When I have a choice, Ada 95, Eiffel, Modula 3,
Sather, Scheme, or Squeak. Sometimes my clients want C, but they explicitly do NOT want C++.
and why?
Mostly for software reliability and programmer productivity.
If you use the best language for the job methodology,
Which I do.
why have you not coded in Qt?
AFAIK, there are no Qt bindings for the languages I use, though someone claims to be developing Ada 95 bindings. The
list of languages with Qt bindings is amazingly short and lopsided,
There are GTK bindings for every language I use, and they work just fine. From my point of view, Qt brings nothing useful to the party. I'm not going to switch to an inferior programming language in order to use a toolkit that only offers very minor benefits (if any).
Code in GTK for a week, then code in Qt for a week. Tell me which one we should throw out then.
I've coded with GTK for a lot longer than a week. I would have been happy to code with Qt for at least a week, but it doesn't have bindings for any of the languages I use. This rather heavily influences my opinion on which one "we should throw out".
I'm not fundamentally opposed to taxing interstate commerce, but why the f*#@ should interstate sales over the internet be taxed in any different way than interstate telephone sales, interstate mail order sales, etc.? The internet is only acting as a communication medium, and does not fundamentally change the nature of an interstate sale: an order is communicated from a buyer to a seller in another state, and goods are shipped from the seller to the buyer.
The hardware is theoretically capable of it.
But even though I'm a Linux advocate, I'm not convinced that there would be much point in running Linux on it. If that's what you want, get a Zaurus or something.
IIRC the 10C/11C/12C/15C/16C were the 'Pioneer' series
No, those are the "Voyager" series. The Pioneer series started with the HP 27S, and included the 10B, 17B, 20S, 21S, 22S, 27S, 32S, 42S, and variants of those.
My second favourite series was "Woodstock" - 21/25/25C/29C
The best industrial design of a calculator ever, with the HP-41C a close second.
Wrong. The 28 and 48 were developed by a team in Corvallis, Oregon, led by Bill Wickes. Around 1994 the director of Corvallis division was told that he had to transfer engineering for either calculators or notebooks to Singapore division so that the latter would do more than just manufacturing.
He thought Corvallis would be better off with laptops, so he sent calculators to Singapore.
As it turns out, Corvallis Division was not successful with notebooks, so now they only design and manufacture inkjet cartridges. (I was fortunate enough to get a tour of the inkjet operation a few years ago, and it's absolutely amazing.)
Singapore division did one calculator product as a collaboration with the engineers in Oregon, the HP 38G, and that was the end of new calculator development for a while.
Around 1997 or so, a new calculator division formed in Australia, and hired as engineers some of the people in France who had written good third-party extensions to the 48 software, such as Metakernel. ACO did NOT employ any of the original HP 48 engineers. ACO took the 48GX source code, added in Metakernel and other enhancements, designed a new case (with a not-so-reliable keyboard), and called it the 49G.
ACO desiged various other products, such as the 10BII business calculator, a redesign of the 10B.
They also made OEM deals with other companies to produce the low-end HP 6S and HP 6S Solar and a few other models.
Some of the most interesting products designed by ACO were cancelled before introduction. The Xpander was a Windows CE based calculator with a geometry application. Preproduction units were made and given to educators, who apparently really liked it. The reason for the cancellation is unclear. Casio has now announced a somewhat similar product, the Classpad.
Another notable product ACO almost launched (right before they were shut down) was the Jornada X25. This was a Linux and Java based PDA with a B&W display, StrongARM processor, and MP3 capabilities which would have been very inexpensive (around $199 retail) and targetted to high school students. They did one preproduction run and seeded developers (I was one). There was a roadmap for an even less expensive model.
Now a new calculator division has been formed in San Diego. They are introducing a new 49G+ which is ARM-based and has many improvements over the 49G. The keyboard is much better, though there is still room for improvement.
But I thought that HP committed the travesty of killing of its wonderful calculator line before Fiona took charge.
She was definitely in charge when HP shut down ACO, the Austrailian Calculator Operation.
HP's new calculator division is based in San Diego. I was just at the HHC conference this
past weekend, and the division director,
marketing director, engineering manager, and
lead software engineer were there. They seemed
to be very concerned about
customer feedback and fixing any problems (or
perceived problems) with product quality.
Their first new high-end RPN calculator in four
years is the HP 49G+, which will be officially announced in the US on October 6. (It's already
available in some parts of the world.)
The HP 49G+ has similar capabilities to the HP 49G, but with a larger display (25% more pixels), faster processor (75 MHz ARM), more memory, better keyboard, USB interface, and an SD slot for flash memory expansion.
Once an hour may not be sufficient for the customer. But once every five minutes or even once every minute would probably be reasonable, and should not be an ureasonable load on a hosting provider.
if you RTFA you'll see that it isn't actually a fusion reactor or reaction at all.
If you RTFA you'll see that it emits four neutrons per minute above the background level. If you're claiming that those four neutrons aren't the result of fusion, pray tell where they are in fact coming from?
While it's obviously possible to write reliable, secure software in C, the language doesn't offer any help to a programmer wanting to do so, and is in many ways a hindrance. And although C++ is
better in some ways, it actually has most of the drawbacks of C since it is basically a superset.
The first step to solving
the OS security problem, IMNSHO, is to build the OS in a real high-level language, instead of a portable assembly language. (Who was it that said that C combines the power and flexibility of assembly language with the ease of use and maintainability of assembly language?)
There are many languages that are much better suited for software engineering. A partial list might contain Ada, Eiffel, Java, Modula 3, Oberon, Sather, Scheme, Self, and Smalltalk. I'm sure there are plenty of other reasonable languages.
There may be some portions of the OS that need to bypass normal type and bounds checking, but these should be few and far between. Most of the languages I've listed above provide facilities for this, but they require you to explicity declare that you're doing it, rather than making it the default behavior as in C and C++.
In copyright law, ownership cannot be transferred without express, written authority of a copyright holder.
Wow, Darl finally gets something right.
Some have claimed that, because SCO software code was present in software distributed under the GPL, SCO has forfeited its rights to this code. Not so - SCO never gave permission, or granted rights, for this to happen.
Irrelevant. The GPL does not transfer ownership. It grants a license. No one has claimed that SCO has given away Unix code or transferred ownership. But by redistributing Linux under the GPL terms, SCO has granted a license to any code in Linux which they may own.
In the unlikely event that the GPL is found to be invalid, as Darl has claimed, SCO will be in even worse shape than they are now. The GPL is the only thing that has allowed SCO to distribute the Linux kernel at all; if the GPL is invalid, SCO is guilty of tens or hundreds of thousands of counts of copyright violation themselves. At up to $150K damages per count, they could have to pay damages comparable to what they've asked from IBM.
I beg to differ. Xerox had several things that were "true" GUIs before Apple started down that path. In fact, it's well documented that Apple decided to develop a GUI inteface after being given a demonstration at Xerox PARC (in exchange for Apple stock). Apple licensed Smalltalk-80 (which was GUI-based) and had it running on 68000-based hardware well before the Lisa's native OS and GUI. Apple's port of Smalltalk-80 was never released as a product, though a version ported to the Macintosh OS was released to developers through APDA.
Whenever I buy a new notebook, the first thing I do is to remove the hard drive and set it aside. Then I put in a new Toshiba or HGST (formerly IBM) 9.5mm drive and install Linux.
To hell with the EULA of the vendor's software, I refuse to use it.
Most recently (a little over a year ago) I did this with a Fujitsu Lifebook P2040.
You didn't read the article did you?
The article states why base 3 is the most efficient.
Yes, I read the article, and it provides no actual
support for the conjecture that base 3 will be more
efficient in practice for useful circuits. It only claims that it is now possible to make simple ternary gates that aren't as inefficent when implemented with heterojunction devices as they were with mainstream semiconductor technology.
Sure, you may be able to make an efficient gate of some sort. But that's not the same thing as being able to build an efficient adder, multiplier, instruction decoder, or other practical structures needed
to build microprocessors and other useful chips.
Base 3 or 4 logic is NOT smaller than base 2.
on
Beyond Binary Computing?
·
· Score: 5, Informative
A move to multi-valued logic provides more computational capability without the standard increase in die size or transistor count.
No, it doesn't. Let's see you design a 16-quat
full adder that takes fewer transistors or less
die area than an 32-bit full adder.
Base 3 or higher are a lose for implementing logic. Base 4 is useful in some kinds of memory, and this has been done by Intel since around 1980-81. Intel used a quaternary ROM (two bits per cell) for the microcode store of the 43203 Interface
Processor, and (IIRC) for the 8087. More recently this technique has been used in flash memory.
If you need to run a backup MX for a lot of
domains, you don't have to configure them all
manually. You can just tell Postfix that it's
allowed to backup domains that have MXes on
specific networks. For instance, my Postfix
main.cf includes:
(specific addresses changed to protect the innocent, and yes, I know that a byte can't exceed 255, that was deliberate)
This tells Postfix to accept mail for any domain
that has an MX in one of the specified networks.
So whenever I add a new domain to one of my
primary MX servers, I don't have to change the
configuration on my backup MX servers at all.
I had an aftermarket lithium ion pack for my
Nokia phone. The pack had a clear housing rather
than opaque, but I bought it only due to it being
less expensive than the Nokia pack.
One day I met a friend for lunch, and he said
that he'd tried to call me but reached my voice
mail. I pulled the phone out of my pocket and
discovered it was dead. A glance at the battery
showed that there had apparently been a short
circuit in the pack that burned insulation on
wiring and vaporized a small piece of the wiring.
It was close to one of the actual lithium-ion
celss, so I am very lucky that the cell did not
catch fire in my pocket.
I don't know for a fact that the official
Nokia packs are any better engineered than the
aftermarket once. But I was very disturbed that
the pack did not contain a fuse in series with
the cells to avoid this type of problem.
Burning lithium cells are very dangerous.
I wish I'd taken photos of the burned pack, but
I didn't yet have a digital camera at the time,
and it didn't seem important enough to justify
buying a disposable camera.
but it also has some very stupid ideas (genetically lucky people)
Wow, thanks for clarifying that for me. I knew that the idea was very unusual, but I'm not smart enough to be able to completely write if off as stupid. It always seemed to me that since we don't really know with complete certainty that the collapse of quantum wavefunctions can never be externally influenced, that unexpectedly biased outcomes to "random" occurances might be possible.
After all, that's what it would take for a person to have "luck". Apparently Greg Egan's novel "Quarantine" must also have some "very stupid ideas".
all tied together with a poorly-written excuse of a story. The plot is dry,
<sarcasm>
Yes, there's just nothing interesting in the novel at all.
</sarcasm>
far too predictable
<sarcasm>
Right. It was immediately obvious from the first page exactly what would happen.
</sarcasm>
, and unoriginal.
<sarcasm>
I know exactly what you mean. There were just so many
other novels that covered exactly
this same concept and plot before.
</sarcasm>
Of course, I've never been able to actually find or identify those novels.
C'mon, it's fine to say that you didn't like the novel, but your criticism of it is extremely weak.
Maybe you could cite some examples to support your assertions of "completely predicable" and "unoriginal".
What a great business model, I think I'll use it too...
If you're running MAME, you owe me $32. Pay up! MAME includes some code I wrote, in violation of the GPL license on my code. Unlike SCO, I'm actually willing to publicly identify which lines of code are at issue.
I'm joking about the $32, although they really did
violate my license. However, I'm NOT going to sue them. In fact, I think I'll grant the MAME project a license to use the code under the MAME license instead.
So much for my chances of making billions of dollars on it!:-)
Slashdot Mirror
However, even if Cingular's network was taken down by a virus, that still demonstrates incompetence on their part.
I'm not fundamentally opposed to taxing interstate commerce, but why the f*#@ should interstate sales over the internet be taxed in any different way than interstate telephone sales, interstate mail order sales, etc.? The internet is only acting as a communication medium, and does not fundamentally change the nature of an interstate sale: an order is communicated from a buyer to a seller in another state, and goods are shipped from the seller to the buyer.
The hardware is theoretically capable of it. But even though I'm a Linux advocate, I'm not convinced that there would be much point in running Linux on it. If that's what you want, get a Zaurus or something.
Singapore division did one calculator product as a collaboration with the engineers in Oregon, the HP 38G, and that was the end of new calculator development for a while.
Around 1997 or so, a new calculator division formed in Australia, and hired as engineers some of the people in France who had written good third-party extensions to the 48 software, such as Metakernel. ACO did NOT employ any of the original HP 48 engineers. ACO took the 48GX source code, added in Metakernel and other enhancements, designed a new case (with a not-so-reliable keyboard), and called it the 49G.
ACO desiged various other products, such as the 10BII business calculator, a redesign of the 10B. They also made OEM deals with other companies to produce the low-end HP 6S and HP 6S Solar and a few other models.
Some of the most interesting products designed by ACO were cancelled before introduction. The Xpander was a Windows CE based calculator with a geometry application. Preproduction units were made and given to educators, who apparently really liked it. The reason for the cancellation is unclear. Casio has now announced a somewhat similar product, the Classpad.
Another notable product ACO almost launched (right before they were shut down) was the Jornada X25. This was a Linux and Java based PDA with a B&W display, StrongARM processor, and MP3 capabilities which would have been very inexpensive (around $199 retail) and targetted to high school students. They did one preproduction run and seeded developers (I was one). There was a roadmap for an even less expensive model.
Now a new calculator division has been formed in San Diego. They are introducing a new 49G+ which is ARM-based and has many improvements over the 49G. The keyboard is much better, though there is still room for improvement.
HP's new calculator division is based in San Diego. I was just at the HHC conference this past weekend, and the division director, marketing director, engineering manager, and lead software engineer were there. They seemed to be very concerned about customer feedback and fixing any problems (or perceived problems) with product quality.
Their first new high-end RPN calculator in four years is the HP 49G+, which will be officially announced in the US on October 6. (It's already available in some parts of the world.)
The HP 49G+ has similar capabilities to the HP 49G, but with a larger display (25% more pixels), faster processor (75 MHz ARM), more memory, better keyboard, USB interface, and an SD slot for flash memory expansion.
Once an hour may not be sufficient for the customer. But once every five minutes or even once every minute would probably be reasonable, and should not be an ureasonable load on a hosting provider.
The first step to solving the OS security problem, IMNSHO, is to build the OS in a real high-level language, instead of a portable assembly language. (Who was it that said that C combines the power and flexibility of assembly language with the ease of use and maintainability of assembly language?)
There are many languages that are much better suited for software engineering. A partial list might contain Ada, Eiffel, Java, Modula 3, Oberon, Sather, Scheme, Self, and Smalltalk. I'm sure there are plenty of other reasonable languages.
There may be some portions of the OS that need to bypass normal type and bounds checking, but these should be few and far between. Most of the languages I've listed above provide facilities for this, but they require you to explicity declare that you're doing it, rather than making it the default behavior as in C and C++.
In the unlikely event that the GPL is found to be invalid, as Darl has claimed, SCO will be in even worse shape than they are now. The GPL is the only thing that has allowed SCO to distribute the Linux kernel at all; if the GPL is invalid, SCO is guilty of tens or hundreds of thousands of counts of copyright violation themselves. At up to $150K damages per count, they could have to pay damages comparable to what they've asked from IBM.
To hell with the EULA of the vendor's software, I refuse to use it.
Most recently (a little over a year ago) I did this with a Fujitsu Lifebook P2040.
Sure, you may be able to make an efficient gate of some sort. But that's not the same thing as being able to build an efficient adder, multiplier, instruction decoder, or other practical structures needed to build microprocessors and other useful chips.
Base 3 or higher are a lose for implementing logic. Base 4 is useful in some kinds of memory, and this has been done by Intel since around 1980-81. Intel used a quaternary ROM (two bits per cell) for the microcode store of the 43203 Interface Processor, and (IIRC) for the 8087. More recently this technique has been used in flash memory.
smtpd_recipient_restrictions = permit_mynetworks, permit_mx_backup, reject
permit_mx_backup_networks = 64.15.260.112/27, 282.66.92.0/22, 67.91.305.33/32
(specific addresses changed to protect the innocent, and yes, I know that a byte can't exceed 255, that was deliberate)
This tells Postfix to accept mail for any domain that has an MX in one of the specified networks. So whenever I add a new domain to one of my primary MX servers, I don't have to change the configuration on my backup MX servers at all.
If she fails, the vendor, and possibly the election officials, will cite this as "proof" that the system is secure.
One day I met a friend for lunch, and he said that he'd tried to call me but reached my voice mail. I pulled the phone out of my pocket and discovered it was dead. A glance at the battery showed that there had apparently been a short circuit in the pack that burned insulation on wiring and vaporized a small piece of the wiring. It was close to one of the actual lithium-ion celss, so I am very lucky that the cell did not catch fire in my pocket.
I don't know for a fact that the official Nokia packs are any better engineered than the aftermarket once. But I was very disturbed that the pack did not contain a fuse in series with the cells to avoid this type of problem. Burning lithium cells are very dangerous.
I wish I'd taken photos of the burned pack, but I didn't yet have a digital camera at the time, and it didn't seem important enough to justify buying a disposable camera.
C'mon, it's fine to say that you didn't like the novel, but your criticism of it is extremely weak. Maybe you could cite some examples to support your assertions of "completely predicable" and "unoriginal".
Well, I'm one of those people that expected "the Simpsons in the future", and IMNSO that's exactly what it's been, and I've been happy with it.
If you're running MAME, you owe me $32. Pay up! MAME includes some code I wrote, in violation of the GPL license on my code. Unlike SCO, I'm actually willing to publicly identify which lines of code are at issue.
I'm joking about the $32, although they really did violate my license. However, I'm NOT going to sue them. In fact, I think I'll grant the MAME project a license to use the code under the MAME license instead.
So much for my chances of making billions of dollars on it! :-)