| Author |
Message |
bachus
Joined: Feb 29, 2004
Posts: 2922
Location: Up in that tree over there.
Audio files: 5
|
 Posted: Mon Oct 29, 2007 6:09 am Post subject:
|
  |
|
| Acoustic Interloper wrote: | His other two points are bullshit. There is no intrinsic correlation of multiple threads to disk IO, and the only villian he really has for paging is a bad Windoze paging algorithm.
|
| Quote: | | "The problem is that when you add multithreading to the mix, you have multiple threads submitting disk requests, which then causes the disk to thrash all over the place. This is especially fun when you boot your computer and all of the garbage that every program has placed in the startup folder decides to load at the same time. It's gotten worse with faster CPUs, because now programmers think they can speed things up by doing tasks in the "background," only to increase the amount of contention on the one poor disk, which ends up lowering overall throughput. (*cough*Intellisense*cough*)" |
I thought he is describing an extrinsic problem not an intrinsic one. But I have not the expertise to argue with him. Would you mind explaining the problem with his argument for those of us without enough depth to get it off the top?
Thanks.
_________________
The question is not whether they can talk or reason, but whether they can suffer. -- Jeremy Bentham |
|
|
Back to top
|
|
 |
Acoustic Interloper
Joined: Jul 07, 2007
Posts: 2070
Location: Berks County, PA
Audio files: 89
|
| Posted: Mon Oct 29, 2007 6:09 am Post subject:
|
|
|
| dewdrop_world wrote: | As far as software-as-art, SuperCollider must be in the running. McCartney has this uncanny ability to make everybody else's code look unbearably clumsy.
James |
Yes, having an identifiable architect (or few) is always a necessary, if not sufficient, precondition for having a good architecture. There are software engineers who really want to make their creations sing! Unfortunately, their priorties often conflict with the priorities of software producing companies.
_________________
When the stream is deep
my wild little dog frolics,
when shallow, she drinks. |
|
|
Back to top
|
|
|
Acoustic Interloper
Joined: Jul 07, 2007
Posts: 2070
Location: Berks County, PA
Audio files: 89
|
| Posted: Mon Oct 29, 2007 7:39 am Post subject:
|
|
|
| bachus wrote: | | Acoustic Interloper wrote: | His other two points are bullshit. There is no intrinsic correlation of multiple threads to disk IO, and the only villian he really has for paging is a bad Windoze paging algorithm.
|
| Quote: | | "The problem is that when you add multithreading to the mix, you have multiple threads submitting disk requests, which then causes the disk to thrash all over the place. This is especially fun when you boot your computer and all of the garbage that every program has placed in the startup folder decides to load at the same time. It's gotten worse with faster CPUs, because now programmers think they can speed things up by doing tasks in the "background," only to increase the amount of contention on the one poor disk, which ends up lowering overall throughput. (*cough*Intellisense*cough*)" |
I thought he is describing an extrinsic problem not an intrinsic one. But I have not the expertise to argue with him. Would you mind explaining the problem with his argument for those of us without enough depth to get it off the top?
Thanks. |
I have to keep going back and rereading, because the author is mixing up several different concepts into a jumble. Can't keep the 'blog' straight in my head.
| Quote: | | While multithreading on the CPU is good, though, it isn't so good for the hard disk. In fact, it makes hard disk performance suck, because seeks are expensive and even a moderate amount of seeking causes throughput to plummet. The problem is that when you add multithreading to the mix, you have multiple threads submitting disk requests, which then causes the disk to thrash all over the place. This is especially fun when you boot your computer and all of the garbage that every program has placed in the startup folder decides to load at the same time. It's gotten worse with faster CPUs, because now programmers think they can speed things up by doing tasks in the "background," |
Typically, multithreading refers to breaking up a single threaded application (1 process) into multiple threads of execution. Believe it or not, it is possible to *simplify* program structure using multithreading. I have had to write a TCP/IP-based server (nowadays a 'web server') using both 16 bit Windows (1994, before multithreading came into Windows), and a later upgrade using 32 bit Windows which has multithreading as well as multiple application processes. The former requires any module that needs to wait for *any* IO (disk or network) to interact with every other module, both IO intensive and CPU intensive modules, within the Windows so-called 'event loop.' It's a terrible software structure that breaks modularity, because every semi-independent module has to chain itself into the OS' 'event loop.' OSX has duplicated this nonsense. It's basically a poor man's thread scheduler, but since the Win16 OS did not support threads, it forced the application programmer to emulate them, with every application task putting code into the event loop handler. (The tasks were semi-independent because the Windows machine was a DSP hardware cards server that served DSP hardware access to users on different Unix & windows machines running DSP software tools; they used the cards independently, but had interactions with respect to PC CPU & memory management and other application-neutral support infrastructure.
When doing a new system using Win32 that supports multithreading, each semi-indendent application task could now run in its own thread. When the need to block for disk IO or network IO arose, it was no longer necessary to clutter up application logic with lower level scheduling code that belongs in the OS. Doing this did not increase the amount of disk IO, or network IO, which was driven entirely by the application. Of course there is now *synchronization overhead* -- two semi-independent threads periodically need to synchronize when accessing shared data -- but in a well structured multithreaded program synchronization occurs infrequently. It was occuring in the single threaded win16 program, anyway, it's just that the app logic had to share in the overhead.
The other potential advantage to correctly structured multithreaded code is that processor with hardware multithreading, such as both the Alienware and MacBook Pro machines I am using, can assign hardware threads to software threads, accelerating the application. The acceleration does not scale linearly with hardware threads, because 1) there is always some (sequential) synchronization overhead among threads, 2) large scale mutlithreading takes more careful, typically profiling-based division of threaded modules, and 3) hardware threads still share some subset of CPU & other resources, e.g., connections to memory.
So, issue one with the blog is that the problem is indeed extrinisic and not intrinsic. So it should be titled "Shitty Multithreading' and not "Multithreading," or maybe "Multithreading Gone Wild."
But, in fact, the blog is not referring to splitting an application process (modular or otherwise) into multiple threads, it is referring to running multiple application processes at the same time, something that's been done since the time sharing systems of the 1960's. UNIX is a time sharing system, and the 'taska <filea> filec' command pipeline is really a way to do multithreading without thread support in the OS; the shell sets up the synchronization, and as long as process scheduling overhead is not severe, you get the same effect *if the software architecture can be cast as a pipeline.* That is not the case for the DSP hardware server, for example.
So, what's really wrong with time sharing? A) SHITTY SCHEDULERS, and B) SHITTY APPLICATIONS. Extrinsic, indeed! My first multi-process application on win32 was a DSP debugger toolset that has an IO intensive GUI running in one process and a CPU intensive DSP simulator running in another, communicating *very infrequently* (from the simulator's perspective), only at user interaction time, via a socket. Windows NT manifest problem (A) by giving the CPU intensive simulator processor consistently higher execution priority than the IO intensive GUI process; what it should have done is, over time, lower the execution priority of the CPU hogging simulation, so that a click to the "Breakpoint" button on the GUI would get immediate priority to stop the simulation. Instead, NT's scheduler, by reversing scheduling priorities that had appeared in textbooks for decades, made it impossible to stop the simulation from within the GUI. This eventually got fixed in NT.
In fact, what the blog really seems to be pointing at is running a bunch of B) SHITTY APPLICATIONS, especially from the get-go at startup time. Yeah, there are lots of those. I used to run Norton Anti-virus and SpySweeper on my Alienware machine. Startup time had become intolerable, so I started watching process times and memory consumption at startup time, and SpySweeper appeared to be a big culprit. In its defense (I suppose), I don't know how much of the problem may have been caused by interaction between the two. Remember the old "Spy vs. Spy" cartoons in Mad Magazine? (Maybe they still have this?) Anyway, Norton is adding more anti-spy stuff, so why have 2? Just before SpySweeper expired, I threw it the hell off the machine, also did a disk defrag, and it is just *amazing* how much faster this machine comes up.
None of this has anything intrinsically to do with multithreading or multiprocess computing, or paging or garbage collection for that matter. It's all extrinsic, stemming from a shitty, careless, cheap attitude about software architecture within software producing companies. Ed Yourdon even codified it with the name "Good Enough Programming" in his 1996 book *Rise and Resurrection of the American programmer* (YES, JESUS!), the sequel to his earlier, other day-late-and-a-dime-short saga, *Decline and Fall of the American Programmer*. (The Internet and Web arose between the two, and somehow astute Ed translated that into, 'Make Money and Salvage Careers by Writing Whatever Shit the Market will Buy.' What that has to do with engineering in a specific locale is beyond me. Seems pretty global, increasingly sweatshop piecework programming turning out shit.) I guess he's working on a third book, *Decline and Fall of the Resurrected Programmers Who Gave a Shit*. Catchy, no 
There is something to be said for avoiding the perfectionist trap of never getting anything out due to the 'polishing my turd' syndrome. I've had to catalyze a few of these in my time. But perfectionism in commodity software is not, I dare say, much of a problem, and 'good enough' has fallen pretty low.
_________________
When the stream is deep
my wild little dog frolics,
when shallow, she drinks. |
|
|
Back to top
|
|
|
bachus
Joined: Feb 29, 2004
Posts: 2922
Location: Up in that tree over there.
Audio files: 5
|
| Posted: Mon Oct 29, 2007 7:43 am Post subject:
|
|
|
Hey, thanks A.I. much appreciated!
_________________
The question is not whether they can talk or reason, but whether they can suffer. -- Jeremy Bentham |
|
|
Back to top
|
|
|
Kassen
Janitor
Joined: Jul 06, 2004
Posts: 7678
Location: The Hague, NL
G2 patch files: 3
|
| Posted: Mon Oct 29, 2007 10:44 am Post subject:
|
|
|
| dewdrop_world wrote: |
As far as software-as-art, SuperCollider must be in the running. McCartney has this uncanny ability to make everybody else's code look unbearably clumsy.
|
It is. It's also one of the few works in it's field that prides itself on efficiency. In my own opinion all the TOPLAP systems ( http://toplap.org/forums.html/ToplapSystems ) are art to some degree and in some way but it's a small niche and "artfullness" drops off steeply as soon as you get out of it. MAX/MSP is a great system but I wouldn't call it "art", for example.
_________________
Kassen |
|
|
Back to top
|
|
|
Kassen
Janitor
Joined: Jul 06, 2004
Posts: 7678
Location: The Hague, NL
G2 patch files: 3
|
| Posted: Mon Oct 29, 2007 11:02 am Post subject:
|
|
|
| Acoustic Interloper wrote: |
I have to keep going back and rereading, because the author is mixing up several different concepts into a jumble. Can't keep the 'blog' straight in my head. |
Thanks! I felt exactly the same way. I'm not nearly as deeply into this stuff as you are but felt I should at least be able to read that blog but...
| Quote: | | There is something to be said for avoiding the perfectionist trap of never getting anything out due to the 'polishing my turd' syndrome. I've had to catalyze a few of these in my time. But perfectionism in commodity software is not, I dare say, much of a problem, and 'good enough' has fallen pretty low. |
Yeah. I wonder how much of that is to blame on modern CPU's being so blazing fast while 95% of their users have very modest demands (editing text, browsing the web, watching a few videos, etc). I can well understand why so many people in the "software-art" scene are interested in programming for mobile phones now.
The prime market for multi-core multi-cpu setups at times appears to be to make sure cute flash intros will run at all (sorry, pet peeve, but what on earth is going on there that makes something barely above a Amiga demo bring down a Pentium 4 or above?)
_________________
Kassen |
|
|
Back to top
|
|
|
Acoustic Interloper
Joined: Jul 07, 2007
Posts: 2070
Location: Berks County, PA
Audio files: 89
|
| Posted: Mon Oct 29, 2007 3:08 pm Post subject:
|
|
|
| dewdrop_world wrote: | | Acoustic Interloper wrote: | | Hire the ones that are young and naive and easy to shit up, and set them to work on piecework. Keep it cheap, just like walmart and prefab housing developments and the rest of it, if not cheap to buy, at least cheap to produce. Follow the law of averages, chopping off the heads that are too low *and* too high. |
That sounds exactly like the company I work for.
James |
Yeah, I just got my head chopped off again by a would-be employer a few minutes ago, and I can only deduce the reason. They don't provide any honest, accurate information from which I could guess better. The nice thing about heads, though, is that I can always grow out another.
_________________
When the stream is deep
my wild little dog frolics,
when shallow, she drinks. |
|
|
Back to top
|
|
|
brinxmat
Joined: Oct 24, 2005
Posts: 262
Location: Norway
|
| Posted: Tue Oct 30, 2007 1:07 am Post subject:
|
|
|
| http://www.apple.com/macosx/developertools/xcode.html wrote: | | Objective-C 2.0 is so compelling, Apple wrote Xcode 3.0 itself using it. Enjoy modern garbage collection, syntax enhancements, runtime performance improvements, and 64-bit support. All at your own pace, since it’s compatible with existing Objective-C source. Write applications more quickly with fewer bugs using Objective-C 2.0 in Xcode 3.0. |
Objective-C didn't previously have automatic garbage collection, I always assumed that there was a good reason for this, though I have found none stated explicitly, I came upon the following by myself: if you're clever enough to create a working program, you're probably clever enough to manage memory as well.
Obviously, I'm not a trained programmer, and my opinion probably counts for zip, but I would suspect that automatic garbage collection would:
a) improve the learnability of a programming language for people who just need to use it
b) contain better, more efficient -- and not least functional -- routines for garbage collection than say your average programmer could supply
c) provide more rapid development, allowing the optimization efforts to go rather to the eternal black hole of -- wait for it -- the display!
d) not be entirely significant for those who objected to its implementation because they presumably have a working knowledge of C, and as Objective-C is a superset of C, they can use C-style memory management.
However, everything would be much faster if we just did it in assembly.
_________________
-- Say "&Eth;onne hit wæs hrenig weðer" |
|
|
Back to top
|
|
|
bachus
Joined: Feb 29, 2004
Posts: 2922
Location: Up in that tree over there.
Audio files: 5
|
| Posted: Tue Oct 30, 2007 6:05 am Post subject:
|
|
|
| brinxmat wrote: |
However, everything would be much faster if we just did it in assembly. |
Yup we only need two languages, Assembler and Common Lisp, or for wusses (like me) C and Common Lisp.
_________________
The question is not whether they can talk or reason, but whether they can suffer. -- Jeremy Bentham |
|
|
Back to top
|
|
|
cappy2112
Joined: Dec 24, 2004
Posts: 2471
Location: San Jose, California
Audio files: 2
G2 patch files: 1
|
| Posted: Tue Oct 30, 2007 8:43 am Post subject:
|
|
|
| brinxmat wrote: |
However, everything would be much faster if we just did it in assembly. |
Sure- when the speed is necessary.
Can you imagine writing the NM1 or G2 editors in assembly?
After having worked with firmware engineers for the last 20-some years, less than 3 want to work with assembly, and are happy to move to a new code base which is in a higher level language. Of course, there is a tiny amount of firmware that must be done is assembly, but it's less than 1 % of the overall project. |
|
|
Back to top
|
|
|
BobTheDog
Joined: Feb 28, 2005
Posts: 4044
Location: England
Audio files: 32
G2 patch files: 15
|
| Posted: Tue Oct 30, 2007 8:49 am Post subject:
|
|
|
| cappy2112 wrote: | | brinxmat wrote: |
However, everything would be much faster if we just did it in assembly. |
Sure- when the speed is necessary.
Can you imagine writing the NM1 or G2 editors in assembly?
After having worked with firmware engineers for the last 20-some years, less than 3 want to work with assembly, and are happy to move to a new code base which is in a higher level language. Of course, there is a tiny amount of firmware that must be done is assembly, but it's less than 1 % of the overall project. |
There is also the argument that decent modern compilers (like the Intel c++ compiler) can generate superior machine code than most humans nowadays. Getting the most out of a processor is a lot harder nowadays as they are pretty complex beasts even though the instructions havn't changed too much over time. |
|
|
Back to top
|
|
|
Antimon
Joined: Jan 18, 2005
Posts: 4145
Location: Sweden
Audio files: 371
G2 patch files: 100
|
| Posted: Tue Oct 30, 2007 11:21 am Post subject:
|
|
|
| BobTheDog wrote: | | cappy2112 wrote: | | brinxmat wrote: |
However, everything would be much faster if we just did it in assembly. |
Sure- when the speed is necessary.
Can you imagine writing the NM1 or G2 editors in assembly?
After having worked with firmware engineers for the last 20-some years, less than 3 want to work with assembly, and are happy to move to a new code base which is in a higher level language. Of course, there is a tiny amount of firmware that must be done is assembly, but it's less than 1 % of the overall project. |
There is also the argument that decent modern compilers (like the Intel c++ compiler) can generate superior machine code than most humans nowadays. Getting the most out of a processor is a lot harder nowadays as they are pretty complex beasts even though the instructions havn't changed too much over time. |
Even more impressive is that, if you turn the optimisation level high enough, many compilers can create faulty code a million times faster than humans.
/Stefan
_________________
Antimon's Window
@soundcloud @Flattr home - you can't explain music |
|
|
Back to top
|
|
|
cappy2112
Joined: Dec 24, 2004
Posts: 2471
Location: San Jose, California
Audio files: 2
G2 patch files: 1
|
| Posted: Tue Oct 30, 2007 11:34 am Post subject:
|
|
|
| Antimon wrote: |
Even more impressive is that, if you turn the optimisation level high enough, many compilers can create faulty code a million times faster than humans.
/Stefan |
That's the job-security level of optimization.
(Only to be used when all other optimizations fail) |
|
|
Back to top
|
|
|
blue hell
Site Admin
Joined: Apr 03, 2004
Posts: 24119
Location: The Netherlands, Enschede
Audio files: 279
G2 patch files: 320
|
| Posted: Tue Oct 30, 2007 11:39 am Post subject:
|
|
|
| cappy2112 wrote: | | Can you imagine writing the NM1 or G2 editors in assembly? |
Sure, why not, a good macro assembler is very capable of making abstractions. As you said there is no real need to do so ... and there is the disadvantage of having to learn a new language for each processor ... but when writing parts in high level and parts in assembly you'd have to learn the assembler for the specific processor anyway (except when working with a team of course, where you could have just one "processor specific" expert probably).
_________________
Jan
also .. could someone please turn down the thermostat a bit.
 |
|
|
Back to top
|
|
|
Kassen
Janitor
Joined: Jul 06, 2004
Posts: 7678
Location: The Hague, NL
G2 patch files: 3
|
| Posted: Sun Nov 04, 2007 12:49 pm Post subject:
|
|
|
Today I ran into a actual Leopard demonstration at a electronics store and after looking at some of the demonstration models (the only one that had a screen that looked good to me turned out to be displaying the memory dump of a crashed OS but it did so in nicely stylized B&W) I got the chance to ask some of the questions that I've been having to a self-proclaimed Apple expert in a jacket stamped with Apples on all sides.
After a few moments I already had to explain to him what Cocoa and X-server are and I think he only had a very vague idea about what Quartz does.
I'm now wondering to what lengths I would have to go to get some solid info on OSX tuning because this is getting quite silly.
_________________
Kassen |
|
|
Back to top
|
|
|
dewdrop_world
Joined: Aug 28, 2006
Posts: 858
Location: Guangzhou, China
Audio files: 4
|
| Posted: Sun Nov 04, 2007 1:05 pm Post subject:
|
|
|
Where was this? An Apple store? If so, they are retail drones, not software engineers.
If it was some kind of trade show, you were probably talking to marketing types, again not engineers.
You'd have more luck on a web forum for Mac devs. (And I don't mean the official Apple-hosted ones.)
James
_________________
ddw online: http://www.dewdrop-world.net
sc3 online: http://supercollider.sourceforge.net |
|
|
Back to top
|
|
|
Kassen
Janitor
Joined: Jul 06, 2004
Posts: 7678
Location: The Hague, NL
G2 patch files: 3
|
| Posted: Sun Nov 04, 2007 1:37 pm Post subject:
|
|
|
A large media store (sells computers, consoles, stereo equipment, televisions, etc) that is now apparently also a Apple store. Some of the people that work there know everything and will answer questions honestly, most of them I wouldn't trust to even plug in a power-chord without a manual nearby and the phone number of a helpdesk.
Oh, and another question, one that doesn't involve any deep system stuff. If I hook up a PC keyboard to a Mac it will lack the "command" key, obviously, so that will need to be re-mapped. I read that MacOS generally takes the "Windows" keys for this but the keyboards that I like predate "Windows" keys by a decade or so.
Can you hook up a 80's PS/2 keyboard through a PS/2 to USB converter and still get the command key functionality, for example through one of the Alt keys? I suppose I could order one of those modern Model M re-makes that do have such keys but those were over a 100 bucks a pop last time I looked and shipping a 2.5+Kg keyboard internationally is probably quite expensive as well. Alternately I could get one of those modern remakes of terminal keyboards that are supposed to be even better (and larger, and heavier, and yet more expensive..) does OSX support old-school terminal keyboard keys? Can you have multiple keyboard profiles and use hotkeys to swap between them?
I can't find any of that stuff on the Apple dev pages and didn't dare ask this poor young man. I'm not being difficult, really, but I absolutely need a Model M or equivalent, when I travel with my laptop I always bring one in a large back-pack because I find modern keyboards and especially laptop ones impossible to type more then a few sentences on. As I understand it all I will absolutely need the command key functionality because so many important keyboard commands depend on it.
I hadn't considered this at all yet :¬(
_________________
Kassen |
|
|
Back to top
|
|
|
Acoustic Interloper
Joined: Jul 07, 2007
Posts: 2070
Location: Berks County, PA
Audio files: 89
|
| Posted: Sun Nov 04, 2007 1:45 pm Post subject:
|
|
|
If it was in an Apple store, then it was probably somebody from the so-called Genius Bar When my daughter had to take her new iMac in for repairs at 2.5 weeks (fewer than 2 weeks would have given her a new machine), 1.5 week ago, a Smarmy Genius (as she described him behind the Genius Bar asked her, "Before it died, wasn't it really great?"
My 27 year old Apple II Plus still works, never had a hardware problem, despite my poking around its insides. They received the replacement parts that they guessed (power supply + some logic board, just in case) at end of the week, told her it might be another 5 days until they had time to do the repair, so my wife called up cross country to tell them what she thought. They fixed it that day. Thank God for Mothers  Of course they had horror stories in order to push the extended warranty. It used to be you never needed those warrantees, but now I guess they are built into the cost/profit structure.
My wife started using my new Mac laptop last night, now that she finally has a digital camera, so I guess we are entering the iTimes. I was sadly disappointed last night, though, when we wanted to watch a "free movie short (15 minutes) available exclusively on iTunes," and the first thing they required for the "free" movie was a credit card number. I would have settled for giving an email address.
So far, no worse than Windows XP for me, although my sister-in-law ignored my advise to buy an OSX machine a few weeks ago, bought a Vista machine with a reasonable amount of memory, then took it back after a week because it was so sluggish, and because she hates Vista look&feel so bad (double the clicks for the same work). I guess I'll have to walk her through buying an XP machine on-line from 1000 miles away; can't get them in the retail stores any more.
I finally figured out one thing that was definitely better in the good old digital days. It was like living in a frontier. Not much support infrastructure, also not much clutter. Like creating on a clean, blank canvas. Now it is like living in a strip mall. Maybe I can figure out the O.S. equivalent of the gated community, for those willing to shell out the money to keep out the riff raff 
_________________
When the stream is deep
my wild little dog frolics,
when shallow, she drinks. |
|
|
Back to top
|
|
|
Kassen
Janitor
Joined: Jul 06, 2004
Posts: 7678
Location: The Hague, NL
G2 patch files: 3
|
| Posted: Sun Nov 04, 2007 1:51 pm Post subject:
|
|
|
Ooooh!
http://www.matias.ca/tactilepro2/forums.html
It's a real keyboard, except made for Mac's! And supposed to be even louder then the IBM! (sadly also over a 100 Euro and it doesn't look like it has a proper springy cable).
_________________
Kassen |
|
|
Back to top
|
|
|
Kassen
Janitor
Joined: Jul 06, 2004
Posts: 7678
Location: The Hague, NL
G2 patch files: 3
|
| Posted: Sun Nov 04, 2007 2:05 pm Post subject:
|
|
|
| Acoustic Interloper wrote: |
I finally figured out one thing that was definitely better in the good old digital days. It was like living in a frontier. Not much support infrastructure, also not much clutter. Like creating on a clean, blank canvas. Now it is like living in a strip mall. |
I find it's mainly the clutter that makes me need support, but maybe when one needs support for something it very quickly starts to look like clutter?
I think there are still plenty of "frontier" ares but they -sadly- do depend on things that are less then exciting. I do quite like this whole virtual machine idea where your playground is nicely fenced off. It's like the outside world doesn't really bother you there that much. Until, of course, the fence keeps you from addressing bits that would be fun but aren't abstracted to it.
_________________
Kassen |
|
|
Back to top
|
|
|
cappy2112
Joined: Dec 24, 2004
Posts: 2471
Location: San Jose, California
Audio files: 2
G2 patch files: 1
|
| Posted: Sun Nov 04, 2007 4:20 pm Post subject:
|
|
|
| Kassen wrote: | | T displaying the memory dump of a crashed OS but it did so in nicely stylized B&W) |
Do you remember when Bill Gates did a public demo of one of the new versions of Windows ( I don't remember which one, but I think it was Pre XP)- The demo crashed while he was doing his presentation. How appropo! |
|
|
Back to top
|
|
|
Acoustic Interloper
Joined: Jul 07, 2007
Posts: 2070
Location: Berks County, PA
Audio files: 89
|
| Posted: Sun Nov 04, 2007 5:22 pm Post subject:
|
|
|
| Kassen wrote: | | I do quite like this whole virtual machine idea where your playground is nicely fenced off. It's like the outside world doesn't really bother you there that much. Until, of course, the fence keeps you from addressing bits that would be fun but aren't abstracted to it. |
Yes, back in the day when I was the architect for a DSP simulator/emulator/debugger that operated under a Tcl/Tk UI, I could spend the whole work day inside the tool, typing into Tcl + tool extensions, regardless of whether running on UNIX or Windows, blissfully detached from outsdie influences. Occasionally I would forget and type 'make' for the tool itself from within the tool, and would get the error when the linker tried to replace the file that I was living inside. There's a paper on it here and a related one here. I like the two figures below, from the respective papers. I miss those days.
| Description: |
|
| Filesize: |
119.85 KB |
| Viewed: |
111 Time(s) |
| This image has been reduced to fit the page. Click on it to enlarge. |
|
_________________
When the stream is deep
my wild little dog frolics,
when shallow, she drinks. |
|
|
Back to top
|
|
|
Kassen
Janitor
Joined: Jul 06, 2004
Posts: 7678
Location: The Hague, NL
G2 patch files: 3
|
| Posted: Sun Nov 04, 2007 5:38 pm Post subject:
|
|
|
Cappy; Sure, such things happen, this was computer that was open to the general public to try so who knows what happened to it. I'm not holding it against Apple, I just wanted to mention it as it was a part of this scenario as I saw it. I also noticed that the non-apple-jacket-wearing shop people had considerable difficulty demonstrating basic features and said things like "oh, this doesn't work on this one, let's move to the next model here...." It's a brand new OS as well, or at least a brand new major update, I'm not sure how I'm supposed to see that as it's not yet called "OSXI" but it still is a paid update. I'm sure all will be fine with a few service packs. I'm used to stuff like that and besides, none of the stuff they were demoing holds any interest to me so for all I care it could crash non-stop.
_________________
Kassen |
|
|
Back to top
|
|
|
blue hell
Site Admin
Joined: Apr 03, 2004
Posts: 24119
Location: The Netherlands, Enschede
Audio files: 279
G2 patch files: 320
|
| Posted: Sun Nov 04, 2007 5:45 pm Post subject:
|
|
|
| Acoustic Interloper wrote: | | There's a paper on it here and a related one[...] |
Looks like a good read (but must do so later), I'm looking into improving a debugging environment currently
_________________
Jan
also .. could someone please turn down the thermostat a bit.
|
|
|
Back to top
|
|
|
Kassen
Janitor
Joined: Jul 06, 2004
Posts: 7678
Location: The Hague, NL
G2 patch files: 3
|
| Posted: Sun Nov 04, 2007 5:54 pm Post subject:
|
|
|
| Acoustic Interloper wrote: | | I miss those days. |
Well, the days of the VM are far from over. I would say that with things like the advance of portable computing and some of the exotic hardware that runs on ( I think I read about one mobile phone that was a 4bit dual processor... ) as well as expectations of consistency they will continue to grow in popularity.
_________________
Kassen |
|
|
Back to top
|
|
|
|
Forum index » Discussion » Schmooze
FAQ
Calendar
Search
Memberlist
Usergroups
Register
Profile
Log in to check your private messages
Log in
Chat Room
