what Apple did with OS X. Build
a completely new operating system using the BSD kernel with a Microsoft
Windows shell on top of it. Abandon Win32 APIs and backward native
compatibility. As a temporary measure, I'd throw some serious resources
at WINE to provide a compatibility layer for legacy Windows
applications until new native BSD/Windows applications were ready.
Strip out all of the DRM, and release a lean, fast, secure, inexpensive
desktop client with the feature that people want, as opposed to those
the MAFIAA wants.
Obviously, that's not going to happen. Also obviously, Windows is doomed.
08:33
- Ha! Found it.
Most
people probably don't remember exactly what they were reading on the
evening of 24 December 1964. I do. I was reading and re-reading the
description of a chemistry set on page 248 of the Sears Christmas
Wish Book. Here it is.
In the preface of Illustrated Guide to Home Chemistry Experiments,
I said that I'd asked for the biggest, most expensive
Lionel-Porter-Chemcraft chemistry set in the Sears Wish Book. I was
wrong. I asked for the second most expensive one.
I don't
remember why I didn't ask for the $50 set. Maybe I figured that my
parents would decide that it was too advanced for me, at age 11. Or
maybe the price scared me. After all, $50 was a lot of money in 1964.
Heck, $30 was a lot of money.
I checked the government CPI page,
which informed me that $30 in 1964 was the equivalent of about $204
today. I don't believe that. Prices have increased more than seven
times in the last 44 years. In terms of real purchasing power, I'd
guess that $30 chemistry set cost my parents the equivalent of at least
$300 in today's money.
But I really, really wanted that
chemistry set and my parents decided to get it for me, for which I will
always be thankful. Their decision paid off, too, and they weren't
alone. Hundreds of thousands of other kids also got chemistry sets that
year, and for a significant number of them that first chemistry
set was the key that opened a door to a new world. Getting that
first chemistry set ultimately led thousands upon thousands of those
kids to choose careers in science.
I don't know if my parents or
those hundreds of thousands of other parents set out consciously
to encourage a new generation of scientists, but that was
unquestionably the effect. My parent's generation nurtured a new crop
of scientists, which blossomed in my generation.
Alas, their
children and their children's children haven't done as well, and I
blame most of that on the demise of the real chemistry set. According
to the best figures I could find, only about a third as many college
freshmen major in chemistry as did so when I entered college in 1971.
The US population in 1971 was about 208,000,000, and is now over
300,000,000. When you consider that a higher percentage of high-school
students now attend college, that means an entering freshman is now
perhaps one fifth as likely to major in chemistry as was his
counterpart in 1971.
And
it's not just chemists. Most physicists
didn't get started in science with physics sets, nor biologists with
biology sets. Nearly all of them got started with chemistry sets,
and the demise
of chemistry sets has inevitably led to fewer people choosing
chemistry or physics or
biology as careers. We're running out of scientists and doing
nothing to
replace them. Without scientists, civilization collapses. We need to
encourage our best and brightest kids to become scientists, and the way
to do that is to grab their interest when they're young by giving them
chemistry sets.
00:00
-
Copyright
© 1998,
1999, 2000, 2001, 2002, 2003, 2004, 2005, 2006, 2007, 2008 by Robert
Bruce
Thompson. All
Rights Reserved.
