First transfer -- music is sold to someone else. Is the watermark ownership transferred?
A bit more complicated -- music is purchased in the US. Buyer travels to Canada. A Canadian copies the music (legally). Now, there are two (legal) copies; one in the US and one in Canada. The Canadian now travels to the US, and has her laptop (with the copy on it) checked. She is detained. What law was broken?
So, you resort to ad-hominem attacks. It doesn't matter who co-sponsored or co-created to make my argument -- indeed, that is what I am railing against. Now, I shall return the ad-hominem attack (in kind). You question my ability to read; you should know that is was PHILIPS and Sony who produced the CD.
Now back to reasoned argument.
Paying extra for combos? The Bourne Ultimatum (I just picked one title): HD DVD and DVD, 27.87 at Walmart (Canada). 16.87 for the DVD. It's difficult finding the exact same title, so I picked WAR on Blu-Ray to be comparable. 28.88 at Walmart. 19.87 for the DVD. Scaling for the "value" of the movie (we have to add 3 dollars to the HD DVD, because the DVD price differential is 3 dollars), and rounding:
HD DVD/DVD 31 dollars, BLU-RAY/DVD 41 dollars. BLU-RAY only: 29 dollars
So, I guess you are right - buying the BLU-RAY will save you 2 dollars a movie, without the DVD compatibility. Want the DVD? It's 10 dollars more. (We could have doubled the price differential, but the price scaling at Walmart appears to be fixed to a base. I could run all the numbers from Walmart through a spreadsheet, but *YOU* made the accusation -- burden of proof is yours).
Email me at my munged email address if you are indeed serious about selling your HD DVD collection.
The HD-DVD format allowed for DVD on one side, HD-DVD on the other... Blu-Ray? And, Blu-Ray is from Sony; arguably the second-most evil corporation (ahead of Microsoft, and just behind Disney).
Blu-Ray -- the Sony/Disney mashup from hell. Of course the other studios feel they have to jump on the bandwagon.
I liked the idea of buying DVD/HD-DVD in the same package...
"Office, Scrivener, Coda, the Adobe suite, Garage Band, iMovie, and iDVD."
Well, not true. By "Office" I presume you mean that useless piece of shite sold by Microsoft. Ok, I have even tried to USE that. That was the word processor that self-immolated when I played with the cursor keys too much (adjusting a table). Never could figure that one... Oh, it ate the document when it happened...
The rest I actually have to "Google":
Scrivener: An editor. With an integrated cork board. And a "vi mode". For fourty bucks. Of course, its from "Literature and Latte".
The Adobe Suite: um... "creative tools for quiche eaters". I went there, and indicated what I do... Here it is: Go to the Product Selector: I create images for publication (graphics, diagrams, I mostly use the troff set currently), also I generally use Tex so I choose Design/Print Publishing; I do simple Web pages (mostly information publication, static web) and a bit of UI design (what use is software WITHOUT a user interface, after all?). And, I prefer the UI to be "behind" either Apache with CGI or TCL/TK, because I hate the work. So I choose WEB/Web Developement and WEB/User interface Design. I don't really deal with Photography (other than tossing JPEG files into a movie or onto the web) so I have no picks in that category. I do author DVDs (simple shite, mostly I just use boilerplate and go), so I choose Video/DVD-authoring. And TADA, the fucking quiche eaters want two thousand five hundred dollars.
Garage Band - Have you EVER heard me "podcast". No? There is a reason. iMovie, iDvd - Not what I do with computers. But, ok, the ONLY sensible thing here. Make family DVDs. But... why don't I just go and buy a direct-to-DVD camcorder, and give the mess to my 17 year old nephew?
So, with Garage Band, iMove and iDvd you have something that will interest my nephew. With a 2500 dollar price tag on the Adobe software, you have something that I will never buy, given that my use is simple, and WORKS NOW. With the recommendation of "Office" -- well that CAN'T be a recommendation, because I have actually used the product and it is shite (Outlook good, Word, Excel, PowerPoint - Shite).
This is going to convince me that an Apple OSX system is worth buying? Maybe you have convinced me to buy one for the nephew, though...
Although I *did* see a feature of the OSX that I liked -- the transparent terminal...
Come on, flame away! I am utterly sure that the Mac Fanbois can't hold back...
Pepsiman - thanks for the reference. I haven't (as yet) read the literature, but it was an interesting issue at the time, and the idea of generalizing to a peephole optimizer is very interesting.
GCC a clone of CC... No its not. It's an implementation of a published computer language. Linux and BSD are implementations of an OS that was, itself, "open source" (Lions Commentary on Unix) (or Minix). And, even if the IMPLEMENTATION was restricted, the specification (POSIX) is open, and even, um... REQUIRED?
Unix has NEVER been "closed source" in the sense of Microsoft. Microsoft Office? Sure, and, yes OpenOffice is a "clone". And, as such, suffers. But, proper typography is the domain of TeX, which has ALWAYS been open source. There is no closed source comparision.
You have to distinguish between closed implementation of an open idea and open implementation of an open idea, and a completely closed idea. Office exemplifies the second. Open source is the first two categories. An example of "closed implementation of an open idea" is Adobe Postscript. Arguably, and EXCELLENT idea, and the premier implementation is Ghostscript (open implementation of that idea). Yes, closing an implementation for a time can make some money, and it doesn't hurt the community much. Eventually, there can be an open implementation.
As to the "windows (tm)" shell -- which one? Windows 3.1? A pale implementation of a shell Windows 95? Better -- finally ripped off that "right mouse button" idea that was put into the "open" realm. Anyway, by now there has been so much inter-breeding in the GUI area that trying to call it on "originally open" vs. "originally closed" is senseless.
Mono?.net, sure, and.net? jvm for sure. And jvm? Published in a book as an "open idea".
Actually, all "got a hold of" computers at the same time.
The history of programming languages is quite instructive. It relates to algorithms, and possible uses.
First, we have machine code - tedious, and then assemblers (autocoders). Designed to take SOME of the burden of counting. (labels, branch targets).
Then, the idea of compiling formulas (math) -> FORTRAN. The experience of FORTRAN leads to BNF format (Backus *is* Mr. Fortran), which brings us to: COMIT and SNOBOL
In SNOBOL4, a BNF description is (more than less) DIRECTLY executable. Its all about patterns, right? FWIW - SNOBOL4 (1966) has associative tables, patterns more powerful than REs, self-modifying code, dynamic memory, recursion.
Also, interest in how computers (should) work, and the "correct" way to do a program led to LISP. At the same time as FORTRAN.
Interest in accounting led to COBOL in 1960.
And, iterest in mathematical programming (arrays, etc.) led to APL (at the same time). APL and LISP have dynamic memory management as well.
So, we have:
FORTRAN (static, no need for gc)-> ALGOL -> PASCAL (dynamic, no gc - new/free) -> C -> JAVA (dynamic, gc), etc. SNOBOL (dynamic, gc) -> ?? (bizarre, the only "dead end" here) LISP (dynamic, gc) -> SCHEME, functional languages COBOL (static, no need for gc) -> well, COBOL APL (dynamic, gc) -> J, Matlab
The reason I believe SNOBOL was "dead-ended" is that it was simply TOO powerful to be taught (I recall a prof saying "to solve this, use ANY language, EXCEPT SNOBOL).
In a sense, LISP is similar, given a rich library (Common LISP)...
Why did only one of these strands become "main-stream"? An excellent question. One reason is that ACM wanted algorithms in ALGOL/FORTRAN. Another is that computers were VERY expensive, and avoidance of garbage collection was almost -- religious -- for a while. A third was the introduction of "FORTRAN in short pants" aka BASIC.
Also the FORTRAN strand seems to model the perceptions of GC efficiency as time goes on (initially, on static allocations, then new/free, then dynamic with gc).
And, I want to point out the practical application -- computer graphics. Our friends at Microsoft had decided that "ROP Functions" should be implemented by display drivers. Indeed, the "ROP Functions" were this set of 256 functions. Windows 3 days; I am not familiar if this is currently the case.
The solution? Since the target was always an Intel processor, generate machine code sequences of increasing time complexity, and, as each is generated, see if it solves one of the Functions. Since the seach is over the complete legal instruction sequence space, and is ordered by speed, it guaranteed the lowest cost operation for each of these functions.
The code was prepared, and run, in 1991. Run time? Minutes. It wrote the ROP compiler (at this point, Windows 3 used a "Blit Compiler", so this would be a meta-compiler). (PS. It gave some counters to Worlfram's Conjectures). Of course, it didn't actually PROVE anything (at least in symbol length terms) because it was driven by least execution cost, without being limited to that set of operations (any x86 opcode was available, ordered by opcode cost on a 80486).
The devs name was Pat (something?) -- Paul Hsieh was also in my group at the time, and may remember Pat's last name (to give proper credit). I just reviewed the results of the program (being a "manager type" at the time, and trying to convince Microsoft that a "DirectX" architecture would be useful, and that framebuffer drivers could be used for underlying support -- we demoed these technologies to Microsoft in '90, but didn't have "permanent" Microsoft resident staff at Redmond till a year later - which ended up being terrifically funny, because our proprietary code was checked in at MS, and was then ripped off by the competition, bugs and all).
I just wanted to record some of the history, and see if anyone else remembers Pat to give him the well-deserved credit for writing the "Function Finder" that ended up on millions of PCs.
"Years yes... but they are still fairly new compared to the ACLs that are present in NTFS. Also, ACL's aren't really universal or standardized in linux."
Linux ACLs are an implementation of POSIX 1003.1e draft 17. Every Linux. Everwhere. Maybe "still fairly new compared to the ACLs that are present in NTFS" is true.
That IS a problem. But it's not the reason ACLs (discretionary access control) isn't used much. Indeed, it isn't used on Windows, either.
The reason(s) its not used are:
1 - too complicated. period. 2 - only NEEDED for 10% or less of the cases. The other 90% is handled nicely by uga permissions. 3 - see (1). Because it's THAT complicated, ACLs cannot be audited. Since the auditing is impractical, its not going to be done. As a DIRECT result, the calculus of ACLs must be such that ACLs can be effectively reasoned about. And with SO many properties in the VMS/Windows world, that ain't happening.
To give you a PRACTICAL example of "ACL failure", let me relate a VMS story. I had set my script that is automatically executed at startup to readable by anyone, and writeable by me. Only. Except that I was a VMS "newbie" (only been using it for a year!) and I hadn't realized that "write" ACTUALLY meant WRITE TO THIS VERSION OF THE FILE.
Think about it -- my default home directory was set to remove the WRITE permission, but CREATE... another story. And, since VMS stores an automatic backup of a file on CREATE (versioning), another user was able to simply CREATE a new startup script in my directory. I was wheeled this way (and schooled, see below). The newly created script sucked my files (executing as me, of course, on my next login). And it then erased itself (now having permission to do so).
The schooling? I tend to like the uga scheme more. There are ONLY 3 permissions to reason about (actually 5), and only 3 classes of users. The permissions are contained.
Anyway, the discussion is about MACs not ACLs, which are a completely different thing. I believe that the NT core tries to apply ACLs to things that are not files (eg. processes). If so, I believe that is a mistake (re-read my horror story above).
"The author of the posting clearly has no knowledge of the state of Microsoft software and development tools today. Take one look at the.NET Framework and not only is it a ripoff of Java, but it made huge improvements like making a language-agnostic programming platform (parially due to CTS and CLI) and allowing multiple syntaxes (yes even Java-like syntax) to interoperate. Programmers can work in their language of choice and the compiled code will interoperate with all the other.NET languages which were other programmer's choices. That's one example."
Just happened to be at the start. Is.NET a ripoff of Java? Is this innovative? Isn't the JVM a "language-agnostic programming platform"? Doesn't JVM allow multiple syntaxes? (REXX comes to mind immediately). About the only thing that was tricky was ensuring that the JVM verifier would accept your byte-code -- for security reasons... And, I thought that was pretty innovative. "Byte code" systems have been around for a LONG time (Pascal P-machine, Smalltalk, etc.).
Which means -- I don't know WHAT your point was in that paragraph. I *think* that you are trying to say that.NET is innovative because it stressed multi-language development. You must have forgotten that JVM (Java) allowed exactely that as well. There is even a JVM assembler available! This then leaves the "Microsoft.NET is a ripoff of JAVA", with no improvement, and renders the rest of your post questionable.
How do you create accounts on OS-X with a known UID/GID? AFAIK, its from the command line
And, how about automount? AFAIK (and, really, I'm no expert at OS-X here, just on Solaris, Linux, AIX and HP/UX), you need to edit/etc/hostconfig, setting AUTOMOUNT=-YES-. Correct me if I am wrong...
As to automounter maps -- I generally serve them over NIS. It just works (tm). Of course LDAP may be more suitable for signon in organizations with LOTS of people, but NIS works for tens to hundreds of entries just fine. And, really, have you ever seen automounter maps with more than a couple of hundred entries? That's a whale-load of servers... Of course they can supplied out of LDAP... BUT (and this is a big one)... the only best practices for OS-X deployment I have seen says "OS-X *might* be able to read maps from LDAP", and goes on to recommend that a script be written to read the information OUT of LDAP, and put it into/etc/auto.nfs (if you have a better best practice for architects, give me a reference). So, OS-X is only able to reliably support automounts from NIS, and you have declared NIS to be dead? I like your style.
OS-X is probably just fine for an "at-home" installation. Probably just fine for a small workgroup (possibly even with some Windows and a Unix/Linux box or two). It doesn't fit into the larger environments, except as a limited client, and even there, can be a serious bear to configure. Please note that Windows has the same problems, and WORSE. But there is enough traction to make IT infrastructure BEND into the Windows Way. This is done by adding Exchange Servers to the central infrastructure, and adding central scripts to replicate Exchange signon into LDAP or NIS (Microsoft even helps with this!). But, given the ludicrously small install base of OS-X there is no such help... and, unlike Linux, where the answer to the question "Does automounter accept LDAP maps?" is simply YES, OS-X makes the audacious demand to "learn netinfo".
I will take a slightly contrarian point here. The "Mac HIG" was originally horrible. In promoting a "common" menu bar that forced applications into a particular mode, whether or not it was suitable, coupled with the need to continually go to the menu, coupled with the idiocy of over-loading the (single) mouse button. HIG? I would say In-humane guidlines.
The situation improved somewhat with OS-X. It STILL has horrible warts, which make it impossible (yes, you did read that correctly) for many of us to use. Well -- not exactly impossible. I imagine that OS-X could be coerced into running under some configuration of vmware... (although it isn't nice).
OS-X does have a VNC server available (thank the godess), so not all is lost. Still, NFS on OS-X *REQUIRES* the command line (by default, I do believe that there are some 3rd party GUI tools). NIS also has problems (pretty intense command line work for setup) -- and NIS auto.master maps for automount don't work.
So, OS-X is a client side (pretty much) Unix based OS that doesn't respect much of the Unix auto configuration. Is the OS-X side better? Not sure, it may be.
As to the much-vaunted "consistency"? It came at great cost in the OS-9 and before days (horrid inefficiencies in USING the GUI). OS-X? I have tried it, but I can't say for sure.
Before hammering Xerox PARC, please have a close look at Squeak (and, yes, it is available on Macs).
"Then go out an try to vi an 800MB file, be it on Linux, Solaris, HP-UX, whatever. It takes ages to load on them too."
It may. That's why we have POSIX compliant programs like sed. "Stream Editor". To be used in precisely these cases.
But, ok, I'm crazy!!! How crazy? Let me tell you... I am running Fedora Core 5 Linux on a Panasonic Toughbook CF-27. 6GB of the slowest hard disk, coupled with 128MB of memory. So far, so good? And, icing on the cake, it's a 350Mhz Pentium II.
Running with the GUI (Gnome, the default) leaving web browser and 3 other windows open (not sure what they are running, and, frankly, I don't care), I am going to use vi (VIM 7, w/ bell and whistles) on a 618MB text file.
Yup! I told you, I am crazy: Crazy Freddy, they call me!!! Let's see how long it takes (because I am not going to just accept "takes ages" without measuring just what ages is):
[fred@pandora snobol4]$ time vi y
real 4m59.040s user 1m36.500s sys 0m35.390s [fred@pandora snobol4]$ ls -hl y -rw-rw-r-- 1 fred fred 618M Dec 13 13:29 y
It took 5 minutes, of which 2 minutes was spent on processor and 3 whole minutes physically reading the file. So, by substitution, takes ages == 5 minutes (actually, = 5 minutes, because I am deliberately timing this on 10 year old hardware).
I wonder what 4 hours is on that timescale?
Or are you telling me that Windows is SO bad as an environment that (1) tools for this are not provided, and (2) the standard system editor supplied by the OS vendor is incapable of handling the problem, forcing me to use 3rd party tools for something this simple?
And, as usual, I don't care. Because I freelance by the hour.
We have companies that want me to do creative work (design and development of a software or hardware/software system).
Instead of being given a quiet area conducive to concentration, I am put into either a cubicle farm or a completely open area. With interruptions. With noise. I have even been put into a cubicle, shared with someone else, *and* three servers, *and* two automated CD pressing machines.
I have a noise-cancelling headset (takes care of low freq noise), and a lot of patience.
Still, I have a theory -- could not the environment be why F/OSS can be soooo much better sometimes?
I wonder what the working conditions at TATA are like?
The BSA does not accept a "proof of license" sticker. You MUST present a receipt for the software. If you go to a trade show, and accept a "free" copy of swagware (say, a copy of Windows) that does NOT have a receipt, you will be counted in violation.
Under THOSE rules, 8% is (translated) comletely in compliance.
[and yes, I do not accept that "free" software because it exposes my business to the BSA, and (reading the EULA fine print) at least one commercial software product put audits in! (possibly on my dime -- I threw the software out, because the terms were so bizarre)]
Sam Bronfman made his fortune by bootlegging whiskey from Canada to the US during Prohibition, dealing with the likes of Al Capone in Chicago. Edgar is his son, born in 1929.
The ISP is in control of the DNS server *and* the proxy server. Either a DNS redirect to the ISP can be done, or an HTTP redirect, thus changing the name of the site to the site approved by the CA. The man in the middle works from there. Most users won't notice, anyway (or care, given the number of trojans around). The "internet" works as usual...
Slashdot Mirror
Assume a perfect watermarking system.
First transfer -- music is sold to someone else. Is the watermark ownership transferred?
A bit more complicated -- music is purchased in the US. Buyer travels to Canada. A Canadian copies the music (legally). Now, there are two (legal) copies; one in the US and one in Canada. The Canadian now travels to the US, and has her laptop (with the copy on it) checked. She is detained. What law was broken?
So of what use IS the watermark?
And what format is Disney using?
So, you resort to ad-hominem attacks. It doesn't matter who co-sponsored or co-created to make my argument -- indeed, that is what I am railing against. Now, I shall return the ad-hominem attack (in kind). You question my ability to read; you should know that is was PHILIPS and Sony who produced the CD.
Now back to reasoned argument.
Paying extra for combos? The Bourne Ultimatum (I just picked one title): HD DVD and DVD, 27.87 at Walmart (Canada). 16.87 for the DVD. It's difficult finding the exact same title, so I picked WAR on Blu-Ray to be comparable. 28.88 at Walmart. 19.87 for the DVD. Scaling for the "value" of the movie (we have to add 3 dollars to the HD DVD, because the DVD price differential is 3 dollars), and rounding:
HD DVD/DVD 31 dollars, BLU-RAY/DVD 41 dollars. BLU-RAY only: 29 dollars
So, I guess you are right - buying the BLU-RAY will save you 2 dollars a movie, without the DVD compatibility. Want the DVD? It's 10 dollars more. (We could have doubled the price differential, but the price scaling at Walmart appears to be fixed to a base. I could run all the numbers from Walmart through a spreadsheet, but *YOU* made the accusation -- burden of proof is yours).
Email me at my munged email address if you are indeed serious about selling your HD DVD collection.
The HD-DVD format allowed for DVD on one side, HD-DVD on the other... Blu-Ray? And, Blu-Ray is from Sony; arguably the second-most evil corporation (ahead of Microsoft, and just behind Disney).
Blu-Ray -- the Sony/Disney mashup from hell. Of course the other studios feel they have to jump on the bandwagon.
I liked the idea of buying DVD/HD-DVD in the same package...
Wireless networks - never been a problem. Just get wired to wireless bridge, and let the hardware deal with the issue.
Printers - never been an issue, either. Just set up a print server, and hump postscript around. Is that so hard?
Because I have NEVER HEARD of those programs.
"Office, Scrivener, Coda, the Adobe suite, Garage Band, iMovie, and iDVD."
Well, not true. By "Office" I presume you mean that useless piece of shite sold by Microsoft. Ok, I have even tried to USE that. That was the word processor that self-immolated when I played with the cursor keys too much (adjusting a table). Never could figure that one... Oh, it ate the document when it happened...
The rest I actually have to "Google":
Scrivener: An editor. With an integrated cork board. And a "vi mode". For fourty bucks. Of course, its from "Literature and Latte".
The Adobe Suite: um... "creative tools for quiche eaters". I went there, and indicated what I do... Here it is: Go to the Product Selector: I create images for publication (graphics, diagrams, I mostly use the troff set currently), also I generally use Tex so I choose Design/Print Publishing; I do simple Web pages (mostly information publication, static web) and a bit of UI design (what use is software WITHOUT a user interface, after all?). And, I prefer the UI to be "behind" either Apache with CGI or TCL/TK, because I hate the work. So I choose WEB/Web Developement and WEB/User interface Design. I don't really deal with Photography (other than tossing JPEG files into a movie or onto the web) so I have no picks in that category. I do author DVDs (simple shite, mostly I just use boilerplate and go), so I choose Video/DVD-authoring. And TADA, the fucking quiche eaters want two thousand five hundred dollars.
Garage Band - Have you EVER heard me "podcast". No? There is a reason.
iMovie, iDvd - Not what I do with computers. But, ok, the ONLY sensible thing here. Make family DVDs. But... why don't I just go and buy a direct-to-DVD camcorder, and give the mess to my 17 year old nephew?
So, with Garage Band, iMove and iDvd you have something that will interest my nephew. With a 2500 dollar price tag on the Adobe software, you have something that I will never buy, given that my use is simple, and WORKS NOW. With the recommendation of "Office" -- well that CAN'T be a recommendation, because I have actually used the product and it is shite (Outlook good, Word, Excel, PowerPoint - Shite).
This is going to convince me that an Apple OSX system is worth buying? Maybe you have convinced me to buy one for the nephew, though...
Although I *did* see a feature of the OSX that I liked -- the transparent terminal...
Come on, flame away! I am utterly sure that the Mac Fanbois can't hold back...
It is perhaps an "unintended consequence", but, yes, correct as published.
Pepsiman - thanks for the reference. I haven't (as yet) read the literature, but it was an interesting issue at the time, and the idea of generalizing to a peephole optimizer is very interesting.
GCC a clone of CC... No its not. It's an implementation of a published computer language. Linux and BSD are implementations of an OS that was, itself, "open source" (Lions Commentary on Unix) (or Minix). And, even if the IMPLEMENTATION was restricted, the specification (POSIX) is open, and even, um... REQUIRED?
.net, sure, and .net? jvm for sure. And jvm? Published in a book as an "open idea".
Unix has NEVER been "closed source" in the sense of Microsoft. Microsoft Office? Sure, and, yes OpenOffice is a "clone". And, as such, suffers. But, proper typography is the domain of TeX, which has ALWAYS been open source. There is no closed source comparision.
You have to distinguish between closed implementation of an open idea and open implementation of an open idea, and a completely closed idea. Office exemplifies the second. Open source is the first two categories. An example of "closed implementation of an open idea" is Adobe Postscript. Arguably, and EXCELLENT idea, and the premier implementation is Ghostscript (open implementation of that idea). Yes, closing an implementation for a time can make some money, and it doesn't hurt the community much. Eventually, there can be an open implementation.
As to the "windows (tm)" shell -- which one? Windows 3.1? A pale implementation of a shell Windows 95? Better -- finally ripped off that "right mouse button" idea that was put into the "open" realm.
Anyway, by now there has been so much inter-breeding in the GUI area that trying to call it on "originally open" vs. "originally closed" is senseless.
Mono?
Sorry for replying to myself -- the fellows name is Pat Maupin. Credits to him!
Actually, all "got a hold of" computers at the same time.
/dev/null, as usual.
The history of programming languages is quite instructive. It relates to algorithms, and possible uses.
First, we have machine code - tedious, and then assemblers (autocoders). Designed to take SOME of the burden of counting. (labels, branch targets).
Then, the idea of compiling formulas (math) -> FORTRAN. The experience of FORTRAN leads to BNF format (Backus *is* Mr. Fortran), which brings us to: COMIT and SNOBOL
In SNOBOL4, a BNF description is (more than less) DIRECTLY executable. Its all about patterns, right? FWIW - SNOBOL4 (1966) has associative tables, patterns more powerful than REs, self-modifying code, dynamic memory, recursion.
Also, interest in how computers (should) work, and the "correct" way to do a program led to LISP. At the same time as FORTRAN.
Interest in accounting led to COBOL in 1960.
And, iterest in mathematical programming (arrays, etc.) led to APL (at the same time). APL and LISP have dynamic memory management as well.
So, we have:
FORTRAN (static, no need for gc)-> ALGOL -> PASCAL (dynamic, no gc - new/free) -> C -> JAVA (dynamic, gc), etc.
SNOBOL (dynamic, gc) -> ?? (bizarre, the only "dead end" here)
LISP (dynamic, gc) -> SCHEME, functional languages
COBOL (static, no need for gc) -> well, COBOL
APL (dynamic, gc) -> J, Matlab
The reason I believe SNOBOL was "dead-ended" is that it was simply TOO powerful to be taught (I recall a prof saying "to solve this, use ANY language, EXCEPT SNOBOL).
In a sense, LISP is similar, given a rich library (Common LISP)...
Why did only one of these strands become "main-stream"? An excellent question. One reason is that ACM wanted algorithms in ALGOL/FORTRAN. Another is that computers were VERY expensive, and avoidance of garbage collection was almost -- religious -- for a while. A third was the introduction of "FORTRAN in short pants" aka BASIC.
Also the FORTRAN strand seems to model the perceptions of GC efficiency as time goes on (initially, on static allocations, then new/free, then dynamic with gc).
Opinions welcome, flames to
Thank you.
And, I want to point out the practical application -- computer graphics. Our friends at Microsoft had decided that "ROP Functions" should be implemented by display drivers. Indeed, the "ROP Functions" were this set of 256 functions. Windows 3 days; I am not familiar if this is currently the case.
The solution? Since the target was always an Intel processor, generate machine code sequences of increasing time complexity, and, as each is generated, see if it solves one of the Functions. Since the seach is over the complete legal instruction sequence space, and is ordered by speed, it guaranteed the lowest cost operation for each of these functions.
The code was prepared, and run, in 1991. Run time? Minutes. It wrote the ROP compiler (at this point, Windows 3 used a "Blit Compiler", so this would be a meta-compiler). (PS. It gave some counters to Worlfram's Conjectures). Of course, it didn't actually PROVE anything (at least in symbol length terms) because it was driven by least execution cost, without being limited to that set of operations (any x86 opcode was available, ordered by opcode cost on a 80486).
The devs name was Pat (something?) -- Paul Hsieh was also in my group at the time, and may remember Pat's last name (to give proper credit). I just reviewed the results of the program (being a "manager type" at the time, and trying to convince Microsoft that a "DirectX" architecture would be useful, and that framebuffer drivers could be used for underlying support -- we demoed these technologies to Microsoft in '90, but didn't have "permanent" Microsoft resident staff at Redmond till a year later - which ended up being terrifically funny, because our proprietary code was checked in at MS, and was then ripped off by the competition, bugs and all).
I just wanted to record some of the history, and see if anyone else remembers Pat to give him the well-deserved credit for writing the "Function Finder" that ended up on millions of PCs.
"Years yes... but they are still fairly new compared to the ACLs that are present in NTFS. Also, ACL's aren't really universal or standardized in linux."
Linux ACLs are an implementation of POSIX 1003.1e draft 17. Every Linux. Everwhere. Maybe "still fairly new compared to the ACLs that are present in NTFS" is true.
That IS a problem. But it's not the reason ACLs (discretionary access control) isn't used much. Indeed, it isn't used on Windows, either.
The reason(s) its not used are:
1 - too complicated. period.
2 - only NEEDED for 10% or less of the cases. The other 90% is handled nicely by uga permissions.
3 - see (1). Because it's THAT complicated, ACLs cannot be audited. Since the auditing is impractical, its not going to be done. As a DIRECT result, the calculus of ACLs must be such that ACLs can be effectively reasoned about. And with SO many properties in the VMS/Windows world, that ain't happening.
To give you a PRACTICAL example of "ACL failure", let me relate a VMS story. I had set my script that is automatically executed at startup to readable by anyone, and writeable by me. Only. Except that I was a VMS "newbie" (only been using it for a year!) and I hadn't realized that "write" ACTUALLY meant WRITE TO THIS VERSION OF THE FILE.
Think about it -- my default home directory was set to remove the WRITE permission, but CREATE... another story. And, since VMS stores an automatic backup of a file on CREATE (versioning), another user was able to simply CREATE a new startup script in my directory. I was wheeled this way (and schooled, see below). The newly created script sucked my files (executing as me, of course, on my next login). And it then erased itself (now having permission to do so).
The schooling? I tend to like the uga scheme more. There are ONLY 3 permissions to reason about (actually 5), and only 3 classes of users. The permissions are contained.
Anyway, the discussion is about MACs not ACLs, which are a completely different thing. I believe that the NT core tries to apply ACLs to things that are not files (eg. processes). If so, I believe that is a mistake (re-read my horror story above).
Somehow I don't think that I will be given access to Windows source, even after the research is published.
So I didn't finish your post. I got stuck on:
.NET Framework and not only is it a ripoff of Java, but it made huge improvements like making a language-agnostic programming platform (parially due to CTS and CLI) and allowing multiple syntaxes (yes even Java-like syntax) to interoperate. Programmers can work in their language of choice and the compiled code will interoperate with all the other .NET languages which were other programmer's choices. That's one example."
.NET a ripoff of Java? Is this innovative? Isn't the JVM a "language-agnostic programming platform"? Doesn't JVM allow multiple syntaxes? (REXX comes to mind immediately). About the only thing that was tricky was ensuring that the JVM verifier would accept your byte-code -- for security reasons... And, I thought that was pretty innovative. "Byte code" systems have been around for a LONG time (Pascal P-machine, Smalltalk, etc.).
.NET is innovative because it stressed multi-language development. You must have forgotten that JVM (Java) allowed exactely that as well. There is even a JVM assembler available! This then leaves the "Microsoft .NET is a ripoff of JAVA", with no improvement, and renders the rest of your post questionable.
"The author of the posting clearly has no knowledge of the state of Microsoft software and development tools today. Take one look at the
Just happened to be at the start. Is
Which means -- I don't know WHAT your point was in that paragraph. I *think* that you are trying to say that
Since *I* don't have such access, there is an obvious restriction. Think about the scientific method...
Sure, NFS from finder...
/etc/hostconfig, setting AUTOMOUNT=-YES-. Correct me if I am wrong...
/etc/auto.nfs (if you have a better best practice for architects, give me a reference). So, OS-X is only able to reliably support automounts from NIS, and you have declared NIS to be dead? I like your style.
How do you create accounts on OS-X with a known UID/GID? AFAIK, its from the command line
And, how about automount? AFAIK (and, really, I'm no expert at OS-X here, just on Solaris, Linux, AIX and HP/UX), you need to edit
As to automounter maps -- I generally serve them over NIS. It just works (tm). Of course LDAP may be more suitable for signon in organizations with LOTS of people, but NIS works for tens to hundreds of entries just fine. And, really, have you ever seen automounter maps with more than a couple of hundred entries? That's a whale-load of servers... Of course they can supplied out of LDAP... BUT (and this is a big one)... the only best practices for OS-X deployment I have seen says "OS-X *might* be able to read maps from LDAP", and goes on to recommend that a script be written to read the information OUT of LDAP, and put it into
OS-X is probably just fine for an "at-home" installation. Probably just fine for a small workgroup (possibly even with some Windows and a Unix/Linux box or two). It doesn't fit into the larger environments, except as a limited client, and even there, can be a serious bear to configure. Please note that Windows has the same problems, and WORSE. But there is enough traction to make IT infrastructure BEND into the Windows Way. This is done by adding Exchange Servers to the central infrastructure, and adding central scripts to replicate Exchange signon into LDAP or NIS (Microsoft even helps with this!). But, given the ludicrously small install base of OS-X there is no such help... and, unlike Linux, where the answer to the question "Does automounter accept LDAP maps?" is simply YES, OS-X makes the audacious demand to "learn netinfo".
Little wonder that I find Apple obnoxious.
Brad-X
I will take a slightly contrarian point here. The "Mac HIG" was originally horrible. In promoting a "common" menu bar that forced applications into a particular mode, whether or not it was suitable, coupled with the need to continually go to the menu, coupled with the idiocy of over-loading the (single) mouse button. HIG? I would say In-humane guidlines.
The situation improved somewhat with OS-X. It STILL has horrible warts, which make it impossible (yes, you did read that correctly) for many of us to use. Well -- not exactly impossible. I imagine that OS-X could be coerced into running under some configuration of vmware... (although it isn't nice).
OS-X does have a VNC server available (thank the godess), so not all is lost. Still, NFS on OS-X *REQUIRES* the command line (by default, I do believe that there are some 3rd party GUI tools).
NIS also has problems (pretty intense command line work for setup) -- and NIS auto.master maps for automount don't work.
So, OS-X is a client side (pretty much) Unix based OS that doesn't respect much of the Unix auto configuration. Is the OS-X side better? Not sure, it may be.
As to the much-vaunted "consistency"? It came at great cost in the OS-9 and before days (horrid inefficiencies in USING the GUI). OS-X? I have tried it, but I can't say for sure.
Before hammering Xerox PARC, please have a close look at Squeak (and, yes, it is available on Macs).
"Then go out an try to vi an 800MB file, be it on Linux, Solaris, HP-UX, whatever. It takes ages to load on them too."
It may. That's why we have POSIX compliant programs like sed. "Stream Editor". To be used in precisely these cases.
But, ok, I'm crazy!!! How crazy? Let me tell you... I am running Fedora Core 5 Linux on a Panasonic Toughbook CF-27. 6GB of the slowest hard disk, coupled with 128MB of memory. So far, so good? And, icing on the cake, it's a 350Mhz Pentium II.
Running with the GUI (Gnome, the default) leaving web browser and 3 other windows open (not sure what they are running, and, frankly, I don't care), I am going to use vi (VIM 7, w/ bell and whistles) on a 618MB text file.
Yup! I told you, I am crazy: Crazy Freddy, they call me!!! Let's see how long it takes (because I am not going to just accept "takes ages" without measuring just what ages is):
[fred@pandora snobol4]$ time vi y
real 4m59.040s
user 1m36.500s
sys 0m35.390s
[fred@pandora snobol4]$ ls -hl y
-rw-rw-r-- 1 fred fred 618M Dec 13 13:29 y
It took 5 minutes, of which 2 minutes was spent on processor and 3 whole minutes physically reading the file. So, by substitution, takes ages == 5 minutes (actually, = 5 minutes, because I am deliberately timing this on 10 year old hardware).
I wonder what 4 hours is on that timescale?
Or are you telling me that Windows is SO bad as an environment that (1) tools for this are not provided, and (2) the standard system editor supplied by the OS vendor is incapable of handling the problem, forcing me to use 3rd party tools for something this simple?
And, as usual, I don't care. Because I freelance by the hour.
We have companies that want me to do creative work (design and development of a software or hardware/software system).
Instead of being given a quiet area conducive to concentration, I am put into either a cubicle farm or a completely open area. With interruptions. With noise. I have even been put into a cubicle, shared with someone else, *and* three servers, *and* two automated CD pressing machines.
I have a noise-cancelling headset (takes care of low freq noise), and a lot of patience.
Still, I have a theory -- could not the environment be why F/OSS can be soooo much better sometimes?
I wonder what the working conditions at TATA are like?
Ok, stop there, I'm already laughing...
Now you FINALLY understand why Consumer Reports doesn't take advertising?
As I am sure will be pointed out by others -- READ THE EULA.
By clicking "I Accept", you give the copyright holder, or proxy, the right to the audit.
Using Windows? At least one Microsoft program had that clause in it...
Please...
The BSA does not accept a "proof of license" sticker. You MUST present a receipt for the software. If you go to a trade show, and accept a "free" copy of swagware (say, a copy of Windows) that does NOT have a receipt, you will be counted in violation.
Under THOSE rules, 8% is (translated) comletely in compliance.
[and yes, I do not accept that "free" software because it exposes my business to the BSA, and (reading the EULA fine print) at least one commercial software product put audits in! (possibly on my dime -- I threw the software out, because the terms were so bizarre)]
Sam Bronfman made his fortune by bootlegging whiskey from Canada to the US during Prohibition, dealing with the likes of Al Capone in Chicago. Edgar is his son, born in 1929.
The ISP is in control of the DNS server *and* the proxy server. Either a DNS redirect to the ISP can be done, or an HTTP redirect, thus changing the name of the site to the site approved by the CA. The man in the middle works from there. Most users won't notice, anyway (or care, given the number of trojans around). The "internet" works as usual...