Anything you need to quanitfy can be measured in some way that is superior to not measuring it at all.
But, how do you know that the way you're measuring it is better than not measuring at all? There are lots of ways to measure things that are worse than no measurement at all, because they reward the wrong activity.
The canonical examples here are paying programmers per bug fixed, or paying testers per bug detected. Either one of these alone is bad - together they allow programmers and testers to print money for themselves.
In theory, nothing is unmeasurable. In practice, some things are so hard to measure that you might as well not even try.
You seem to be saying "eventually the network will be fast enough that the speed advantage of portable media won't matter." The problem is that "fast enough" is a moving target - the size of the data objects we want to move does track with Moore's Law.
Example: in 1987, my house was connected to the Internet via a Telebit 19.2 KB modem. Today, in 2004, it's a 3MB/256KB cable link.
Seventeen years is roughly eleven eighteen-month periods, so if network speeds had kept up with Moore's Law, I would have seen a 2048-fold speedup. Instead, I got a 156-fold speedup (or only 13-fold if you look at upload speed).
Portable physical media will probably always be faster than fetching data via a network.
But I often wonder why Linux vendors and Apple don't directly attack the numerous security shortfalls of Microsoft products.
As godawful as all the recent security problems with MS have been, they really haven't been bad enough to get either business users or the general public to switch. They've both been conditioned by years of exposure to Windows to just expect this kind of annoyng behavior from computers.
And that's all the recent exploits have been, annoyances. At worst, you get to scrape your disk off and reinstall, with very little harm done - a little worse than a BSOD, but not costly.
Now, if a worm got loose with properties like I suggested here, people would switch in droves.
And, how much will they pay? Between discounts to keep first worlders from switching, and discounts to get third worlders to pay anything at all, their realized income per copy has nowhere to go but down, fast.
Granted, he didn't get much of anywhere at Xerox, either - defining "anywhere" as "broad acceptance and commercialization".
Both Xerox and HP are schizophrenic organizations. In the 80's Xerox was bipolar, oscillating between "we put marks on paper/toner pays our salaries" and "we are the information company", taking about three years to complete a cycle. HP used to be an "it's all about the engineering" company - now after ther merger with Compaq, their strategy seems to be "good enough is good enough to be the last man standing".
Part of the reason MS thinks of itself as an underdog is their inability to really innovate. They've never been first in any software category - they're good enough to be the last man standing, but that requires competence and persistence, not innovation.
Their marketing and sales force has the general public convinced they're brilliant innovators, but among their technical peers, they're behind the curve. We know it, they know it, and it gives them an inferiority complex a mile wide.
Does this mean that Word 5.1 was the last release of Word/Mac where MS was actually competing with other Mac word processors?
From 6.0 upward, was Word/Mac's selling point by-ghod compatibility with Word/Windows, which was judged more important than competing on merit (performance or features) in the Mac market?
Lisp community, OTOH, seems to think that only commercial implementations are worthwhile.
CMUCL and SBCL and OpenMCL and CLISP aren't exactly chopped liver. Steel Bank Common Lisp might be of particular interest to you, since it runs on most modern architectures and pretty much anything Unix-ish, and a Windows port is limping (see here for details). Or maybe CLISP - it compiles to byte codes, like Python, so it runs everywhere, at about Python speed.
Twist it any way you want, one standard open source implementation would guarantee better library and tool support.
You're probably right, but the other languages you mention that have one standard implementation also have a benevolent dictator/gatekeeper that "owns" that implementation, and by extension the standard for the language. They can evolve faster than Common Lisp or Scheme, which are defined by published specifications that aren't owned by one person, but that one person becomes a single point of failure for the language. None of them has yet had to survive the loss of their gatekeeper, and it's not clear to me that they can do so. All of them have problems that might have been avoided with a slower, spec-based evolution (rant - all of them do variable scoping wrong).
Being defined by a reference implementaton also holds languages back - the language is defined by the reference implementaton, warts and all. Take Python speed - CPython is slow compared to Lisps that compile to native code. Python compilers help, but as long as CPython is the reference, everything else (including compilers) will be non-"standard", and Python the language will still be thought of as slow. The Lisp community can't coalesce around one portable implementation like CLISP, because they'd have to give up the high-performance high ground.
Lisp turns 46 this year. It's behind the curve in recent protocols and libraries, but it's still here, and will likely still be around 46 years from now.
... from Road Tools. Same price range. Durable as hell. Fits in your carrying case, and takes up less room than a foldable LapPad. Targus rebrands them and sells them in Staples, or you can order them online through the website above.
Guaranteed forever. Our family has two iBooks, each with a small CoolPad. My kids have managed to pop off some of the rubber feet and unscrew the rivet-and-screw that holds the two swiveling parts together. Sent email to the manufacturer, and they're sending me replacement parts, no charge. I'm definitely going to put a little thread lock on that screw...
My preferred laptop working stack is:
lap/laptop case/CoolPad/iBook
Raises the machine up to closer to eye level, the CoolPad lets air circulate under the iBook, and the whole rig is in my lap, so I can put it away and be ready to roll in 10 seconds or so.
I think that applying for a patent when prior art is known about should be a criminal offense and punishable by a court of law.
Companies defend themselves against that by saying "we didn't know, we didn't look, and you can't prove otherwise." Bizarrely enough, this works in court.
Right now, the USPTO has incentives to grant patents, because their budget comes mostly from fees. There's no penalty for them when they grant a bogus patent, because bogosity gets determined in the courts, down the line.
What's needed is a feedback mechanism - say, patent examiners get annual reviews on both their efficiency (patents reviewed/year) and their accuracy (patents upheld/invalidated).
Back in the days when dinosaurs ruled the machine room, big customers with big problems and big accounts could get operating system and compiler source code.
The biggest reason Unix became the OS of choice in universities in the 1980s was the availability of source, which made it possible to port it to new platforms.
Linux today is all about hardware flexibility - "Linux on my wristwatch" is still a little funny, but "Linux on my pocket PC/XBox/random-hardware" is not a joke.
The parent post has it right - the bleeding edge usually is on non-commodity hardware, so it will always tend toward I-can-port-it-myself OS solutions.
So the full price would be taxed by the relevant jurisdiction's sales tax.
I can live with that - that's no worse than paying sales tax on a shrinkwrap software package. At least it keeps software from being subject to special taxes.
The software is free, and copyrighted, and useless without a dongle.
The dongle costs $$$. The only "custom" software in it is the authentication key, and if they're going to tax that, they'll have to tax RFID chips too.
... for those of us who are tech support for blind friends. I would get my blind friend onto an OS X Mac in a heartbeat if there were a usable screen reader, just to get him away from the Windows virus and worm wars.
If Apple does this right, it will enable a lot more independence in mainstream blind computer users - having screen reader support in the OS should allow blind users to do a lot more of their own system maintenance than is possible now
spread fast for the first few hours or days, until it saturated the vulnerable population, then cut way back on network traffic and hide.
not crash machines or trash all their files - instead, it would slowly and subtly modify user data files (see here for a few suggestions).
Imagine what would happen to modern business if they discovered that they couldn't trust any document that had ever touched a Windows machine... the world's economy would grind to a halt. Not even Microsoft has enough money to pay damages for an event like that, though the combined law firms of the world would try to get it from them.
Slashdot Mirror
But, how do you know that the way you're measuring it is better than not measuring at all? There are lots of ways to measure things that are worse than no measurement at all, because they reward the wrong activity.
The canonical examples here are paying programmers per bug fixed, or paying testers per bug detected. Either one of these alone is bad - together they allow programmers and testers to print money for themselves.
In theory, nothing is unmeasurable. In practice, some things are so hard to measure that you might as well not even try.
I think you meant to say the multi-billion dollar cell-phone ring tone market.
I didn't believe it either...
Well, maybe not right, but he called the stock buyback over a week ago. Maybe Gates is going to take his ball and go home.
You seem to be saying "eventually the network will be fast enough that the speed advantage of portable media won't matter." The problem is that "fast enough" is a moving target - the size of the data objects we want to move does track with Moore's Law.
Your own numbers above show this:
1987 Floppy 1.2MB 25 minutes
2004 CD-ROM 660MB 2-3 hours
The size of an interesting collection of data has gone up by a factor of 500, but network speeds have only gone up 20-200 times.
Example: in 1987, my house was connected to the Internet via a Telebit 19.2 KB modem. Today, in 2004, it's a 3MB/256KB cable link.
Seventeen years is roughly eleven eighteen-month periods, so if network speeds had kept up with Moore's Law, I would have seen a 2048-fold speedup. Instead, I got a 156-fold speedup (or only 13-fold if you look at upload speed).
Portable physical media will probably always be faster than fetching data via a network.
As godawful as all the recent security problems with MS have been, they really haven't been bad enough to get either business users or the general public to switch. They've both been conditioned by years of exposure to Windows to just expect this kind of annoyng behavior from computers.
And that's all the recent exploits have been, annoyances. At worst, you get to scrape your disk off and reinstall, with very little harm done - a little worse than a BSOD, but not costly.
Now, if a worm got loose with properties like I suggested here, people would switch in droves.
And, how much will they pay? Between discounts to keep first worlders from switching, and discounts to get third worlders to pay anything at all, their realized income per copy has nowhere to go but down, fast.
Granted, he didn't get much of anywhere at Xerox, either - defining "anywhere" as "broad acceptance and commercialization".
Both Xerox and HP are schizophrenic organizations. In the 80's Xerox was bipolar, oscillating between "we put marks on paper/toner pays our salaries" and "we are the information company", taking about three years to complete a cycle. HP used to be an "it's all about the engineering" company - now after ther merger with Compaq, their strategy seems to be "good enough is good enough to be the last man standing".
Sad, really...
Weird. Granted, Apple did a lot of weird, relatively unproductive stuff in the mid-80s...
Part of the reason MS thinks of itself as an underdog is their inability to really innovate. They've never been first in any software category - they're good enough to be the last man standing, but that requires competence and persistence, not innovation.
Their marketing and sales force has the general public convinced they're brilliant innovators, but among their technical peers, they're behind the curve. We know it, they know it, and it gives them an inferiority complex a mile wide.
Does this mean that Word 5.1 was the last release of Word/Mac where MS was actually competing with other Mac word processors?
From 6.0 upward, was Word/Mac's selling point by-ghod compatibility with Word/Windows, which was judged more important than competing on merit (performance or features) in the Mac market?
CMUCL and SBCL and OpenMCL and CLISP aren't exactly chopped liver. Steel Bank Common Lisp might be of particular interest to you, since it runs on most modern architectures and pretty much anything Unix-ish, and a Windows port is limping (see here for details). Or maybe CLISP - it compiles to byte codes, like Python, so it runs everywhere, at about Python speed.
You're probably right, but the other languages you mention that have one standard implementation also have a benevolent dictator/gatekeeper that "owns" that implementation, and by extension the standard for the language. They can evolve faster than Common Lisp or Scheme, which are defined by published specifications that aren't owned by one person, but that one person becomes a single point of failure for the language. None of them has yet had to survive the loss of their gatekeeper, and it's not clear to me that they can do so. All of them have problems that might have been avoided with a slower, spec-based evolution (rant - all of them do variable scoping wrong).
Being defined by a reference implementaton also holds languages back - the language is defined by the reference implementaton, warts and all. Take Python speed - CPython is slow compared to Lisps that compile to native code. Python compilers help, but as long as CPython is the reference, everything else (including compilers) will be non-"standard", and Python the language will still be thought of as slow. The Lisp community can't coalesce around one portable implementation like CLISP, because they'd have to give up the high-performance high ground.
Lisp turns 46 this year. It's behind the curve in recent protocols and libraries, but it's still here, and will likely still be around 46 years from now.
It's been tried - see Dylan. As near as I can tell, Dylan didn't take off because:
The Lisp people saw no major advantages to it other than the syntax, and they'd already gotten past that barrier
The non-Lisp people apparently didn't understand that it really was better than C++/Java
Like the one growing here?
Now I'm really confused. Multiple implementations with a common specification hurt Lisp, but Jython and stackless python are not drawbacks for Python?
... the most sincere form of self-criticism.
You have to physically carry the data from point A to point B
You have to hand the data to the recipient, so both of you have to be space-time coincident
This will just add another step in the old one-upmanship communication chain:
"I need a copy of that."
"Can I fax it to you?"
"Can you email it to me?
"Do you have a web site where I can drop it?"
"Here, just drop it on my PDA"
Feh.
... from Road Tools. Same price range. Durable as hell. Fits in your carrying case, and takes up less room than a foldable LapPad. Targus rebrands them and sells them in Staples, or you can order them online through the website above.
Guaranteed forever. Our family has two iBooks, each with a small CoolPad. My kids have managed to pop off some of the rubber feet and unscrew the rivet-and-screw that holds the two swiveling parts together. Sent email to the manufacturer, and they're sending me replacement parts, no charge. I'm definitely going to put a little thread lock on that screw...
My preferred laptop working stack is:
lap/laptop case/CoolPad/iBook
Raises the machine up to closer to eye level, the CoolPad lets air circulate under the iBook, and the whole rig is in my lap, so I can put it away and be ready to roll in 10 seconds or so.
Companies defend themselves against that by saying "we didn't know, we didn't look, and you can't prove otherwise." Bizarrely enough, this works in court.
Right now, the USPTO has incentives to grant patents, because their budget comes mostly from fees. There's no penalty for them when they grant a bogus patent, because bogosity gets determined in the courts, down the line.
What's needed is a feedback mechanism - say, patent examiners get annual reviews on both their efficiency (patents reviewed/year) and their accuracy (patents upheld/invalidated).
Thank you, no. What if the other driver has had a really bad day, and wants to take a few other people with him over the cliff?
Back in the days when dinosaurs ruled the machine room, big customers with big problems and big accounts could get operating system and compiler source code.
The biggest reason Unix became the OS of choice in universities in the 1980s was the availability of source, which made it possible to port it to new platforms.
Linux today is all about hardware flexibility - "Linux on my wristwatch" is still a little funny, but "Linux on my pocket PC/XBox/random-hardware" is not a joke.
The parent post has it right - the bleeding edge usually is on non-commodity hardware, so it will always tend toward I-can-port-it-myself OS solutions.
... or maybe "satirical tribute".
... don't get the implant over your kidney or you might lose two things at once!
I can live with that - that's no worse than paying sales tax on a shrinkwrap software package. At least it keeps software from being subject to special taxes.
The software is free, and copyrighted, and useless without a dongle.
The dongle costs $$$. The only "custom" software in it is the authentication key, and if they're going to tax that, they'll have to tax RFID chips too.
Any other problems?
... for those of us who are tech support for blind friends. I would get my blind friend onto an OS X Mac in a heartbeat if there were a usable screen reader, just to get him away from the Windows virus and worm wars.
If Apple does this right, it will enable a lot more independence in mainstream blind computer users - having screen reader support in the OS should allow blind users to do a lot more of their own system maintenance than is possible now
spread fast for the first few hours or days, until it saturated the vulnerable population, then cut way back on network traffic and hide.
not crash machines or trash all their files - instead, it would slowly and subtly modify user data files (see here for a few suggestions).
Imagine what would happen to modern business if they discovered that they couldn't trust any document that had ever touched a Windows machine... the world's economy would grind to a halt. Not even Microsoft has enough money to pay damages for an event like that, though the combined law firms of the world would try to get it from them.