|
|
 |
|
Reader
Bill Lindley sent me a pointer to an intriguing article in the Scientific
American Web site about a research project that aims to rank Web sites
by their importance, according to the number of times that other pages link
to them. This may be one way to get around the spamdexers, who use various
means to subvert the traditional robot-based search engines. It's tough
to create a multitude of sites that point to another site, and it's easy
to detect such subversion. (Fifty pointers in one HTML document pointing
to the same place…hmmmm.) I've got to ponder this one a little, but in the
meantime, go give it a read and see what you think: http://www.sciam.com/1999/0699issue/0699raghavan.html
|
|
|
47 today. No regrets, though there are days when I miss my hair—and other days when I don't. I ponder my position in history at times, and realize that I may well bracket the most incredible period of advancement in technology that the human race will ever see. When I was in grade school, transistors were novel and TVs were mostly vacuum tubes. Kids played with Erector Sets and made kites out of newspaper. My obsession with the king of technologies, electronics, led me through vacuum tech, transistor tech, IC tech, and now computer tech. I made my own radios, my own digital frequency counters (with individual 7400 family ICs), and eventually my own computers, with a wire-wrap gun and a lot of patience. Kids coming of age today are born with mouse in hand. They do well in computers, but will they ever truly understand gear trains, like I did, courtesy my Erector/Meccano construction set? Will they ever understand how kites fly, and what makes airplane wings provide lift? I can hope. But truthfully, any child who really wants to know these things has tools to help him or her learn that I never even dreamed of, starting, of course, with the Web. As for me, well, in the wake of 1969's Apollo mission to the Moon, our nerd's lunch table at Lane Tech High in Chicago decided it would hold a reunion on the Moon in 2000. After all, the Luna Hilton would be completed by 1985, no? No. I have some faint hope of achieving orbit as a tourist before I die—and some brighter hope of owning an "air car" like so many SF stories have assumed would be commonplace by now. (See VDM Diary for May 24, 1999.) On the other hand, I have done a lot of interesting things with these two hands, learned to allow myself to be amazed, and (perhaps most amazingly, for an ex-nerd who lost his father young) I met and married a beautiful woman, who has been my best friend now for almost thirty years. Like I said, no regrets. None! |
|
|
One
of my readers called my attention to KDevelop, a graphical IDE for C++ programmers
that is both hosted by and targeted to the K Desktop Environment under Linux.
KDevelop is conceptually similar to C++Builder or Visual C++, and consists
of an IDE and several graphical tools, some of which stand on their own,
and others that act as front ends for the standard GNU tools like GCC and
GDB (the GNU debugger.) There's a class browser, an application wizard,
an icon editor, a form-oriented dialog editor, a documentation browser,
a help editor, and lots more. Not sure how I missed it while I was researching
KDE, but there you have it. As with KDE, it's being developed in Germany,
and while the Web site cautions that "only unstable alpha versions are available"
it's actually very slick and professional looking. I have not run it yet,
but the reader I mentioned has agreed to write an article for VDM, and my
hope is that you'll read all about it in the November/December issue. In
the meantime, go to the KDevelop Web site: http://www.cs.uni-potsdam.de/~smeier/kdevelop/
Now what we need is something similar for GNOME—something that allows
the use of FreePascal or GNU Pascal if possible.
|
|
|
Speedchoice
is going 2-way wireless! (See VDM Diary for March 20, 1999.) That means
I'll be able to free up the phone line for the uplink and be connected to
the net 24 X 7. This is more important to me than the uplink speed (though
it is occasionally important, as when I'm throwing manuscript files around)
which in the current system is typically 26.4 Kbps. The two-way service
is another $10 per month over what I pay now, and comparable to what I pay
for the cable modem service in my Chicago satellite office. I'm scheduled
for install in early August (!!) and I'll report here after it's up and
running. It'll be interesting to see whether the new agreement explicitly
prohibits mounting servers. I have a server mounted on my Chicago machine,
but it's for my own use (I experiment with Web content there) and not for
public consumption.
|
|
|
Apple
just doesn't learn. They're now suing Daewoo for making a PC that looks
too much like an IMac. Apple is famous for using the courts to fend off
competition, and it's interesting to note that every time they've actually
gotten to trial on something like this, they've lost bigtime. Their suit
against Microsoft was a horrible waste of their money and energy, and I
believe it was the main reason they ceased to be a significant part of the
market back a few years ago. They usually manage to force smaller companies
to stop doing what they don't like. Microsoft fought back, and won. Daewoo
is not a bit player in the PC world. If they fight, Apple will spend precious
money and energy again, and will emerge from the fight just a little bit
smaller. How often can they do this?
|
|
|
The
new Sony Vaio superslim laptops haven't attracted the attention they deserve.
I have an odd and very rare little book published in 1978 by Xerox Palo
Alto Research Center, which contains a photo of a mockup of what PARC's
research scientists considered the "ideal" DynaBook. It was a clamshell-format
computer like all our current laptops, with the keyboard on one half and
the display on the other. It was less than an inch thick. PARC had no idea
in 1978 how such a thing might be created, but they were sure that that
was what we needed, and they set out to create software that would eventually
run on that species of electronic notebook. It was PARC, not Apple, that
created all the concepts that underlie both the Mac and Windows. I know;
I was there; I saw it happen—though I was not part of PARC itself but
in another Xerox unit. We used PARC workstations and read PARC reports.
When I saw that little notebook mockup, I remember thinking, Someday
we'll all have computers like that. Someday's here. It took 20 years,
but hey—the world has changed. Let us not forget who thought of it
all first.
|
|
|
I bought an 18.2 GB hard drive for my main system at home, having in mind to replace the existing 6.4 GB drive that came with the Compaq Deskpro 6300. The vendor assured me that the new drive was compatible with the old drive, and that they could both live on the same SCSI cable. Electrically true...but physically impossible. I discovered that the new drive was about 1/8" too tall to fit in the teeny-weeny little slot Compaq allows for hard drives in the great big honking heavy sheet metal all-steel 6300. The only way to get it in there was to remove the Jaz drive from one of the exposed bays and mount the 18.2 GB drive in its place. This I did, and it works fine—and now I have a box with 24.8 GB of hard drive space. That oughta keep me awhile, I suppose. But it's a real puzzle. The 6300 is huge—and most of its internal space is simply wasted. There is easily room for two more full-size drives inside, but the internal components are arranged so that all that extra space is simply wasted. About all I can figure is that Compaq is deliberately keeping their big DeskPro machines from being used as servers, so that they can't compete with the more expensive Compaq ProLiant server line. The "power-down off" main power switch is one way to do this (see VDM Diary for April 11, 1999) and arbitrarily limiting hard drive space is another. This is almost certainly the last Compaq I will ever own. Next one is a Dell. |
|
|
Every
so often, a computing challenge turns up for me for which there is no solution
short of programming. None! Here's the latest: I want to draw a spiral and
save it as a graphic so that I can print it to laser transfer film and apply
it to a printed circuit board. Spirals can be used as "printed inductors"
in electronic circuits, and I want to experiment with these. But I have
had no luck in searching for a drawing program that will allow me to create
parameterized spirals. I intuit that Visio is capable of handling the task,
but I haven't been able to figure out how. So I guess it's back to Delphi,
look up equations for drawing spirals, and create a special utility. Every
time I think our apps are without bounds in their power, something slaps
me upside the head to remind me that there is no substitute for being able
to program.
|
|
|
Travelling
for a couple of days. Back on the 14th.
|
|
|
Work
on my Aardmarks bookmark manager continues apace. I'll explain the concept
in detail in an article I'm working on that should appear in the September/October
issue of VDM. I intend to distribute the source to the utility itself,
but it can't be "open source" because it uses commercial software components
from three different companies. This is an interesting problem, and it'll
be enlightening to see if people want to work on Aardmarks badly enough
to buy the libraries. My guess is they won't…but we'll see. There will be
some new free components to come of the project, however. More on those
later on, as things firm up.
|
|
|
People
keep pushing Python at me. I don't get it. We have Perl—why do we need
Python? I must be missing something.
|
|
|
This
has nothing to do with computing, but it's cool nonetheless: There's a company
producing a wind-up flashlight and wind-up radio, both containing their
own little generators. I stopped in at a Discovery Channel retail store
down at the mall the other night, and played with both of them. You can
get them in bright yellow plastic cases, or, most marvelously, in transparent
plastic cases that let you watch what the innards are doing. A crank on
the side of both units winds up a wide stainless steel spring, which then
turns a small generator at uniform speed. There are lots of plastic gears
that turn. The radio plays for an hour on a single wind-tight of the spring.
I didn't write down the name of the units, but I think it was FreePlay or
something like that.
|
|
|
I've
got a new product here called WinDriver, and it looks promising. The product
helps developers create Windows 9x and NT drivers for specific hardware
peripherals. It's got a wizard that gathers data and then generates a skeletal
driver for the peripheral, which seems pretty routine—except that this
wizard goes out and looks directly at the hardware and tries to detect interrupt
numbers, memory map ranges, and such. The documentation doesn't suggest
this, but my experiments imply that you could possibly write a driver for
a piece of hardware that you didn't develop. This is mighty tempting—I
have more than one gadget that won't talk to NT4 because the vendors were
too lazy to implement NT drivers. It's cleanly done and very impressive,
and I suggest you take a look if you play the drivers game. (The brochure
indicates that a Linux version is in the works…and that could definitely
make a difference!) www.krftech.com.
|
|
|
Once again, somebody has sent me snotty email about how the Mac was a more user-friendly system than Wintel. Predictably, he's never used a PC. That's a common thread with such letters, which I get on a spotty basis, mostly because of things I wrote years ago. There was a time when the Mac was easier to use than Windows, but that time has passed. My best example of Windows' superiority is the context menu that appears when you right-click on something. The Mac has nothing like it—can't have, since the Mac mouse has only one button. I admit, three buttons may be more than most people can coordinate, though I learned the mouse on a three-button system at Xerox and have always preferred three buttons. But with two buttons you may have the ideal compromise: One button for selection, and another button for context. It works beautifully, especially once people get out of the "Dummies" category and become experts. Mac people don't seem to understand this, and they can't, because their hardware has no way to do it. I've watched over the shoulders of some of our Mac people here, and they take a lot more mousing around to get things done than PC people do. Apple has always believed their own PR a little too much. Sooner or later it's going to kill them. |
|
|
There is a GNU Pascal. But in the wild, wild world of Linux, which depends utterly on the GNU toolset, it's almost completely invisible. It's almost like it's a conspiracy. C and C++ people seem to fear Pascal above all else. Deny it if you will. I know it's true. GNU Pascal can be found here: http://agnes.dida.physik.uni-essen.de/~gnu-pascal/ And, of course, there's FreePascal 32, which is more mature and more like Borland Pascal 7, which might as well be the world standard: http://gd.tuwien.ac.at/languages/pascal/fpc/www/ |
|
|
The
real danger in open-source software development may be that development
stops before the product is really done. The remorseless competition in
the Windows software arena has delivered us some extremely polished and
innovative software—with perhaps more bugs than we'd like. Open source
products, by contrast, tend to be simpler, less polished, but more robust.
RPM is great—but it needs to go the last 20% of the way toward being
a really turnkey install system. It's good enough for the nerds, I suppose,
but it's not yet in the same league as Installshield. After spending a month
or two in the Linux world, I have to say, it's mostly bug-free, but more
than a little spartan. Maybe this is good. Creeping featurism is certainly
a problem in the Windows world, particularly with mega-apps like Office—but
I, for one, would like to see more attention paid to the niceties of user
interfaces. I've been away from the command line for a long time—and
that's really the way I like it.
|
|
June 30, 1999: