I've been looking at eiffel recently. I need to sit down and code something in it to see whether it's really what I want and read more than the basic tutorial, but I'm facinated by some of the things it does.
It generates very fast code. Other than C/C++ and ML, it's right up there. No Java sluggishness.
It's got *smart* assertions. I'll bet the compiler can have tons of fun with optimization. Lots of room to make it blisteringly fast.
It's safe. No writing to random memory.
The templating system is to *die* for. It does what I wish ML could do. Basically, from my brief skimming of the thing, leave a type out on a parameter, and when you pass in different types it creates templated functions. Cool.
The biggest thing is pretty much the performance. You can use C/C++ (not safe languages, not all that designed around application development), ML (not great if you aren't one of those functional language wackos:-) ), or eiffel to get decent performance.
After that, the hit starts to become uncomfortable to me. A brief stint with QuickBASIC and later MATLAB left me hating slow languages -- it always ends up with you doing 10x the work to get the one bit of the program that must run quickly to zip along.
Uh...what else. The build system is integrated into the language, like SML. Definitely weird if you're used to GNU Make, but on the upside you don't have to use that godawful autoconf!
There's a free (fairly good) eiffel compiler, called (IIRC) SmallEiffel if you want to play with it. The one listed in the article is commercial.
[sigh]. Things were so much simpler in the Dark Ages. You and the EveryAuction guy and the other authors go out and beat the snot out of the guy who illegally grabbed all your code.
This push the FSF does to "give us ownership of your code so that you can get legal defense" is a little disturbing. In most cases they already have legal power to do revisions to your license (GPL v3). I'm more than a bit uncomfortable with how much IP power they're accumulating. Oh, it's fine now, but what about 50 years from now when Stallman's dead, and some big company manages to buy off a majority of the members in the thing? They already have more IP power than any other single organization I can think of, and it's growing rapidly.
The FSF is a single point of failure for the GPL. They're also the source of my single sense of unhappiness with it.
Linus feels the same way -- he's gone GPL v2 only, which eliminates their revision power as regards the Linux kernel.
I'm curious. Is licensing like water? That is, if I GPL a complete work, and then part of that code isn't copyrighted by me, is the GPLing still valid as regards the part that I do own?
I wonder if this has even ever been hammered out before. Probably, but it would take a lawyer to track down the case law.
It could be that the entire deal was bad the moment the guy tried to GPL the code he didn't own, so you'd want a revised agreement first.
Also, if he wrote the code under NDA from the propriatary people, they *might* be able to claim that his code exposes important knowledge about their own software that constitutes a trade secret (If someone wrote part of a driver for, say, Nvidia, I suspect this could be the case).
You know, if you aren't keeping it private for one reason or another, it might be interesting to know how you figured out that the guy was using GPLed software. Whether it was an inadvertent forum post (and he didn't understand the GPL) or whether you did an Xvid-style binary code comparison...
I'm also not a lawyer. However, at the very least I suspect this would have to be certified mail, and I strongly suspect that you cannot nullify someone's copyright by doing this.
I think what would happen if you tried this is that you'd piss off/scare the parties involved, they'd retain a lawyer, they'd start sending nasty notes and look at taking legal action with you. Which is what they should do.
Basically, it comes down to this. The original guy infringed the GPL. He needs to stop distributing the code. You may be able to sue him for damages for using your code without license. It doesn't mean that he can give away other code that he doesn't own, however.
If I steal the Mona Lisa and give it to you and the police can't find me, that doesn't mean that you can keep the painting openly. Just because I'm legally liable for all sorts of crimes does *not* mean that you then have the right to do whatever you want with the Mona Lisa. If you attempt to, I dunno, chop up the Mona Lisa *knowing* that it's stolen, *you* will be liable for that crime (though I'm still liable for all the things I've done).
Please do not start this. Once a couple of nationalistic types start going on this, it can go on forever, will get bitter, and won't accomplish anything.
I dunno. That was my initial thought too, but how neat would it be to know that your significant other is soliciting opinions from 250k other people to try to get you the perfect Christmas gift?
Actually, how neat would it be to be married to someone that reads tech forums, period?:-)
That's gotta suck based just on the Mac OS X pixmaps. People I know that use VNC regularly use a solid color background, solid color widgets, solid color titlebars to keep things light and usable...all that glitzy translucency and racing strips across everything wreak havoc on attempts to do an efficient pixel-pushing system.
And, frusteratingly, I hear calls for X to do the same. An irritating but vocal minority keeps yelling that X is bloated (it's not, and RAM usage requirements are because of server-side pixmap storage), slow (it's not, and if you wanted it to it could buffer all client windows to get "smooth dragging" if you wanted to blow huge quantities of RAM like Quartz does), and complicated (because they aren't using the straightforward, reliable front ends to set it up). A lot of this gets blamed on the remote capabilities of X, and I keep hearing calls for a windowing system using a local framebuffer. Ick.
It's good to see that some people *took* this step and regretted it. I would hope that X users can *learn* from this.
Re:Mesa is not hardware accelerated
on
Mesa 5.0 Released
·
· Score: 4, Informative
Mesa is not hardware accelerated
You mean Mesa's software driver is not hardware accelerated. Take a look at the Mesa FAQ, point 1.2.
Yes. Specifically, Mesa serves as the OpenGL core for the XFree86/DRI OpenGL drivers. ***
Now, Mesa does *include* a software rendering engine...
This could be what you meant, but this is the first post along the "Mesa is nothing more than a software renderer" lines, and there are a *lot* on here, some of which are definitely wrong.
Re:I don't know what they did
on
Mesa 5.0 Released
·
· Score: 3, Informative
Umm...that's *bad*. glxgears is a benchmark, and *should* be eating everything.
Which means that your GPU now isn't accepting above some level for some reason.
This guy is complaining about what the MVP designation means...uh, no kidding it's politicized and inaccurate. It's a *marketing tool*. It has nearly zero real world utility -- it just gets MS free tech support workers. Hell, other companies should do this. "Maytag MVP".
Doing tech/dev assistance has always been amorphous. You earned the respect of the community through helping alone. Trying to control the process by controlling the database that marks people as "helpful" or not is doomed to failure.
Because it makes no *sense* at all for MS to be losing money in every product except for two. MS is enormous, and would have been cutting things to save money well before this. Dropping some personnel. It might be possible if MS was really into takin gambles *and* all their other markets were expanding markets, but peripherals? A bunch of their software markets, like SQL Server, or Exchange? Those are stable, mature markets.
I suppose "bullshit" was a little strong, and it wasn't really the original poster's fault. It just blew me away that almost the entire body of comments immediately seized upon MS massively exploiting monopoly powers, and not the much more obvious and plausible accounting fudging.
Microsoft is leveraging their high prices to enable them to give away other products, thus undercutting their competition.
Without doing research, I can pretty reasonably put this in two words -- "bull" and "shit".
I'll be willing to be that two years ago, three years ago, all those markets were reported as profitable. And it isn't because of a "tech downturn" that dropped *everything* into red ink without managers doing any cost saving. No, you'd hear about divisions being cut, layoffs, everything if there were real losses.
It's pretty obvious what's going on. MS is making money, just as usual. A while ago, a big company went belly up because of "loss hiding" -- our old friend Enron. As a result of this, lots of laws were passed making executives and auditors legally liable for hiding losses, inflating profits, and tucking them into future good years. Perhaps more importantly, the current public opinion is to crucify execs doing this, and not to let the government let them off the hook easily.
What's happened is that our buddy MS has, like most large companies over the past few years, has been tucking away a few too many losses under the rug and artificially jacked up reported profits.
Now, all of a sudden, Bill G. and Co. could be doing hard jail time (to say nothing of their auditing firm) if they can be shown to be deliberately hiding losses for another year. So they want to get rid of their losses *now*. It can't wait for another year -- they have to show all those unreported losses and inflated profit immediately. Well, they can't say that Windows is losing money -- 2k to XP migration is critical right now, Linux is a threat, and looking less than stable would be an awful idea. They can't say that Office is losing money -- for the first time in years, competitors have just sprung up, including Open Office and even WordPerfect pulled a comeback. The Office product also has to be rock solid. So where are all those losses going? Right into these non-core markets. Everything else loses money to clear up the balance sheets.
This isn't just MS, either. You're going to see a *lot* of big companies doing this, and a *lot* of negative filings, as companies have to avoid giving away past reporting falsehoods.
Now, I haven't looked at their past sheets. If this is consistent with past filings, I'm wrong. But I'd quite confidently bet that I'm not.
The FastTrack protocol is vastly superior to Gnutella, especially the original Gnutella.
Okay. What grounds for your claim?
DirectConnect interface non-idiotic?!?
You'll notice that I was comparing it to the slapped-together, script-running IRC+DCC combination, usually done via fservs, which are quite idiotic compared to DirectConnect.
You're actively running a server that's taking requests for the file and handing them out. More like having a mounted Polaroid camera in front of the picture in your window with a sign and instructions.
Re:A *real* anti-leech/anti attacker system propos
on
Gnutella2?
·
· Score: 2
The second issue is taken care of -- if HoneyPot isscrews me over, I zero his trust.
The transitive trust is a good point. A is legitimate, but A trusts B, and B is bogus, and A (deliberately or otherwise) does not update his local trust list upon B going bad. Hm.
Okay, here's a possible modification. Allow negative trust values. If I'm C, I trust A, user A trusts user B, and I consider user B a "bad guy" (negative trust), reduce C's trust of user A periodically.
Re:200 entry trust list?
on
Gnutella2?
·
· Score: 2
You can partially automate this -- on a download, give a point of trust automatically, say. If the user says "this person screwed me", they can block the person from showing up in future results, and in doing so, set their trust to 0.
Slashdot Mirror
I've been looking at eiffel recently. I need to sit down and code something in it to see whether it's really what I want and read more than the basic tutorial, but I'm facinated by some of the things it does.
:-) ), or eiffel to get decent performance.
It generates very fast code. Other than C/C++ and ML, it's right up there. No Java sluggishness.
It's got *smart* assertions. I'll bet the compiler can have tons of fun with optimization. Lots of room to make it blisteringly fast.
It's safe. No writing to random memory.
The templating system is to *die* for. It does what I wish ML could do. Basically, from my brief skimming of the thing, leave a type out on a parameter, and when you pass in different types it creates templated functions. Cool.
The biggest thing is pretty much the performance. You can use C/C++ (not safe languages, not all that designed around application development), ML (not great if you aren't one of those functional language wackos
After that, the hit starts to become uncomfortable to me. A brief stint with QuickBASIC and later MATLAB left me hating slow languages -- it always ends up with you doing 10x the work to get the one bit of the program that must run quickly to zip along.
Uh...what else. The build system is integrated into the language, like SML. Definitely weird if you're used to GNU Make, but on the upside you don't have to use that godawful autoconf!
There's a free (fairly good) eiffel compiler, called (IIRC) SmallEiffel if you want to play with it. The one listed in the article is commercial.
[sigh]. Things were so much simpler in the Dark Ages. You and the EveryAuction guy and the other authors go out and beat the snot out of the guy who illegally grabbed all your code.
This push the FSF does to "give us ownership of your code so that you can get legal defense" is a little disturbing. In most cases they already have legal power to do revisions to your license (GPL v3). I'm more than a bit uncomfortable with how much IP power they're accumulating. Oh, it's fine now, but what about 50 years from now when Stallman's dead, and some big company manages to buy off a majority of the members in the thing? They already have more IP power than any other single organization I can think of, and it's growing rapidly.
The FSF is a single point of failure for the GPL. They're also the source of my single sense of unhappiness with it.
Linus feels the same way -- he's gone GPL v2 only, which eliminates their revision power as regards the Linux kernel.
I'm curious. Is licensing like water? That is, if I GPL a complete work, and then part of that code isn't copyrighted by me, is the GPLing still valid as regards the part that I do own?
I wonder if this has even ever been hammered out before. Probably, but it would take a lawyer to track down the case law.
It could be that the entire deal was bad the moment the guy tried to GPL the code he didn't own, so you'd want a revised agreement first.
Also, if he wrote the code under NDA from the propriatary people, they *might* be able to claim that his code exposes important knowledge about their own software that constitutes a trade secret (If someone wrote part of a driver for, say, Nvidia, I suspect this could be the case).
You know, if you aren't keeping it private for one reason or another, it might be interesting to know how you figured out that the guy was using GPLed software. Whether it was an inadvertent forum post (and he didn't understand the GPL) or whether you did an Xvid-style binary code comparison...
I'm also not a lawyer. However, at the very least I suspect this would have to be certified mail, and I strongly suspect that you cannot nullify someone's copyright by doing this.
I think what would happen if you tried this is that you'd piss off/scare the parties involved, they'd retain a lawyer, they'd start sending nasty notes and look at taking legal action with you. Which is what they should do.
Basically, it comes down to this. The original guy infringed the GPL. He needs to stop distributing the code. You may be able to sue him for damages for using your code without license. It doesn't mean that he can give away other code that he doesn't own, however.
Um...no.
If I steal the Mona Lisa and give it to you and the police can't find me, that doesn't mean that you can keep the painting openly. Just because I'm legally liable for all sorts of crimes does *not* mean that you then have the right to do whatever you want with the Mona Lisa. If you attempt to, I dunno, chop up the Mona Lisa *knowing* that it's stolen, *you* will be liable for that crime (though I'm still liable for all the things I've done).
In this case, you are definitely incorrect.
Please do not start this. Once a couple of nationalistic types start going on this, it can go on forever, will get bitter, and won't accomplish anything.
I dunno. That was my initial thought too, but how neat would it be to know that your significant other is soliciting opinions from 250k other people to try to get you the perfect Christmas gift?
:-)
Actually, how neat would it be to be married to someone that reads tech forums, period?
It doesn't let you drag windows across desktops, though...just stops when you hit the edge, IIRC.
the bitmap of the screen
That's gotta suck based just on the Mac OS X pixmaps. People I know that use VNC regularly use a solid color background, solid color widgets, solid color titlebars to keep things light and usable...all that glitzy translucency and racing strips across everything wreak havoc on attempts to do an efficient pixel-pushing system.
the non-remoteness of the Quartz APIs
And, frusteratingly, I hear calls for X to do the same. An irritating but vocal minority keeps yelling that X is bloated (it's not, and RAM usage requirements are because of server-side pixmap storage), slow (it's not, and if you wanted it to it could buffer all client windows to get "smooth dragging" if you wanted to blow huge quantities of RAM like Quartz does), and complicated (because they aren't using the straightforward, reliable front ends to set it up). A lot of this gets blamed on the remote capabilities of X, and I keep hearing calls for a windowing system using a local framebuffer. Ick.
It's good to see that some people *took* this step and regretted it. I would hope that X users can *learn* from this.
Mesa is not hardware accelerated
You mean Mesa's software driver is not hardware accelerated. Take a look at the Mesa FAQ, point 1.2.
Let me point readers to the Mesa FAQ:
***
1.2 Does Mesa support/use graphics hardware?
Yes. Specifically, Mesa serves as the OpenGL core for the XFree86/DRI OpenGL drivers.
***
Now, Mesa does *include* a software rendering engine...
This could be what you meant, but this is the first post along the "Mesa is nothing more than a software renderer" lines, and there are a *lot* on here, some of which are definitely wrong.
Umm...that's *bad*. glxgears is a benchmark, and *should* be eating everything.
Which means that your GPU now isn't accepting above some level for some reason.
This guy is complaining about what the MVP designation means...uh, no kidding it's politicized and inaccurate. It's a *marketing tool*. It has nearly zero real world utility -- it just gets MS free tech support workers. Hell, other companies should do this. "Maytag MVP".
Doing tech/dev assistance has always been amorphous. You earned the respect of the community through helping alone. Trying to control the process by controlling the database that marks people as "helpful" or not is doomed to failure.
Because it makes no *sense* at all for MS to be losing money in every product except for two. MS is enormous, and would have been cutting things to save money well before this. Dropping some personnel. It might be possible if MS was really into takin gambles *and* all their other markets were expanding markets, but peripherals? A bunch of their software markets, like SQL Server, or Exchange? Those are stable, mature markets.
I suppose "bullshit" was a little strong, and it wasn't really the original poster's fault. It just blew me away that almost the entire body of comments immediately seized upon MS massively exploiting monopoly powers, and not the much more obvious and plausible accounting fudging.
Microsoft is leveraging their high prices to enable them to give away other products, thus undercutting their competition.
Without doing research, I can pretty reasonably put this in two words -- "bull" and "shit".
I'll be willing to be that two years ago, three years ago, all those markets were reported as profitable. And it isn't because of a "tech downturn" that dropped *everything* into red ink without managers doing any cost saving. No, you'd hear about divisions being cut, layoffs, everything if there were real losses.
It's pretty obvious what's going on. MS is making money, just as usual. A while ago, a big company went belly up because of "loss hiding" -- our old friend Enron. As a result of this, lots of laws were passed making executives and auditors legally liable for hiding losses, inflating profits, and tucking them into future good years. Perhaps more importantly, the current public opinion is to crucify execs doing this, and not to let the government let them off the hook easily.
What's happened is that our buddy MS has, like most large companies over the past few years, has been tucking away a few too many losses under the rug and artificially jacked up reported profits.
Now, all of a sudden, Bill G. and Co. could be doing hard jail time (to say nothing of their auditing firm) if they can be shown to be deliberately hiding losses for another year. So they want to get rid of their losses *now*. It can't wait for another year -- they have to show all those unreported losses and inflated profit immediately. Well, they can't say that Windows is losing money -- 2k to XP migration is critical right now, Linux is a threat, and looking less than stable would be an awful idea. They can't say that Office is losing money -- for the first time in years, competitors have just sprung up, including Open Office and even WordPerfect pulled a comeback. The Office product also has to be rock solid. So where are all those losses going? Right into these non-core markets. Everything else loses money to clear up the balance sheets.
This isn't just MS, either. You're going to see a *lot* of big companies doing this, and a *lot* of negative filings, as companies have to avoid giving away past reporting falsehoods.
Now, I haven't looked at their past sheets. If this is consistent with past filings, I'm wrong. But I'd quite confidently bet that I'm not.
Also, I'm sure that MS could charge less than $45 and still make a profit - since they'd sell more copies.
I never thought I'd see the dot com fallacy in use in real life.
The FastTrack protocol is vastly superior to Gnutella, especially the original Gnutella.
Okay. What grounds for your claim?
DirectConnect interface non-idiotic?!?
You'll notice that I was comparing it to the slapped-together, script-running IRC+DCC combination, usually done via fservs, which are quite idiotic compared to DirectConnect.
Yes, I know. LimeWire's additions to the Gnutella protocol put it on par with FastTrack by adding superpeers.
I'm guessing that "SlashFemale" just sounds less snazzy.
You're actively running a server that's taking requests for the file and handing them out. More like having a mounted Polaroid camera in front of the picture in your window with a sign and instructions.
The second issue is taken care of -- if HoneyPot isscrews me over, I zero his trust.
The transitive trust is a good point. A is legitimate, but A trusts B, and B is bogus, and A (deliberately or otherwise) does not update his local trust list upon B going bad. Hm.
Okay, here's a possible modification. Allow negative trust values. If I'm C, I trust A, user A trusts user B, and I consider user B a "bad guy" (negative trust), reduce C's trust of user A periodically.
You can partially automate this -- on a download, give a point of trust automatically, say. If the user says "this person screwed me", they can block the person from showing up in future results, and in doing so, set their trust to 0.