None of that changes the fact that most software today is written for Windows...
I'm not so sure about this. A significant number of UNIX users constantly write new, say, perl programs. The total amount of code being generated there may be significant.
I'll agree that Windows is certainly dominant in the closed-source, horizontal market area -- but while horizontal markets produce a lot of sales, they don't necessarily mean that all that much code gets written.
From the Aegis webpage: Most sites using Aegis and Windows NT together do so by running Aegis on the Unix systems, but building and testing on the NT systems. The work areas and repository are accessed via Samba or NFS.
If HDTV adoption deadlines *don't* get pushed back like they did 80 bazillion times previously, I'll be amazed. We're in the middle of a recession, people aren't that interested in HDTV period, yadda, yadda, yadda.
define Thing:
Leg left = new Leg(length=0.85m);
Leg righ = new Leg(length=0.85m);
Corpus corpus = new Corpus(weight = 55kg);
Head head = new Head(haircolor = new Color(red));... end;
I don't see how this relates to text being well understood or not. I was referring to how people have good tools for looking through and processing text (like the UNIX text utils and perl and emacs), and there's broad familiarity with how to deal with problems that come up. There is broad support for it. Moving to UML (which, incidently, *still* does not have a universal interchange format, despite attempts) doesn't help at all on at least that front.
Could you name one tool?
grep. I'd go bonkers without it when doing development.
I mean... sure there ARE compilers.. I've heared there are compilers:-) I even have used one... and there are text editors... fine.
*shrug* I wasn't really referring to compilers, though there seem to be more implementations for any given language than there are for EUML -- that isn't really an issue, as long as the EUML implementations are good. The problem is with the transformation and analysis stuff.
But can you transform any text in any other format? Just simply with some easy scripting?...I'm not sure what you mean by "in any other format". "To any other format"...yes. I've used generated graphs to visualize control flow and interfile depdendencies, though it's true that I rarely translate text to a non-text format when developing. However, I don't see the problem there.
As for "easy scripting"...I think your terms are a bit loaded here. Scripting is, well, pretty powerful.
E.g., can YOU, not some hypotetical person, transform a C program into a Java program, with a tool, of course? In a reasonable amount of time?
Well...technically yes, using a VM, but I suppose the answer you're probably looking for is no. I'd ask...why would I want to do so, and why would EUML offer me anything in this field that an equivalently high-level text-based language could not?
Nope you cant. You need to write a compiler like tool for every sinlge transformation problem you may think off.
For C and Java, perhaps. C and Java are rather different from EUML or a very high level text-based language -- they're a real-world, here *now* (not in ten years when compilers get smart enough to make their use viable).
You can only ease that burden if you add your own coding style on top of the language structure you use... then you can use sed/perl or similar for scripting.
Oh, I see what you're saying. No -- I can just ram the thing through indent or another source code formatter.
Oops, since when?
Uh....the introduction of ASCII?
Ever heared about internationalisation?
Yup. Ever heard of all the software written in text-based languages that provides internationalization support?
Ever had a problem with attachments, uuencode, uudecode, splitting or concatanation?
Umm...not that I really remember. How is uuencoding going to be less of an issue for a non text-based language? The only way I can see it being relevant is if you're talking about encoding source, in which case (a) ASCII is 7-bit clean and all the programming languages I use are ASCII-based, (b) EUML, which could be stored in God-knows-what binary format, *is* going to run into uuencoding problems. As for splitting or concatanation, I don't see why there would be any more or less problems for ASCII than binary files.
Or simply end of line styles?
My dev tools and editors seem pretty happy with both CRLF and LF, though I haven't tried crossing over to a Mac.
I really think that data interchange is more of an issue with UML than ASCII.
Ever used text mode accidently by a binary ftp download?
Christ, the last time I manually specified ascii or binary for ftp transfers was, I think, on an ancient VAX.
Hm.... seems you do not get the point aboput UML at all if you get down to things like EMACS.
I must not.
Simple: a picture says more than 1000 words.
Well...a picture *can* have more content than a thousand words. But in the case of UML diagrams, there's hardly a pixel's worth of entropy per pixel. I don't stare at raw bitmaps to get an idea of what's going on.
Plus, as I pointed out earlier, there *are* times when it's convenient to graphically visualize parts of a software package. Text can be transformed into a picture for that brief visualization quite easily, though as I've said, I've only needed to do so a handful of times.
I'm pretty sure you realize the perfect mating partner in a fraction of a second by a simple look. He/she just looks perfect.... I'm pretty sure you are not even able to describe in TEXT what you are looking for in such a situation.
Sure -- but that's an area where the end data *is* stored and processed in essentially a visual format. If I want to deal with *images*, it's frusterating to transform images into text, and then back into images again. Code, however, is not natively an image.
The human mind can understand graphics/pictures/diagramms 10 to 100 times faster than text.
That's a ridiculous statement. There's no kind of reasonable metric that you can use to make a claim like that. If I want to convey what a river looks like from above, sure, it's a lot faster to show an image. If I want to make an argument about languages (as I am now), then text is far more effective. It's why Slashdot doesn't consist of a series of little pictograms.
And you can express everything you express with programming languages with graphics at least 3 times faster.
Again, I claim that you cannot provide reasonable grounds for making such a broad and essentially meaningless claim.
Unfortunatly "writing programs" in UML needs as much practice as writing porgrams in C.
If EUML provides the level of detail necessary to fully generate machine code, that's one thing. I haven't used EUML. However, I can decidedly say that UML is nowhere near that level of detail. You can generate the structure of a program, but you have nowhere near enough data to "write a program" in UML.
What I want to say is: use UML for a year. With the same enthusiasm you learned your most favorite language. As long as you did not do that you are simply not qualifying to give a statement about UML.
That is ridiculous, and an argument from authority -- an attempt to avoid having to defend your claims. I cannot simply leave for a year, study UML, then return to the argument. Instead of saying "I am experienced with UML (or, in this case, EUML), and I say it should replace text-based languages, so that *must be the case*", you should say "I am experienced with UML (or, in this case, EUML), and I say it should replace text-based languages, and *this is why*". So far, you've taken a number of stabs at text-based languages, many of which are definitely wrong, and then made an argument from authority. That's hardly convincing.
Do I think UML is a useful tool? Yes, in some circumstances. It's the only standard way of graphically modeling OO program structure. That's reasonable. It might even be useful for generating skeletons to work with. However, I find major fault with a few of the things it lacks. It does *not* have a implemented-by-all interchange language -- I cannot swap data between Visio, Rational Rose, and Dia. In its attempts to be completely abstract, it takes a least-common-denominator approach, and throws out a few language features that *I* find very important -- there is no standard way to model a typedef, for example, even though just about every language I can think of supports something like aliasing or typedefing. Now, that doesn't make UML useless -- just has a few warts. It doesn't appeal much to me from my stint with it. Granted, some of that may be because the UML tools I used are very poorly designed, which is hardly UML's fault. Creating a UML diagram in Visio is a painstaking, carpal-tunnel-inducing series of opening and closing dialogs and flipping through tabs. Dia is somewhat better, but still not perfect. So perhaps I'm a bit biased by my tools.
However, I'm very much dubious when you claim that EUML is the future. You've taken two ideas, and interwoven them -- and yet I feel that the two are very different. First, that very high-level languages are the way of the future. I agree with you, though I don't find them currently viable for real-world use. Second, that UML (or this EUML thing) is an ideal base to provide said high-level functionality. *That* I do not agree with. I have a hard time thinking of benefits that a UML-based system would provide that a (high level -- not like Java or C) text-based system would not provide, and I can definitely see drawbacks to the UML-based system (not good tools to analyze UML code, much UML-related software is priced more highly than I feel that it's worth, graphs tend to break down sooner than text when you're working with very large scale systems, it's easier to change text-based systems (at least with the present interface in Visio and Dia)). I can see making a high level text system that can be converted to UML's semi-standard XML representation, but not just using UML alone. I like to write with the keyboard, not the mouse.
P.S. before the anonymous coward who allways answers to my UML/OO posts is encouraged again: YES, I'm an UML/OO Consultant. You can hire me for $1000 a day plus hotel and traveling. In the last 4 years only one person(company) was not pleased with my job conduction... but that was guy who himself allready thought he was a UML expert, all other companies still hire me from time to time.
*shrug* I don't have a problem with that. If you really feel that something is a good idea, then it would seem a bit hypocritical if you *avoided* working in the field -- I'd expect a Java fan to use Java.
I just have somewhat different feelings about whether next gen languages and compilers should use UML or a derivation thereof.
However, at some point early in the life of this account, one the Slashdot editors, in a random and bitchy mood, permabanned me from M1 moderation. I've *never* been able to mod.
Now, however, the silver lining comes through. I am completely innocent of involvement with flooding Dave with stupid questions.:-)
I think the funniest part was the discussion of Jerry Pournelli's article in byte magazine, painting him as the archetypal PC user, illustrating why we didn't use Macs. I mean, if a computer always works, it just isn't fun:)
That would certanly explain why so many people use Linux on the desktop these days *ducks*...I dunno. Rather than an insult, I suspect that both Mac and Linux users would take that as a compliment.
You know, I know next to nothing about the Oscars, and have never really cared one way or the other who wins what, but the fact that the newest example you listed is from '97 stands as a pretty solid counterexample...
Come on guys, the National Security Agency is one of the good guys.
I know you intended to be sarcastic, and I generally think of Google as a "good company". However, they also have never fallen upon hard times. They're used for almost everything, and there are zero restrictions that I know of on corporate use of Google at any companies I can think of. How much do you think it would be worth to Acme Rubber (i.e. how much would they be willing to pay Google) to find out that FizBaz Rubber employees are searching for "Norwegian greenhouses"? Perhaps FizBaz is moving production from the Amazon to a bunch of greenhouses in Norway.
I started thinking about this a while ago -- Google (well, and other search engines, but Google is the most popular) is a tremendously large information leak to most companies.
It might be a good move for Google to open a "Corporate Program". Subscribers ensure that *no* data, not even aggregate data (well, perhaps barring some specific exceptions), is stored by Google for more than, say, a week, and it does not leave Google premises. It would make Google a lot of money, it would be a pretty obviously intelligent investment for companies that care about security...
pfft, back in my day, we could generate as many valid credit card numbers as we wanted. of course, those usually got used fraudulently....
I think the moral of the story is that CCs are *really* bad from an authentication point of view. For chrissake, the *number* is enough to let you bypass the thing.
A replacement (probably public key/smartcard) system would be a *much* better idea -- you'd have to physically steal a card to abuse it. No more grabbing a database or a recipt and having free rein.
There are only two drawbacks to this: first, there's a *huge* installed base of CC users and support, and second, anyone instituting it (VISA, whatever) is going to have to overcome temptation to try charging percentages of transactions (the reason we don't have e-cash now is because of overly greedy financial services companies who couldn't manage this).
The whole idea of using an application-specific editor seems bizarre to me. I use emacs for just about all editing, and while I can understand not using emacs programs to replace *all* programs (mutt knocks the socks off of vm), it'd be a very difficult argument to say that one should give up all the incredibly powerful features and familiarity with one's normal editor to using an application-specific one.
Okay, I had to look this one up. It seems that EUML is pretty much what it sounds like -- some companies trying to build a really high level language -- fine, that I'll understand -- using UML, which comes off as a bad idea. Frankly, the whole thing smacks of a "solution without a problem" being aimed at corporate purchasers who think "UML == good, so EUML == good".
First, I'm not a tremendous fan of UML. While the next generation of languages may hit within a decade, I'm not holding my breath, and I rather doubt that they'll be based on UML.
Second of all, I don't see why you dislike text so much. You can have the benefits of EUML without going to some mouse-based development environment. A text-based system would work fine as well.
Third, text is well understood by now. Text is stored in a standard format, and there are extremely powerful tools available to process it. Furthermore, it survives interchange fairly well, deals well with different output devices, and can be printed easily. There are a number of excellent, extensible text processing systems like emacs.
if, rather than instantly seeing that it's TCP-based, he spends countless hours pushing raw binary around, eventually coming up with something which, to TCP, is broken, then yes.
So his program is a bit of a hack, and he's not using::XML_foobar. That's *hardly* a reason to go after the author and claim that he was trying to "reverse engineer" XML or HTTP. Many, *many* programs that generate simple XML don't use libxml, and the same is even more true for HTTP. Hell, I'll bet that less than.1% of software that generates HTML uses any kind of a library to do it. There's no need for it.
I suppose you could say that he "reimplemented" a tiny part of HTTP, but he's certainly not reverse engineering it -- what he reverse engineered was the protocol that Streamium uses that sits on XML. To bash the author for this is attacking someone that just handed you a piece of software. If you don't like it, fine, write a "better" version that uses w3c-libwww and libxml. Don't try to insult him.
The Resynthasizer (linked to in parent post) is an incredibly cool research project that I looked at some time ago. It takes a massive number of CPU cycles, but the imagery it can generate is beyond impressive. Take a look, if you've never seen it before -- take a texture, and it can generate "new" texture based on that data. Have a 640x480 grass texture? It can make an 800x600 grass texture. You've seen images rerendered using a different palette...but what about rerendered using a different type of image -- text made out of grass, etc. *Very* cool.
Do you have any idea what you're talking about?
on
Hacking the Streamium
·
· Score: 1
Okay, "Lord Bitman", I'll assume that you're just sleepy and not stupid. We can wait for the next article about a new P2P app and for you to derisively claim that the author "only reverse-engineered TCP" to do that.
The platypus is closer to having a cold-blooded metabolism than most mammals, but it's still warm-blooded, contrary to what the story submissions claims.
So does IPv4 -- it's just that no one actually *uses* them.
The main thing that I *really* don't like about IPv6 is that, while it isn't a mandated part of the protocol, it seems that the overwhelming direction being pushed is to make the last 48 bits of your address your MAC address. Which *really* has nasty privacy implications -- 'slike a universal cookie, visible to everyone, that anyone can see (not just http servers).
There was some todo about patenting business models a while ago. I'm not sure whether it was accepted or not, but if it was, it might be interesting to patent the approach of "acquiring or registering obvious patents for the purpose of extorting money from other companies".
Then, once you own the patent, grant a free, one month license to any company getting sued by someone like NCR.
What's the benefit?
NCR (and similar companies) have little reason *not* to go after companies like this one. If they lose, the only cost to them is of legal fees.
Granted, a challenge to their patent can destroy it, but that costs the challenger money and risk, and doesn't cost NCR anything other than the patent. However, if you were able to sue them for patent infringement, they'd be looking at damages. The only way out of the patent infringement case would be for them to prove that their patent is non-obvious -- essentially that it's a legitimate patent, which they wouldn't be able to do.
The gameplay was sweet and it had some cool interactive music too, though unfortunately it only had two "themes", where Dune II had five or six at least.
Yup. And they had zero quality to work with. Sounded like 8 bit, 11 kHz sound or so, and it has one of the best "feels" of any game I've played.
And the sound of that mini-gun! One of the most amazing sounds ever to come out of a computer game. How come no other game mini-guns just tore shit out like that? It sounds like it's blasting several hundred bullets a minute
Yup. It evokes the phrase "tearing through things" wonderfully well.
Also, Syndicate *Wars* kind of sucked. I'd *love* to see a new Syndicate-style game (Syndicate seemed to be a pretty much one-game-genre, although there are vaguely similar games. Nothing quite mixes the strategic and tactical components with the incredibly gritty cyberpunk feel), and I'd buy it in a second.
The Syndicate engine, unfortunately, was not timer-based, so it stayed pretty jerky through the years, and never got ported to new machines.
A Syndicate-style game would make a pretty cool OSS project, because it takes a relatively small amount of art, and can be played with few system requirements.
Almost all the major DNS-related shifts in the last few years (addition of more TLDs being the most recent previous change) have been pushed by the registrars, and are designed not to drive technical improvement, but to sell more domains.
All that happened is that someone figured out that instead of just having a little checkbox for "also buy foo.net and foo.org" when you register "foo.com", "foo.edU" could be added into the mix.
Slashdot Mirror
None of that changes the fact that most software today is written for Windows...
I'm not so sure about this. A significant number of UNIX users constantly write new, say, perl programs. The total amount of code being generated there may be significant.
I'll agree that Windows is certainly dominant in the closed-source, horizontal market area -- but while horizontal markets produce a lot of sales, they don't necessarily mean that all that much code gets written.
From the Aegis webpage: Most sites using Aegis and Windows NT together do so by running Aegis on the Unix systems, but building and testing on the NT systems. The work areas and repository are accessed via Samba or NFS.
Yes, DVD is better quality, but except for videophiles, people don't care much about the negligible difference.
DVD looks *much* better to me -- whe comparing VHS on a TV to DVD on a computer screen. I'm hardly a videophile.
OTOH, buying a DVD player and hooking it up to a TV seems like a bit of a waste.
If HDTV adoption deadlines *don't* get pushed back like they did 80 bazillion times previously, I'll be amazed. We're in the middle of a recession, people aren't that interested in HDTV period, yadda, yadda, yadda.
define Thing: ...
... sure there ARE compilers .. I've heared there are compilers :-) I even have used one ... and there are text editors ... fine.
...I'm not sure what you mean by "in any other format". "To any other format"...yes. I've used generated graphs to visualize control flow and interfile depdendencies, though it's true that I rarely translate text to a non-text format when developing. However, I don't see the problem there.
... then you can use sed/perl or similar for scripting.
.... seems you do not get the point aboput UML at all if you get down to things like EMACS.
.... I'm pretty sure you are not even able to describe in TEXT what you are looking for in such a situation.
... but that was guy who himself allready thought he was a UML expert, all other companies still hire me from time to time.
Leg left = new Leg(length=0.85m);
Leg righ = new Leg(length=0.85m);
Corpus corpus = new Corpus(weight = 55kg);
Head head = new Head(haircolor = new Color(red));
end;
I don't see how this relates to text being well understood or not. I was referring to how people have good tools for looking through and processing text (like the UNIX text utils and perl and emacs), and there's broad familiarity with how to deal with problems that come up. There is broad support for it. Moving to UML (which, incidently, *still* does not have a universal interchange format, despite attempts) doesn't help at all on at least that front.
Could you name one tool?
grep. I'd go bonkers without it when doing development.
I mean
*shrug* I wasn't really referring to compilers, though there seem to be more implementations for any given language than there are for EUML -- that isn't really an issue, as long as the EUML implementations are good. The problem is with the transformation and analysis stuff.
But can you transform any text in any other format? Just simply with some easy scripting?
As for "easy scripting"...I think your terms are a bit loaded here. Scripting is, well, pretty powerful.
E.g., can YOU, not some hypotetical person, transform a C program into a Java program, with a tool, of course? In a reasonable amount of time?
Well...technically yes, using a VM, but I suppose the answer you're probably looking for is no. I'd ask...why would I want to do so, and why would EUML offer me anything in this field that an equivalently high-level text-based language could not?
Nope you cant. You need to write a compiler like tool for every sinlge transformation problem you may think off.
For C and Java, perhaps. C and Java are rather different from EUML or a very high level text-based language -- they're a real-world, here *now* (not in ten years when compilers get smart enough to make their use viable).
You can only ease that burden if you add your own coding style on top of the language structure you use
Oh, I see what you're saying. No -- I can just ram the thing through indent or another source code formatter.
Oops, since when?
Uh....the introduction of ASCII?
Ever heared about internationalisation?
Yup. Ever heard of all the software written in text-based languages that provides internationalization support?
Ever had a problem with attachments, uuencode, uudecode, splitting or concatanation?
Umm...not that I really remember. How is uuencoding going to be less of an issue for a non text-based language? The only way I can see it being relevant is if you're talking about encoding source, in which case (a) ASCII is 7-bit clean and all the programming languages I use are ASCII-based, (b) EUML, which could be stored in God-knows-what binary format, *is* going to run into uuencoding problems. As for splitting or concatanation, I don't see why there would be any more or less problems for ASCII than binary files.
Or simply end of line styles?
My dev tools and editors seem pretty happy with both CRLF and LF, though I haven't tried crossing over to a Mac.
I really think that data interchange is more of an issue with UML than ASCII.
Ever used text mode accidently by a binary ftp download?
Christ, the last time I manually specified ascii or binary for ftp transfers was, I think, on an ancient VAX.
Hm
I must not.
Simple: a picture says more than 1000 words.
Well...a picture *can* have more content than a thousand words. But in the case of UML diagrams, there's hardly a pixel's worth of entropy per pixel. I don't stare at raw bitmaps to get an idea of what's going on.
Plus, as I pointed out earlier, there *are* times when it's convenient to graphically visualize parts of a software package. Text can be transformed into a picture for that brief visualization quite easily, though as I've said, I've only needed to do so a handful of times.
I'm pretty sure you realize the perfect mating partner in a fraction of a second by a simple look. He/she just looks perfect
Sure -- but that's an area where the end data *is* stored and processed in essentially a visual format. If I want to deal with *images*, it's frusterating to transform images into text, and then back into images again. Code, however, is not natively an image.
The human mind can understand graphics/pictures/diagramms 10 to 100 times faster than text.
That's a ridiculous statement. There's no kind of reasonable metric that you can use to make a claim like that. If I want to convey what a river looks like from above, sure, it's a lot faster to show an image. If I want to make an argument about languages (as I am now), then text is far more effective. It's why Slashdot doesn't consist of a series of little pictograms.
And you can express everything you express with programming languages with graphics at least 3 times faster.
Again, I claim that you cannot provide reasonable grounds for making such a broad and essentially meaningless claim.
Unfortunatly "writing programs" in UML needs as much practice as writing porgrams in C.
If EUML provides the level of detail necessary to fully generate machine code, that's one thing. I haven't used EUML. However, I can decidedly say that UML is nowhere near that level of detail. You can generate the structure of a program, but you have nowhere near enough data to "write a program" in UML.
What I want to say is: use UML for a year. With the same enthusiasm you learned your most favorite language. As long as you did not do that you are simply not qualifying to give a statement about UML.
That is ridiculous, and an argument from authority -- an attempt to avoid having to defend your claims. I cannot simply leave for a year, study UML, then return to the argument. Instead of saying "I am experienced with UML (or, in this case, EUML), and I say it should replace text-based languages, so that *must be the case*", you should say "I am experienced with UML (or, in this case, EUML), and I say it should replace text-based languages, and *this is why*". So far, you've taken a number of stabs at text-based languages, many of which are definitely wrong, and then made an argument from authority. That's hardly convincing.
Do I think UML is a useful tool? Yes, in some circumstances. It's the only standard way of graphically modeling OO program structure. That's reasonable. It might even be useful for generating skeletons to work with. However, I find major fault with a few of the things it lacks. It does *not* have a implemented-by-all interchange language -- I cannot swap data between Visio, Rational Rose, and Dia. In its attempts to be completely abstract, it takes a least-common-denominator approach, and throws out a few language features that *I* find very important -- there is no standard way to model a typedef, for example, even though just about every language I can think of supports something like aliasing or typedefing. Now, that doesn't make UML useless -- just has a few warts. It doesn't appeal much to me from my stint with it. Granted, some of that may be because the UML tools I used are very poorly designed, which is hardly UML's fault. Creating a UML diagram in Visio is a painstaking, carpal-tunnel-inducing series of opening and closing dialogs and flipping through tabs. Dia is somewhat better, but still not perfect. So perhaps I'm a bit biased by my tools.
However, I'm very much dubious when you claim that EUML is the future. You've taken two ideas, and interwoven them -- and yet I feel that the two are very different. First, that very high-level languages are the way of the future. I agree with you, though I don't find them currently viable for real-world use. Second, that UML (or this EUML thing) is an ideal base to provide said high-level functionality. *That* I do not agree with. I have a hard time thinking of benefits that a UML-based system would provide that a (high level -- not like Java or C) text-based system would not provide, and I can definitely see drawbacks to the UML-based system (not good tools to analyze UML code, much UML-related software is priced more highly than I feel that it's worth, graphs tend to break down sooner than text when you're working with very large scale systems, it's easier to change text-based systems (at least with the present interface in Visio and Dia)). I can see making a high level text system that can be converted to UML's semi-standard XML representation, but not just using UML alone. I like to write with the keyboard, not the mouse.
P.S. before the anonymous coward who allways answers to my UML/OO posts is encouraged again: YES, I'm an UML/OO Consultant. You can hire me for $1000 a day plus hotel and traveling. In the last 4 years only one person(company) was not pleased with my job conduction
*shrug* I don't have a problem with that. If you really feel that something is a good idea, then it would seem a bit hypocritical if you *avoided* working in the field -- I'd expect a Java fan to use Java.
I just have somewhat different feelings about whether next gen languages and compilers should use UML or a derivation thereof.
I've been a mod. You've probably been a mod.
:-)
However, at some point early in the life of this account, one the Slashdot editors, in a random and bitchy mood, permabanned me from M1 moderation. I've *never* been able to mod.
Now, however, the silver lining comes through. I am completely innocent of involvement with flooding Dave with stupid questions.
I mean, how many people REALLY "Love" spaghetti?
It's a public good problem.
You can use less extreme words, but the people around you will sound more forceful. [shrug]
I think the funniest part was the discussion of Jerry Pournelli's article in byte magazine, painting him as the archetypal PC user, illustrating why we didn't use Macs. I mean, if a computer always works, it just isn't fun :)
...I dunno. Rather than an insult, I suspect that both Mac and Linux users would take that as a compliment.
That would certanly explain why so many people use Linux on the desktop these days *ducks*
You know, I know next to nothing about the Oscars, and have never really cared one way or the other who wins what, but the fact that the newest example you listed is from '97 stands as a pretty solid counterexample...
Come on guys, the National Security Agency is one of the good guys.
I know you intended to be sarcastic, and I generally think of Google as a "good company". However, they also have never fallen upon hard times. They're used for almost everything, and there are zero restrictions that I know of on corporate use of Google at any companies I can think of. How much do you think it would be worth to Acme Rubber (i.e. how much would they be willing to pay Google) to find out that FizBaz Rubber employees are searching for "Norwegian greenhouses"? Perhaps FizBaz is moving production from the Amazon to a bunch of greenhouses in Norway.
I started thinking about this a while ago -- Google (well, and other search engines, but Google is the most popular) is a tremendously large information leak to most companies.
It might be a good move for Google to open a "Corporate Program". Subscribers ensure that *no* data, not even aggregate data (well, perhaps barring some specific exceptions), is stored by Google for more than, say, a week, and it does not leave Google premises. It would make Google a lot of money, it would be a pretty obviously intelligent investment for companies that care about security...
pfft, back in my day, we could generate as many valid credit card numbers as we wanted. of course, those usually got used fraudulently....
I think the moral of the story is that CCs are *really* bad from an authentication point of view. For chrissake, the *number* is enough to let you bypass the thing.
A replacement (probably public key/smartcard) system would be a *much* better idea -- you'd have to physically steal a card to abuse it. No more grabbing a database or a recipt and having free rein.
There are only two drawbacks to this: first, there's a *huge* installed base of CC users and support, and second, anyone instituting it (VISA, whatever) is going to have to overcome temptation to try charging percentages of transactions (the reason we don't have e-cash now is because of overly greedy financial services companies who couldn't manage this).
The whole idea of using an application-specific editor seems bizarre to me. I use emacs for just about all editing, and while I can understand not using emacs programs to replace *all* programs (mutt knocks the socks off of vm), it'd be a very difficult argument to say that one should give up all the incredibly powerful features and familiarity with one's normal editor to using an application-specific one.
Okay, I had to look this one up. It seems that EUML is pretty much what it sounds like -- some companies trying to build a really high level language -- fine, that I'll understand -- using UML, which comes off as a bad idea. Frankly, the whole thing smacks of a "solution without a problem" being aimed at corporate purchasers who think "UML == good, so EUML == good".
First, I'm not a tremendous fan of UML. While the next generation of languages may hit within a decade, I'm not holding my breath, and I rather doubt that they'll be based on UML.
Second of all, I don't see why you dislike text so much. You can have the benefits of EUML without going to some mouse-based development environment. A text-based system would work fine as well.
Third, text is well understood by now. Text is stored in a standard format, and there are extremely powerful tools available to process it. Furthermore, it survives interchange fairly well, deals well with different output devices, and can be printed easily. There are a number of excellent, extensible text processing systems like emacs.
if, rather than instantly seeing that it's TCP-based, he spends countless hours pushing raw binary around, eventually coming up with something which, to TCP, is broken, then yes.
::XML_foobar. That's *hardly* a reason to go after the author and claim that he was trying to "reverse engineer" XML or HTTP. Many, *many* programs that generate simple XML don't use libxml, and the same is even more true for HTTP. Hell, I'll bet that less than .1% of software that generates HTML uses any kind of a library to do it. There's no need for it.
So his program is a bit of a hack, and he's not using
I suppose you could say that he "reimplemented" a tiny part of HTTP, but he's certainly not reverse engineering it -- what he reverse engineered was the protocol that Streamium uses that sits on XML. To bash the author for this is attacking someone that just handed you a piece of software. If you don't like it, fine, write a "better" version that uses w3c-libwww and libxml. Don't try to insult him.
*the* mission critical service
It's not that it's *unreliable*. It's "performance critical" that I was addressing, not "mission critical".
The Resynthasizer (linked to in parent post) is an incredibly cool research project that I looked at some time ago. It takes a massive number of CPU cycles, but the imagery it can generate is beyond impressive. Take a look, if you've never seen it before -- take a texture, and it can generate "new" texture based on that data. Have a 640x480 grass texture? It can make an 800x600 grass texture. You've seen images rerendered using a different palette...but what about rerendered using a different type of image -- text made out of grass, etc. *Very* cool.
Okay, "Lord Bitman", I'll assume that you're just sleepy and not stupid. We can wait for the next article about a new P2P app and for you to derisively claim that the author "only reverse-engineered TCP" to do that.
This way, your replication process can be over the slow SSH tunnel, but your transactions can be fast as they are local.
Fortunately, anything anyone is planning to do over LDAP (and particularly OpenLDAP) is extremely unlikely to be performance-critical.
The platypus is closer to having a cold-blooded metabolism than most mammals, but it's still warm-blooded, contrary to what the story submissions claims.
To me this reads: "I'm using an OS with a really shitty file management system and it's finally gotten to the point of really pissing me off."
Yes, this approach is well known about, and used by some people. It's also a pain in the ass for everyone else.
- ... supports QoS features.
So does IPv4 -- it's just that no one actually *uses* them.
The main thing that I *really* don't like about IPv6 is that, while it isn't a mandated part of the protocol, it seems that the overwhelming direction being pushed is to make the last 48 bits of your address your MAC address. Which *really* has nasty privacy implications -- 'slike a universal cookie, visible to everyone, that anyone can see (not just http servers).
There was some todo about patenting business models a while ago. I'm not sure whether it was accepted or not, but if it was, it might be interesting to patent the approach of "acquiring or registering obvious patents for the purpose of extorting money from other companies".
Then, once you own the patent, grant a free, one month license to any company getting sued by someone like NCR.
What's the benefit?
NCR (and similar companies) have little reason *not* to go after companies like this one. If they lose, the only cost to them is of legal fees.
Granted, a challenge to their patent can destroy it, but that costs the challenger money and risk, and doesn't cost NCR anything other than the patent. However, if you were able to sue them for patent infringement, they'd be looking at damages. The only way out of the patent infringement case would be for them to prove that their patent is non-obvious -- essentially that it's a legitimate patent, which they wouldn't be able to do.
Syndicate was AWESOME.
Yes. A thousand times yes.
The gameplay was sweet and it had some cool interactive music too, though unfortunately it only had two "themes", where Dune II had five or six at least.
Yup. And they had zero quality to work with. Sounded like 8 bit, 11 kHz sound or so, and it has one of the best "feels" of any game I've played.
And the sound of that mini-gun! One of the most amazing sounds ever to come out of a computer game. How come no other game mini-guns just tore shit out like that? It sounds like it's blasting several hundred bullets a minute
Yup. It evokes the phrase "tearing through things" wonderfully well.
Also, Syndicate *Wars* kind of sucked. I'd *love* to see a new Syndicate-style game (Syndicate seemed to be a pretty much one-game-genre, although there are vaguely similar games. Nothing quite mixes the strategic and tactical components with the incredibly gritty cyberpunk feel), and I'd buy it in a second.
The Syndicate engine, unfortunately, was not timer-based, so it stayed pretty jerky through the years, and never got ported to new machines.
A Syndicate-style game would make a pretty cool OSS project, because it takes a relatively small amount of art, and can be played with few system requirements.
Almost all the major DNS-related shifts in the last few years (addition of more TLDs being the most recent previous change) have been pushed by the registrars, and are designed not to drive technical improvement, but to sell more domains.
All that happened is that someone figured out that instead of just having a little checkbox for "also buy foo.net and foo.org" when you register "foo.com", "foo.edU" could be added into the mix.