"Consequently the GPL implies that every binary bought from Sveasoft must come with a free subscription to the source code."
No. It must come with access to the source code that generated the particular binary. "subscription" is a continuous service including updates. GPL gives you no rights to updates.
"Sveasoft are not free to cancel this subscription if the source is redistributed."
Sveasoft can cancel any subscription service as long as the terms of the subscription say that they can. The subscription service is nothing to do with the GPL.
If car manufacturers were obliged to provide workshop manuals upon request, that wouldn't mean they were obliged to replace your car each time they released a new model. You're, like many others in this thread, conflating two unrelated things.
"I would never even think of installing anything in [...]/usr/local (because it "belongs" to Apple)"
Do what?/usr/local is mine, all mine, as _I'm_ the sysadmin on this machine. If OSX touches/usr/local, then Apple are a pretty crap bunch of bodgers. Can anyone verify?
In which case his argument is little more than "you can't do the chinese character set in 8-bits" which is barely worthy of mention, let alone a whole article.
I program (PHP) for a small, but international, company which sells GSM gimicks (ringtones/logos/games/services). Presently we sell content to Germans, Danes, Norwegians, Swedes, Estonians, and other countries whose character sets are a subset of the above (Finland, UK, etc).
And we "blithely" use 8-bits to represent the characters, no doubt much to Joel's chagrin. And everything works. Perfectly. If Joel is unable to get things to work without using UTF, then maybe, just maybe, he should stop pretending that he's a competant coder. Most of our coders are newbs, but are all from places with wacky character sets, and do not make the same assumptions that UK or US programmers would. It's not PHP that's at fault, it's false assumptions that are the problem.
Joel, as always, is full of false assumptions. And arrogant, but that's almost his middle name.
Oh, and someone needs to tell him that "This platonic A is different than B" makes him sound like an ignorant hick. Different _from_, not different _than_.
For reference, I don't like PHP, but I criticise it for its real faults (of which there are many) not its imagined ones.
But issues of copyright infringement only concern whether he produced it from memory or not. It's _not_ from scratch if you produce it from memory. He'd seen it before, he wasn't the copyright owner, he reproduced it, he infringed copyright. It's black and white.
Do what?!?! You're confusing hardware with software. My references are sound:
""" Clearly three fragmented efforts could not be continued. There was also a realisation that a serious corporate effort would be needed. In April 1985, Dave Cutler was asked to lead a cross-functional team to define a corporate RISC architecture and develop implementation plans for hardware and software. The architecture team consisting of Dave Cutler, Dileep Bhandarkar (VAX Architecture Group), Wayne Cardoza (VAX/VMS Operating System Group), Dave Orbits (High Preformance Systems Group), and Rich Witek (Semiconductor Engineering Group) published the first draft of the PRISM (Parallel Reduced Instruction Set Machine) architecture specification in August 1985.
PRISM was initially defined as a 64-bit RISC architecure.... """
i.e. PRISM was the _hardware_, not the software.
Who am I quoting there? Dileep Bhandarkar. You may have heard of him, perhaps?
Alpha Implementations and Architecure, Complete Reference and Guide. Dileep Bhandarkar. Digital Press, 1995. ISBN 1-55558-130-7
I think you need to add "Mica" to your google searches to find the several references to Mica being the _software_ platform that MS ripped off DEC via Cutler.
If you're lucky you might encounter things like the following: """ Cutler had been developing a follow-on to VMS at DEC called Mica, and when DEC dropped the project he brought the expertise and some engineers with him to Microsoft. DEC also believed he brought Mica's code to Microsoft and sued. """
I thought Prism was the processor line that never came to market, not the software. The DEC code name for the Alpha processor family was "AXP", and one joke was that AXP stood for "almost exactly prism", IIRC.
When I worked in a mobile phone conslutancy, we wanted to remain flexible, and targeted both Hitachi and ARM processors. To do this, we wrote everything apart from a tiny abstracted kernel in portable C. All the DSP and layer 1 stuff took place on ASICS, obviously. That's reflected in other places where I've worked too - assembly just isn't required for 99% of tasks nowadays.
One is a reserved region of screen where representations of running programs can be found, the other is a reserved region of screen where representations of running programs can be found.
And given that the "taskbar" can be hidden, is there any real difference (I'll get to the "yes" answer later) between it and my Sawfish popup menu which not only contains representations of running programs, but, pertinant to this patent, groups all like programs together in order to save space at the expense of having to do a bit more mouse-work. Very little, apart from the fact that my popup can appear anywhere, and MS's only resides along one side of the screen.
The MS, Apple, KDE etc. weenies and patent lawyers who get off on pretending that there's anything novel about what's little more than making a selection by mouse movement and clicking need a serious dose of reality.
Is there anyone I've not included? Hmmm,actually the old Amiga and Atari/ST guys have got off scott free, I don't remember them (us, I was one, not saying which side, as it was a friendly rivalry) getting so self-indulgent and obsessive.
Hexagonally oriented auto-hiding recusive popup icon menus, with gestures of course, that's where the future of GUI development lies, I's sure.
Note that dynamic programming isn't even what real hardcore computer scientists would call it anyway.
Those with a Lisp or ML background (i.e. the real hardcore beardie-weirdie computer scientists) would simply "thunk" the results. I.e. they'd put the results in a theoretically infinite list implemented using thunks. (Note, these are nothing to do with MS's Win/Win32 cludges)
Fibonacci is a canonical undergrad Lisp thunk exercise.
"a Turing machine, which has no concept of a 'stack'"
Total nonsense.
A turing machine has two stacks: - the tape to the left of the current position - the tape to the right of the current position
""" just compile any recursive code in any language to MIPS or some other RISC instruction set without push/pop and call functions like those found on X86 chips. You don't "call" functions directly, you just move around in memory and jump all over the place, """
Total nonsense.
Why has my Alpha have JSR (jump to subroutine) and RET (return from subroutine) instructions?
Why does MIPS (_your_ example), PPC/POWER and TI's C80MP have JAL (jump and link) or BL/BLR (branch and link, branch to link register) instructions? It's just a call stack with extra overhead if you want to go more than 1 deep. (and less overhead if you only go 1 deep. It's therefore only useful for leaf-nodes, at which point function inlining is often a better option.)
You're of course right, but do remember that IBM have had 5000+ pins on their chips since last year. (~half are power.) And yes, these systems do not come cheap. However, AMD would not be breaking any new ground by heading this way.
Slashdot Mirror
"Consequently the GPL implies that every binary bought from Sveasoft must come with a free subscription to the source code."
No. It must come with access to the source code that generated the particular binary.
"subscription" is a continuous service including updates. GPL gives you no rights to updates.
"Sveasoft are not free to cancel this subscription if the source is redistributed."
Sveasoft can cancel any subscription service as long as the terms of the subscription say that they can. The subscription service is nothing to do with the GPL.
If car manufacturers were obliged to provide workshop manuals upon request, that wouldn't mean they were obliged to replace your car each time they released a new model. You're, like many others in this thread, conflating two unrelated things.
FP.
I'm a user of various OSes.
/usr/bin/gzip /usr/local/bin/zip
/usr/bin/telnet /usr/local/bin/ssh
/usr/bin/ftp /usr/local/bin/sftp
/usr/bin/talk /usr/local/bin/irc
asdf ~/public_html uname
FreeBSD
asdf ~/public_html which gzip
asdf ~/public_html which zip
asdf ~/public_html which telnet
asdf ~/public_html which ssh
asdf ~/public_html which ftp
asdf ~/public_html which sftp
asdf ~/public_html which talk
asdf ~/public_html which irc
If that's in any way logical, then I'm a Dutchman.
If talk is part of the base OS, then so is internet explorer.
FP.
"I would never even think of installing anything in [...] /usr/local (because it "belongs" to Apple)"
/usr/local is mine, all mine, as _I'm_ the sysadmin on this machine. If OSX touches /usr/local, then Apple are a pretty crap bunch of bodgers. Can anyone verify?
Do what?
FP.
In which case his argument is little more than "you can't do the chinese character set in 8-bits" which is barely worthy of mention, let alone a whole article.
FP.
Mitä?
http://fi2.php.net/manual/fi/function.eval.php
So are you claiming there's no PHP interpreter in the compiled PHP?
If so, then you do realise that this means that the authors of the compiler have solved the halting problem?
FP.
I program (PHP) for a small, but international, company which sells GSM gimicks (ringtones/logos/games/services). Presently we sell content to Germans, Danes, Norwegians, Swedes, Estonians, and other countries whose character sets are a subset of the above (Finland, UK, etc).
And we "blithely" use 8-bits to represent the characters, no doubt much to Joel's chagrin. And everything works. Perfectly. If Joel is unable to get things to work without using UTF, then maybe, just maybe, he should stop pretending that he's a competant coder. Most of our coders are newbs, but are all from places with wacky character sets, and do not make the same assumptions that UK or US programmers would. It's not PHP that's at fault, it's false assumptions that are the problem.
Joel, as always, is full of false assumptions.
And arrogant, but that's almost his middle name.
Oh, and someone needs to tell him that "This platonic A is different than B" makes him sound like an ignorant hick. Different _from_, not different _than_.
For reference, I don't like PHP, but I criticise it for its real faults (of which there are many) not its imagined ones.
FP.
Sorry for being so abrupt.
I get a bit 'defensive' when challenged on Alpha issues.
Phil
But issues of copyright infringement only concern whether he produced it from memory or not. It's _not_ from scratch if you produce it from memory.
He'd seen it before, he wasn't the copyright owner, he reproduced it, he infringed copyright. It's black and white.
Do what?!?! You're confusing hardware with software. My references are sound:
....
"""
Clearly three fragmented efforts could not be continued. There was also a realisation that a serious corporate effort would be needed. In April 1985, Dave Cutler was asked to lead a cross-functional team to define a corporate
RISC architecture and develop implementation plans for hardware and software. The architecture team consisting of Dave Cutler, Dileep Bhandarkar (VAX Architecture Group), Wayne Cardoza (VAX/VMS Operating System Group), Dave Orbits (High Preformance Systems Group), and Rich Witek (Semiconductor Engineering Group) published the first draft of the PRISM (Parallel Reduced Instruction Set Machine) architecture specification in August 1985.
PRISM was initially defined as a 64-bit RISC architecure
"""
i.e. PRISM was the _hardware_, not the software.
Who am I quoting there?
Dileep Bhandarkar. You may have heard of him, perhaps?
Alpha Implementations and Architecure, Complete Reference and Guide.
Dileep Bhandarkar.
Digital Press, 1995.
ISBN 1-55558-130-7
I think you need to add "Mica" to your google searches to find the several references to Mica being the _software_ platform that MS ripped off DEC via Cutler.
If you're lucky you might encounter things like the following:
"""
Cutler had been developing a follow-on to VMS at DEC called Mica, and when DEC dropped the project he brought the expertise and some engineers with him to Microsoft. DEC also believed he brought Mica's code to Microsoft and sued.
"""
FP.
You're confusing
"from memory"
with
"from scratch"
If he did it in the past, then it's the former, not the latter.
FP.
"Prism"? Are you sure it wasn't "Mica"?
I thought Prism was the processor line that never came to market, not the software. The DEC code name for the Alpha processor family was "AXP", and one joke was that AXP stood for "almost exactly prism", IIRC.
FP.
When I worked in a mobile phone conslutancy, we wanted to remain flexible, and targeted both Hitachi and ARM processors. To do this, we wrote everything apart from a tiny abstracted kernel in portable C.
All the DSP and layer 1 stuff took place on ASICS, obviously. That's reflected in other places where I've worked too - assembly just isn't required for 99% of tasks nowadays.
FP.
Really, is there any difference?
One is a reserved region of screen where representations of running programs can be found, the other is a reserved region of screen where representations of running programs can be found.
And given that the "taskbar" can be hidden, is there any real difference (I'll get to the "yes" answer later) between it and my Sawfish popup menu which not only contains representations of running programs, but, pertinant to this patent, groups all like programs together in order to save space at the expense of having to do a bit more mouse-work. Very little, apart from the fact that my popup can appear anywhere, and MS's only resides along one side of the screen.
The MS, Apple, KDE etc. weenies and patent lawyers who get off on pretending that there's anything novel about what's little more than making a selection by mouse movement and clicking need a serious dose of reality.
Is there anyone I've not included? Hmmm,actually the old Amiga and Atari/ST guys have got off scott free, I don't remember them (us, I was one, not saying which side, as it was a friendly rivalry)
getting so self-indulgent and obsessive.
Hexagonally oriented auto-hiding recusive popup icon menus, with gestures of course, that's where the future of GUI development lies, I's sure.
Either that or it's the DWIW button.
FatPhil
I'm sure I'm not the only one who's learnt that:
/var/www/acmqueue.com/htdocs/db/db.php on line 88
Fatal error: Call to undefined function: message_die() in
is indicative of bad programming. Thanks ACM Queue for an enlightening 2-line article!
FP.
There's a difference between
- C++ doesn't have bounds checking, that's in the standard
and
- C++ doesn't have bounds checking in the standard.
You meant the latter, you wrote the former.
FP.
It would always require one bounds check, as how else would it know that array A is of size >= 10.
FP.
Or it means he can't write C++ worth a damn.
FP.
HHL!
Note that dynamic programming isn't even what real hardcore computer scientists would call it anyway.
Those with a Lisp or ML background (i.e. the real hardcore beardie-weirdie computer scientists) would simply "thunk" the results. I.e. they'd put the results in a theoretically infinite list implemented using thunks. (Note, these are nothing to do with MS's Win/Win32 cludges)
Fibonacci is a canonical undergrad Lisp thunk exercise.
FP.
"a Turing machine, which has no concept of a 'stack'"
Total nonsense.
A turing machine has two stacks:
- the tape to the left of the current position
- the tape to the right of the current position
"""
just compile any recursive code in any language to MIPS or some other RISC instruction set without push/pop and call functions like those found on X86 chips. You don't "call" functions directly, you just move around in memory and jump all over the place,
"""
Total nonsense.
Why has my Alpha have JSR (jump to subroutine) and RET (return from subroutine) instructions?
Why does MIPS (_your_ example), PPC/POWER and TI's C80MP have JAL (jump and link) or BL/BLR (branch and link, branch to link register) instructions? It's just a call stack with extra overhead if you want to go more than 1 deep. (and less overhead if you only go 1 deep. It's therefore only useful for leaf-nodes, at which point function inlining is often a better option.)
FP.
You're of course right, but do remember that IBM have had 5000+ pins on their chips since last year.
(~half are power.) And yes, these systems do not come cheap. However, AMD would not be breaking any new ground by heading this way.
FP.
Don't be a fool!
Don't you realise that means he's running _5_ web servers off this thing. Surely that's 5 times as impressive!
FP.
It invites the question. It begs the question.
/begging the question/, but that doesn't mean that those three words can't be used in that order.
It may not be
Stop being so blinkered^Wproscriptive.
FP.
Yes, my girlfriend of nearly 6 years.
(Hmmm, and my previous girlfriend, who's been with her current partner for 8 years now without sprogging.)
FP.
Some would say that intolerance is an after-effect of fear, and fear is an after-effect of ignorance.
FP.
Comander Taco, Please BAN this EXTREEMIST LOONY NOW!
FP.