You will find that many off-the-shelf devices, like NAT/Routers from Linksys, Netgear, etc. use 192.168.x.x by default; some of them don't let you use anything else (I think Linksys locks you in to 192.168, but you can change the lower two octets).
Just FYI, but on my LinkSys BEFSR41 (firmware rev 1.44.2), all four octets are configurable.
Mind you, with only five hosts at the moment, I haven't bothered to change it out of the 192.168.x.x address space.
PS2 are probably mostly dialup. Broadband is the future; if you want good gaming without lag, you need broadband. And if you want a broadband platform where all your opponents have broadband, you must choose Xbox.
Ever play Frequency or Amplitude online? Both of these games support a mix of broadband and dialup users, and I have never seen an issue playing with people with different connection types.
Sony was smart. They made it possible for a game to specifically function with only one connection type. Take SOCOM for example: it doesn't support the dialup element of the adapter, because it just won't work.
Sony made it so the game developers can decide wether or not to target just broadband users, or everyone. There are a lot of styles of games that work quite well on a dialup connection.
And in this big old world, there are still lots of places where telephone lines are readily available, but where broadband is not. Personally, I applaud a company that is willing to embrace players in remote (or odd) areas where broadband simply isn't an option.
I own an Xbox and I used to own a ps2. I broke it and took it back to best buy under my service plan and got an xbox. It is sooooo much better as a gaming platform and the online gaming is superb. On the ps2 you get to pay a monthly fee for each game you play unless the publisher is nice enough to give you a flat rate for all the games from that publisher/cartel(sony).
Feel glad I just used up the rest of my modpoints. Otherwise, I would have rated this -1: Troll.
It's blatently false. Out of the (quoted) 50 PS2 titles that will support online gaming by the end of the year, only TWO will charge any sort of fee: FFXI, and Everquest Online Adventures. This is as opposed to the Xbox, where you can't play _any_ games without paying a monthly fee to Microsoft.
Oh and just to add...the ps2 only supports 5.1 in cutscenes whereas the xbox supports it ingame
This is also false. A number of PS2 titles support 5.1 surround via DTS encoding, wheras the Xbox only supports Dolby Digital encoding. Grand Theft Auto: Vice City, SSX: Tricky, and NHL 2003 are three titles that support in-game DTS surround.
For someone who "used to own a PS2", you certainly don't know much about it, do you?
The only thing responsible for SCO's bad press in SCO acting like a corporate idiot. They keep making statements and arguments that are contrary to both logic and law. Any wonder why even the inept computer press is beating up on them?
They hardly need IBM's help in making them look bad. They do a great job all by themselves. IBM's best strategy so far has been to keep quiet, and watch SCO implode on their own FUD.
Microsoft takes the 20-year-old PC architecture, puts it in a PC-sized case, designs it around a Windows-based core using common PC components, then goes around trying to get PC game developers to write titles for their console -- and they expect the _games_ to be original???
Microsoft designed the most un-original gaming console ever. They wound up with what many people predicted -- a completely un-inspired software library that looks just like your typical PC game (without getting any of the actual worthwhile PC games). Can anyone truly say they're suprised at the state of the Xbox's library?
You could run NT's other original filesystem, HPFS. Linux has decent HPFS support available, the allocation unit is a 512b sector, and it's organized for fast searches and minimal fragmentation. It can also format up to 64GB partitions (although it still has a 2GB filesize limit).
The trick is to get HPFS support for Windows. To do this, you need to get the driver files from Windows NT v3.x (something that, admittedly I'm not sure works with Windows versions > NT 4. I don't do Windows personally, so I haven't tested it -- like I said above, FWIW). That will give you two of the three OSs supported. HPFS has been around for a while, (circa 1988), so you might be able to find something from the Intel FreeBSD world you could port to OS X.
I use HPFS for my 100MB ZIP disks (which I admittedly rarely use anymore for anything than quick archival purposes). It's not journalled, but it uses a bidirectional sector pointer system, so chaining errors are amost always fixable. The big downside is that if the filesystem is dirty, checking it can take a huge amount of time.
It's probably not a practical solution (I didn't and won't claim it is), but it's still a slightly more constructive answer than "Your stuck with FAT32":).
It features a web server and Java applet for tune addition (along with other methods) - looks extremely cool.
If it doesn't act like a hard drive through standard protocols (NFS, NetBIOS/NetBEUI), then IMO it looks quite a bit less cool.
I want to be able to plug it in and use it as a big old hard drive. In that way it could serve as not only a player, but when connected as the core MP3/Ogg repository on my home network. There would no longer be any need to keep seperate copies of such things on the actual computers themselves -- if I'm ripping CDs, I could use whatever encoding software I want to and let it write directly to the mount point. If I want to playback on my PC, just plug in the player, mount it, and stream the data across the network.
IMO, that's quite a bit more cool than having to go through a web interface to move files back and forth.
Okay -- admittedly, ethernet support is a very good thing, but IMO that just means it has a useful interconnect system available.
The more important question IMO is what protocols does it support over the ethernet connection? NetBIOS/NetBEUI? TCP/IP? Some custom protocol? If TCP/IP, does it support NFS? FTP? NetBIOS over TCP/IP? Something else?
The poorly designed/formatted website doesn't give much information in this regard. I'm assuming it's TCP/IP based, but even then, you need something that will allow you to transfer files to it.
Being NFS based would be ideal here. Being able to mount it from any of my OS/2 or Linux boxes as a drive would be perfect. NetBIOS/NetBEUI/etc. would be of secondary usefulness -- I don't run any of these protocols (as I don't need to -- they add overhead and don't do anything NFS and lpr/lpd don't), but have them available on all my systems. If they're FTP based, well at least NetDrive allows me to treat them as filesystem mounts. If they're using some "custom" protocol that only works with their Windows-based client software, then no thanks.
I don't completely agree. Sure, CPU load doesn't matter in a single-tasking system, but if I'm running a multi-tasking system I don't want it to noticably "hiccup" each and every time some background task has to do some disk I/O. The system should _not_ bog down just because Mozilla is cleaning out its cache -- it should still be responsive.
Obviously, priority levels can have a big effect on this sort of thing -- but in a multitasking environment where you are utilizing most of the CPU with various tasks, having high CPU usage for I/O can be detremental. If it's sufficiently high enough, the system will start to feel like Windows 3.1, where everything stoped until I/O is complete.
I've been tempted to get back into teaching classes such as this. One of my previous employers, an ISP, held such sessions for its customers to teach all sorts of interesting things, and they were generally well attended and well received.
To get people to show up, however, you have to teach them stuff they're not going to just be able to fumble their way through on their own (or with their neighbours kids help:) ). So, some of the ideas I can think up off the top of my head include:
Home Networking. I've been known to answer some home networking questions on some of the PlayStation 2 newsgroups, and you'd be amazed by the number of e-mails I get from people saying "I've just bought my son a PS2 network adapter, but don't know how to connect it to my existing cable/DSL modem". Alot of basic computer users out there are afraid of things we find simple like home routers, even though they're pretty much plug-and-play. Just having a course to cover the simple stuff like "how to select a router" and "how to connect everything" would probably help a whole lot of people. A more "advanced" course that showed people how to safely setup file and print sharing would probably also be well attended.
Setting up a Website. Lots of people tend to be interested in things like this as well, but don't know where to start. Even if you just show them a common page creation utility, they'll still need to know how to upload the content to their web space provider.
How to use search engines more effectively. How many of us here have a friend/relative that can never find things they're looking for online, even through Google? Probably most of us. I routinely get calls from family members who don't know how to narrow their searches to ensure they get more relevent information, and who ask me to find things for them. Things like "I wanted to look for a good local restaurant, so I typed in 'restaurants' and got results from halfway around the world!". Most basic users I know don't even know the basics of how to use a search engine effectively.
Playing with Linux. You might be suprised. While most people do still equate "computer" with "Microsoft Windows", there are people out there who have heard of Linux, and are sufficiently interested to see what it looks like. Such people probably don't know anyone running it, and thus haven't been exposed to it, but would be willing to fiddle with it in a "safe" environment (ie: where they don't have to install anything onto their computers and risk losing all their data;) ). A "course" where people could play around with Linux and some of its applications might just be popular. You could even hand out Knoppix CD's as a part of the course sign-up fee (assuming there is one, of course). People will sign up for anything that hands out free programs they can take home and use;).
Online Console Gaming. Okay -- similar to "Home Networking" above, but in this case you could specifically offer a course on how to hook up a PlayStation 2 or Xbox to the Internet. Lots of kids have these devices and ask their parents for the necessary online kits to play with their friends on the 'net, but lots of parents wouldn't know where to start in hooking such a system up.
Tutoring. If you're in an area with a University, Community College, or even a high school that teaches programming courses, you can always offer a tutoring service for students.
Home Internet Security. Lots of people out there are either afraid of hackers, or too ignorant to be afraid. There is a lot of need for courses like this, although you have to walk a fine line when advertising them that you don't wind up scare-mongering.
I was there on a foggy day. There wasn't much fog; in fact, I had no idea that it was foggy until I got up into the tower. Then I couldn't see the ground anywhere. Disappointing.
If you're not queasy, the best time to go is when it's windy -- the tower is designed to sway in the wind, and the top can transcribe a 12m circle:).
Naturally, the size of Canada dwindles roughly to that of a medium pizza if you discount all the uninhabated/under ice portions.
Yeah, but you still can't readily backpack across it, which was really my point (and no, I didn't miss your attempt at humour:) ). Unless, of course, you really enjoy hiking through uninhabited and iced-over areas (which some people quite enjoy). Personally, I prefer to live in a country that is dominated by wide-open spaces, as opposed to one that is dominated by other human bodies:).
On a serious note though, I find it interesting that people seem to be pointing out a good bit of cultural activities and sites as well as 'geek' ones.
Well, I guess one of the difficult things is that the OP didn't really mention what type of geek they are. A history geek would probably be more interested in a fort from the War of 1812 than a science centre, which might be heaven for a science geek.
Of course, some things transcend the different stratii of geekness, and might even, parry the thought -- appeal to NON-GEEKS!:)
Linux's JFS support doesn't come from AIX -- it comes from OS/2, to which SCO holds no copyright nor any other IP rights.
The OS/2 implementation of JFS was a ground-up rewrite, based on the JFS specification, which is owned completely by IBM. SCO might claim that they have copyrights to the original AIX JFS sources (a dubious claim), but they can't claim they own the JFS _specification_. That's owned by IBM.
As mentioned above, the OS/2 implementation was a ground-up rewrite based on this spec. The OS/2 version of the code was then ported and integrated into Linux by IBM.
Thus I can't see how SCO could have any sort of claim on JFS in Linux. SCO has no contracts with IBM pertaining to OS/2 technologies.
Let me be the first to welcome you to Canada (considering you're not here yet, I _assume_ I'm the first at least:) ).
First things first. Canada is a REALLY BIG PLACE. You do not backpack across Canada. I know that Australia is a big place (a whole continent in fact...), and the US has a decent size, but Canada is in a whole different ballpark. Think of Australia. Now think of another 1/5 of Australia. Stick them together, and you get a bit closer to Canada's size. Canada is nearly 10/million/ square kilometres of land, sprawling across 7 seperate time zones. It's a big place to walk across:).
As such, a good geek travel system to your trip would probably to take the train from coast to coast, getting off in major cities of interest.
Once you've figured out how to get around, where to go? Some good suggestions include (in no particular order, and probably leaving out all sorts of funky places in between...):
The CN Tower in Toronto, Ontario. Take the trip up to Sky Pod, which is the highest public observation deck in the world, at 447m (1465'). Or head down to the first observation deck and stand on the glass floor hanging over, well, nothing but air at 342m (1122') up.
The Royal Ontario Museum in Toronto, Ontario. The ROM is undergoing some renovations at the moment, but it still has some extremely impressive collections.
Canadian Aviation Museum in Ottawa, Ontario. Go see the parts of the Avro Arrow that they've been able to recover, and shed a tear for the greatest interceptor ever built, that never even got to fly a single mission (the Canadian government decided to scrap it after 5 test models were built to instead invest in BOMARC missles. The museum also has one of these. The BOMARCs were bought from the US, and 95% of them were useless. Not a single one was ever used. The engineering team that designed the Arrow mostly made their way to NASA in the US, and were instrumental in the first manned space missions).
Ontario Science Centre, Toronto, Ontario. The Science Centre has an Omnimax theatre which, if you've never seen one, is something you have to experience. Imagine a bowl 6 storys tall sitting in front of you, on a 30deg angle, with a high-definition 70mm film being projected onto it via a fish eye lens. Now imagine that the screen is filled with millions of tiny holes, and the speakers are mounted directly behind it. There are always good geek films playing -- I particularily recently enjoyed the one about the International Space Station. While you're at the OSC, they have a rather large (and free) Internet Cafe-type area setup, so you can get online and check out Slashdot:).
Maritime Museum in Halifax, Nova Scotia. If you're a sea geek, be sure to stop here. In the harbour just outside you can tour a pair of decomissioned Canadian warships.
Canadian Museum of Civilization, Hull (Quebec). Not only a fantastic museum of human civilization in Canada, but if you missed seeing the Omnimax theatre in Toronto, they have a combination Imax/Omnimax screen (although the OSC Omnimax system in Toronto is much better IMO).
Dinosaur Provincial Park, Alberta. I've always wanted to visit this place, but haven't made it yet. New species of dinosaurs are discovered here every year. It's so chock-full of dinosaur remains that it's a UNESCO World Heritage Site. Some of the most impressive T-Rex skeletons ever found were found here.
Lake Superiour Provincial Park, Ontario. A bit of a trek to make it up here, but if you do, not only can you sit on the shores of the world's largest freshwater lake, but you can check out Agawa Rock, where hundreds of years ago the native Ojibwa painted pictographs on the cliffs at the waters edge. If this is up your alley, see them now -- they're already quite faded, and are expected to disappear due to weathering within the next 50 years.
This was (and is) actually quite a cool game, and it's just about to make a big comeback.
However, when it originally came out back in 1993, it was published for OS/2 2.1 by a company called "Advanced Idea Machines", and it had the absolute _worst_ box art of anything ever produced in the history of the world.
I have the box, and it's UGLY ugly. I've seen boxes that I'd call ugly, fugly, pug ugly, and even pug fugly -- but this is the only one I'd ever think of calling UGLY ugly. It truly is that bad:).
I wish I had a flatbed scanner so everyone else here could see how bad this box really was. Even the GalCiv lead developer has commented on several occassions how bad the box art was. It was truly horrid.
The good news is, now that that Windows has finally caught up to the point where it can run a 1993 OS/2 game, their box art budget has apparently been expanded, and the new box art is amazingly cool (actually, there have been several OS/2 versions since the release of GalCiv 1.0 that also had some excellent box art). But that v1.0 box art was embarassingly bad -- particularily for such an otherwise amazing game.
If anyone knows of an online picture of the GalCiv 1.0 box, post it up here (I Googled for it for about 20 minutes to no avail -- 1993 is a bit early for stuff like this).
As the FSF's attorney has pointed out, the issue is not their code becoming LGPLed because of your code. It's the clause that requires that they allow reverse egineering and modification.
Congratulations -- you've missed my point completely!:)
My point was that _if_ a user misinterprets your license in a way that is more restrictive to them, and you tell them that _your_ interpretation is less restrictive, you're _not_ going to have a dispute on your hands as the OP suggested. Indeed -- you're going to have one happy user who, at worst, might ask you to clarify your position in writing (as an ammendment to the license).
They're hardly going to take you to court over something like this. This only ever happens when they interpret the license as being _less_ restrictive, and you retroactively decide they've overstepped their contractual rights.
I would say that that's not true. However, certainly if the two are in the same memory space, they are linked.
Define "same memory space".
On (modern) Intel-based CPUs, most OS's run in a single 32-bit memory segment. Is all code loaded into the same segment (which would be everything) now considered "linked"?
Or how about programs written for PC (or MS) DOS, where there is absolutely no memory virtualization or protection. _EVERYTHING_ lives in the same memory space -- does that make the command interpreter, the BIOS table overlay, and everything else now "linked" to my program or library?
Your definition of "linking" appears far too broad to me to make the LGPL license even usuable for anything, as it would require non-related programs on certain platforms to be under its stipulations just for being loaded into memory.
For anyone who is checking up (or anyone coming into this story _really_ late -- maybe someday some historian will be writing about me and will stumble upon an archive of this conversation on some IPv4 backwater in the far-off future...:) ), for July 14th and 15th, my website got a total of 186 hits, and _zero_ downloads.
At the same time, Dominic's Audacity project got 17634 hits, and 1179 downloads.
Dominic assumed I'd have no marketing problems with the project, seeing as how it got posted to Slashdot, but now I'm the one in the position of envy -- I'd love to have had his projects stats, even for just _one_ day:).
I say, "I use the GPL to encourage open source development. If I were to use the LGPL, then you could use my libraries without giving me the source to your program.".
They still can -- the GPL only requires a party to provide their changes to your code _if_ they redistribute the code. And then, they only have to provide it to the people they redistribute the binaries to.
You may never see their changes if they don't redistribute your code, or if they decide to redistribute it only to those people they've redistributed their binaries to (which doesn't have to be publically online -- they could do it via postal mail if they wanted to).
You've chosen a bad reason to not use the LGPL license...
Distributing a jar, along with some code which will end up calling code from that jar, is linking to that jar. The actual bindings are resolved at runtime (although as I understand it, the jar needs to be present at compile time too), but it's still linking.
If you use this as your definition of "linking", then "linking" loses all purpose and meaning.
According to your definition, it appears that any two binaries loaded in memory, where on calls the other, are "linked", regardless of wether or not the code is mingled in its distribution.
Thus, if my shell script calls on "grep", that's considered "linking", because a call interface is constructed between the two pieces of code, and the logical execution of the program flows from my code to "grep" and then back to my code.
I think you have a terrible definition of "linking" that doesn't coincide with that of anyone in the computer industry.
There's a flaw in your logic. You cannot retro-actively interpret a license. The license is an agreement. If you then decide that the agreement "really means" something other than what I think it means, we have a dispute. We can settle it between us by altering the agreement to be clear on the point, or either one of us can take the matter to court for resolution. This happens all the time. What you meant for the agreement to say is beside the point unless we had some form of verbal or implied understanding at the time.
You are correct in a technical sense, however I don't recall there _ever_ being a case where such a dispute arose when the licensor decided that their interpretation of the license was _more_ leinient than what the licensee thought. Typically, most licensees are thrilled to find out that they have _more_ rights than they assumed to be the case due to the licensors interpretation of the license.
I can't ever see this being the case:
LICENSEE: Hi, we're calling today because we're concerned about using your library in our project. We feel that section 6 of your license means that we have to license our project under the same license as yours.
LICENSOR: Hmmm, well I don't read it that way. It looks to me that you can license your code however you want, and I'm happy to let you do so.
LICENSEE: Now you've done it -- you've given us what we wanted, so now we have to take you to court.
Now I _can_ see a licensee wanting to change the license to make this right more clear, but they aren't about to get into a nasty legal dispute with your because you feel they have _more_ rights than they assumed.
I run an Open Source Java project that is using an LGPL license. It's an API set that is intended for developers. If any developer is concerned that the LGPL license requires them to Open Source their code that links my API, I'll tell them I don't read things that way, and that they have my permission to use my API without open sourcing their own code. And I'll do it in writing.
And if that isn't good enough for them, I'll distribute to them a relicensed version on a case-by-case basis.
Try to think realistically here. At this point in time, there are only 3 or 4 desktop platforms that are important enough that you might want to make sure the installation experience is painless. You can afford to compile and ship 3 curl binaries with your package in order to improve the installation experience for that large set of people. Statically link the binaries.
I'll admit to being almost militant when it comes to ensuring my project is properly cross-platform capable. My developers know it quite well -- I don't make assumptions, and I try to ensure that our code has no platform preferences whatsoever. If I wanted to target specific platforms, I would have used native code, and not Java.
A big part of this has come from my participation in the OS/2 _and_ Linux communities from a time when both were fairly new onto the market. I know what it's like to want to run a program, only to have a vendor say "Sorry, we're not interested in supporting your platform".
It's that level of "thinking realistically" that has allowed Microsoft to take over the desktop. If that were my attitude, I'd just be done with it and write nothing but Windows software:).
I got tired a long time ago of having ISVs tell me what OS I should run to use their software, so I don't (and won't) do that to my users if I can avoid it.
That said, there isn't any reason why such a system as you propose need have any native code outside the JRE itself. The libcurl Java bindings you've pointed out don't make any sense to me, because the task they achieve can be done in pure Java. Maybe I'll try to put one or two developers to the task of putting together a system to auto-download missing libraries, based on property files that have the URLs for platform-specific libraries (where they're needed).
Well then, use the libcurl bindings for Java, rather than the curl command line tool. You can still suck down any needed libraries for them. You might have to whip up dialogs to ask them for password(s) during the install.
Unfortunately, this code relies upon JNI to access libcurl. This is fine for Linux, where this library is present, but once again, this is _not_ a library that ships on many other Java-enabled platforms.
So now, in order to get the libraries using such a theoretical automated tool, we'd either have to port libcurl to every platform a user _might_ possibly want to install onto, or we'll have to have them go out and download a libcurl implementation for their platform -- which defeats the purpose of having such a tol in the first place.
Slashdot Mirror
Just FYI, but on my LinkSys BEFSR41 (firmware rev 1.44.2), all four octets are configurable.
Mind you, with only five hosts at the moment, I haven't bothered to change it out of the 192.168.x.x address space.
Yaz.
Ever play Frequency or Amplitude online? Both of these games support a mix of broadband and dialup users, and I have never seen an issue playing with people with different connection types.
Sony was smart. They made it possible for a game to specifically function with only one connection type. Take SOCOM for example: it doesn't support the dialup element of the adapter, because it just won't work.
Sony made it so the game developers can decide wether or not to target just broadband users, or everyone. There are a lot of styles of games that work quite well on a dialup connection.
And in this big old world, there are still lots of places where telephone lines are readily available, but where broadband is not. Personally, I applaud a company that is willing to embrace players in remote (or odd) areas where broadband simply isn't an option.
Yaz.
Feel glad I just used up the rest of my modpoints. Otherwise, I would have rated this -1: Troll.
It's blatently false. Out of the (quoted) 50 PS2 titles that will support online gaming by the end of the year, only TWO will charge any sort of fee: FFXI, and Everquest Online Adventures. This is as opposed to the Xbox, where you can't play _any_ games without paying a monthly fee to Microsoft.
This is also false. A number of PS2 titles support 5.1 surround via DTS encoding, wheras the Xbox only supports Dolby Digital encoding. Grand Theft Auto: Vice City, SSX: Tricky, and NHL 2003 are three titles that support in-game DTS surround.
For someone who "used to own a PS2", you certainly don't know much about it, do you?
Yaz.
All together now...
"Aw, poor SCO!"
The only thing responsible for SCO's bad press in SCO acting like a corporate idiot. They keep making statements and arguments that are contrary to both logic and law. Any wonder why even the inept computer press is beating up on them?
They hardly need IBM's help in making them look bad. They do a great job all by themselves. IBM's best strategy so far has been to keep quiet, and watch SCO implode on their own FUD.
Yaz.
Microsoft takes the 20-year-old PC architecture, puts it in a PC-sized case, designs it around a Windows-based core using common PC components, then goes around trying to get PC game developers to write titles for their console -- and they expect the _games_ to be original???
Microsoft designed the most un-original gaming console ever. They wound up with what many people predicted -- a completely un-inspired software library that looks just like your typical PC game (without getting any of the actual worthwhile PC games). Can anyone truly say they're suprised at the state of the Xbox's library?
Yaz.
So if SCO does convince some court of law that this is a valid argument, they'll self-destruct.
How on earth can SCO get theri own product CDs pressed if the CD production company is only allowed to make one copy from the master? ;)
Yaz.
Okay -- this one is purely for the "FWIW" file...
You could run NT's other original filesystem, HPFS. Linux has decent HPFS support available, the allocation unit is a 512b sector, and it's organized for fast searches and minimal fragmentation. It can also format up to 64GB partitions (although it still has a 2GB filesize limit).
The trick is to get HPFS support for Windows. To do this, you need to get the driver files from Windows NT v3.x (something that, admittedly I'm not sure works with Windows versions > NT 4. I don't do Windows personally, so I haven't tested it -- like I said above, FWIW). That will give you two of the three OSs supported. HPFS has been around for a while, (circa 1988), so you might be able to find something from the Intel FreeBSD world you could port to OS X.
I use HPFS for my 100MB ZIP disks (which I admittedly rarely use anymore for anything than quick archival purposes). It's not journalled, but it uses a bidirectional sector pointer system, so chaining errors are amost always fixable. The big downside is that if the filesystem is dirty, checking it can take a huge amount of time.
It's probably not a practical solution (I didn't and won't claim it is), but it's still a slightly more constructive answer than "Your stuck with FAT32" :).
Yaz.
If it doesn't act like a hard drive through standard protocols (NFS, NetBIOS/NetBEUI), then IMO it looks quite a bit less cool.
I want to be able to plug it in and use it as a big old hard drive. In that way it could serve as not only a player, but when connected as the core MP3/Ogg repository on my home network. There would no longer be any need to keep seperate copies of such things on the actual computers themselves -- if I'm ripping CDs, I could use whatever encoding software I want to and let it write directly to the mount point. If I want to playback on my PC, just plug in the player, mount it, and stream the data across the network.
IMO, that's quite a bit more cool than having to go through a web interface to move files back and forth.
Yaz.
Okay -- admittedly, ethernet support is a very good thing, but IMO that just means it has a useful interconnect system available.
The more important question IMO is what protocols does it support over the ethernet connection? NetBIOS/NetBEUI? TCP/IP? Some custom protocol? If TCP/IP, does it support NFS? FTP? NetBIOS over TCP/IP? Something else?
The poorly designed/formatted website doesn't give much information in this regard. I'm assuming it's TCP/IP based, but even then, you need something that will allow you to transfer files to it.
Being NFS based would be ideal here. Being able to mount it from any of my OS/2 or Linux boxes as a drive would be perfect. NetBIOS/NetBEUI/etc. would be of secondary usefulness -- I don't run any of these protocols (as I don't need to -- they add overhead and don't do anything NFS and lpr/lpd don't), but have them available on all my systems. If they're FTP based, well at least NetDrive allows me to treat them as filesystem mounts. If they're using some "custom" protocol that only works with their Windows-based client software, then no thanks.
Anyone know?
Yaz.
I don't completely agree. Sure, CPU load doesn't matter in a single-tasking system, but if I'm running a multi-tasking system I don't want it to noticably "hiccup" each and every time some background task has to do some disk I/O. The system should _not_ bog down just because Mozilla is cleaning out its cache -- it should still be responsive.
Obviously, priority levels can have a big effect on this sort of thing -- but in a multitasking environment where you are utilizing most of the CPU with various tasks, having high CPU usage for I/O can be detremental. If it's sufficiently high enough, the system will start to feel like Windows 3.1, where everything stoped until I/O is complete.
Yaz.
I've been tempted to get back into teaching classes such as this. One of my previous employers, an ISP, held such sessions for its customers to teach all sorts of interesting things, and they were generally well attended and well received.
To get people to show up, however, you have to teach them stuff they're not going to just be able to fumble their way through on their own (or with their neighbours kids help :) ). So, some of the ideas I can think up off the top of my head include:
Some thoughts, FWIW.
Yaz.
Hmmm, it looks to me like they missed the page at the beginning that reads:
Yaz.
I was there on a foggy day. There wasn't much fog; in fact, I had no idea that it was foggy until I got up into the tower. Then I couldn't see the ground anywhere. Disappointing.
If you're not queasy, the best time to go is when it's windy -- the tower is designed to sway in the wind, and the top can transcribe a 12m circle :).
Yaz.
Naturally, the size of Canada dwindles roughly to that of a medium pizza if you discount all the uninhabated/under ice portions.
Yeah, but you still can't readily backpack across it, which was really my point (and no, I didn't miss your attempt at humour :) ). Unless, of course, you really enjoy hiking through uninhabited and iced-over areas (which some people quite enjoy). Personally, I prefer to live in a country that is dominated by wide-open spaces, as opposed to one that is dominated by other human bodies :).
On a serious note though, I find it interesting that people seem to be pointing out a good bit of cultural activities and sites as well as 'geek' ones.
Well, I guess one of the difficult things is that the OP didn't really mention what type of geek they are. A history geek would probably be more interested in a fort from the War of 1812 than a science centre, which might be heaven for a science geek.
Of course, some things transcend the different stratii of geekness, and might even, parry the thought -- appeal to NON-GEEKS! :)
Yaz.
Linux's JFS support doesn't come from AIX -- it comes from OS/2, to which SCO holds no copyright nor any other IP rights.
The OS/2 implementation of JFS was a ground-up rewrite, based on the JFS specification, which is owned completely by IBM. SCO might claim that they have copyrights to the original AIX JFS sources (a dubious claim), but they can't claim they own the JFS _specification_. That's owned by IBM.
As mentioned above, the OS/2 implementation was a ground-up rewrite based on this spec. The OS/2 version of the code was then ported and integrated into Linux by IBM.
Thus I can't see how SCO could have any sort of claim on JFS in Linux. SCO has no contracts with IBM pertaining to OS/2 technologies.
Score one for the good guys...
Yaz.
Let me be the first to welcome you to Canada (considering you're not here yet, I _assume_ I'm the first at least :) ).
First things first. Canada is a REALLY BIG PLACE. You do not backpack across Canada. I know that Australia is a big place (a whole continent in fact...), and the US has a decent size, but Canada is in a whole different ballpark. Think of Australia. Now think of another 1/5 of Australia. Stick them together, and you get a bit closer to Canada's size. Canada is nearly 10 /million/ square kilometres of land, sprawling across 7 seperate time zones. It's a big place to walk across :).
As such, a good geek travel system to your trip would probably to take the train from coast to coast, getting off in major cities of interest.
Once you've figured out how to get around, where to go? Some good suggestions include (in no particular order, and probably leaving out all sorts of funky places in between...):
Well, that's what I can think of off
Two words: Galactic Civilizations.
This was (and is) actually quite a cool game, and it's just about to make a big comeback.
However, when it originally came out back in 1993, it was published for OS/2 2.1 by a company called "Advanced Idea Machines", and it had the absolute _worst_ box art of anything ever produced in the history of the world.
I have the box, and it's UGLY ugly. I've seen boxes that I'd call ugly, fugly, pug ugly, and even pug fugly -- but this is the only one I'd ever think of calling UGLY ugly. It truly is that bad :).
I wish I had a flatbed scanner so everyone else here could see how bad this box really was. Even the GalCiv lead developer has commented on several occassions how bad the box art was. It was truly horrid.
The good news is, now that that Windows has finally caught up to the point where it can run a 1993 OS/2 game, their box art budget has apparently been expanded, and the new box art is amazingly cool (actually, there have been several OS/2 versions since the release of GalCiv 1.0 that also had some excellent box art). But that v1.0 box art was embarassingly bad -- particularily for such an otherwise amazing game.
If anyone knows of an online picture of the GalCiv 1.0 box, post it up here (I Googled for it for about 20 minutes to no avail -- 1993 is a bit early for stuff like this).
Yaz.
As the FSF's attorney has pointed out, the issue is not their code becoming LGPLed because of your code. It's the clause that requires that they allow reverse egineering and modification.
Congratulations -- you've missed my point completely! :)
My point was that _if_ a user misinterprets your license in a way that is more restrictive to them, and you tell them that _your_ interpretation is less restrictive, you're _not_ going to have a dispute on your hands as the OP suggested. Indeed -- you're going to have one happy user who, at worst, might ask you to clarify your position in writing (as an ammendment to the license).
They're hardly going to take you to court over something like this. This only ever happens when they interpret the license as being _less_ restrictive, and you retroactively decide they've overstepped their contractual rights.
Yaz.
I would say that that's not true. However, certainly if the two are in the same memory space, they are linked.
Define "same memory space".
On (modern) Intel-based CPUs, most OS's run in a single 32-bit memory segment. Is all code loaded into the same segment (which would be everything) now considered "linked"?
Or how about programs written for PC (or MS) DOS, where there is absolutely no memory virtualization or protection. _EVERYTHING_ lives in the same memory space -- does that make the command interpreter, the BIOS table overlay, and everything else now "linked" to my program or library?
Your definition of "linking" appears far too broad to me to make the LGPL license even usuable for anything, as it would require non-related programs on certain platforms to be under its stipulations just for being loaded into memory.
Yaz.
For anyone who is checking up (or anyone coming into this story _really_ late -- maybe someday some historian will be writing about me and will stumble upon an archive of this conversation on some IPv4 backwater in the far-off future... :) ), for July 14th and 15th, my website got a total of 186 hits, and _zero_ downloads.
At the same time, Dominic's Audacity project got 17634 hits, and 1179 downloads.
Dominic assumed I'd have no marketing problems with the project, seeing as how it got posted to Slashdot, but now I'm the one in the position of envy -- I'd love to have had his projects stats, even for just _one_ day :).
Yaz.
I say, "I use the GPL to encourage open source development. If I were to use the LGPL, then you could use my libraries without giving me the source to your program.".
They still can -- the GPL only requires a party to provide their changes to your code _if_ they redistribute the code. And then, they only have to provide it to the people they redistribute the binaries to.
You may never see their changes if they don't redistribute your code, or if they decide to redistribute it only to those people they've redistributed their binaries to (which doesn't have to be publically online -- they could do it via postal mail if they wanted to).
You've chosen a bad reason to not use the LGPL license...
Yaz.
Distributing a jar, along with some code which will end up calling code from that jar, is linking to that jar. The actual bindings are resolved at runtime (although as I understand it, the jar needs to be present at compile time too), but it's still linking.
If you use this as your definition of "linking", then "linking" loses all purpose and meaning.
According to your definition, it appears that any two binaries loaded in memory, where on calls the other, are "linked", regardless of wether or not the code is mingled in its distribution.
Thus, if my shell script calls on "grep", that's considered "linking", because a call interface is constructed between the two pieces of code, and the logical execution of the program flows from my code to "grep" and then back to my code.
I think you have a terrible definition of "linking" that doesn't coincide with that of anyone in the computer industry.
Yaz.
There's a flaw in your logic. You cannot retro-actively interpret a license. The license is an agreement. If you then decide that the agreement "really means" something other than what I think it means, we have a dispute. We can settle it between us by altering the agreement to be clear on the point, or either one of us can take the matter to court for resolution. This happens all the time. What you meant for the agreement to say is beside the point unless we had some form of verbal or implied understanding at the time.
You are correct in a technical sense, however I don't recall there _ever_ being a case where such a dispute arose when the licensor decided that their interpretation of the license was _more_ leinient than what the licensee thought. Typically, most licensees are thrilled to find out that they have _more_ rights than they assumed to be the case due to the licensors interpretation of the license.
I can't ever see this being the case:
Now I _can_ see a licensee wanting to change the license to make this right more clear, but they aren't about to get into a nasty legal dispute with your because you feel they have _more_ rights than they assumed.
I run an Open Source Java project that is using an LGPL license. It's an API set that is intended for developers. If any developer is concerned that the LGPL license requires them to Open Source their code that links my API, I'll tell them I don't read things that way, and that they have my permission to use my API without open sourcing their own code. And I'll do it in writing.
And if that isn't good enough for them, I'll distribute to them a relicensed version on a case-by-case basis.
Yaz.
Try to think realistically here. At this point in time, there are only 3 or 4 desktop platforms that are important enough that you might want to make sure the installation experience is painless. You can afford to compile and ship 3 curl binaries with your package in order to improve the installation experience for that large set of people. Statically link the binaries.
I'll admit to being almost militant when it comes to ensuring my project is properly cross-platform capable. My developers know it quite well -- I don't make assumptions, and I try to ensure that our code has no platform preferences whatsoever. If I wanted to target specific platforms, I would have used native code, and not Java.
A big part of this has come from my participation in the OS/2 _and_ Linux communities from a time when both were fairly new onto the market. I know what it's like to want to run a program, only to have a vendor say "Sorry, we're not interested in supporting your platform".
It's that level of "thinking realistically" that has allowed Microsoft to take over the desktop. If that were my attitude, I'd just be done with it and write nothing but Windows software :).
I got tired a long time ago of having ISVs tell me what OS I should run to use their software, so I don't (and won't) do that to my users if I can avoid it.
That said, there isn't any reason why such a system as you propose need have any native code outside the JRE itself. The libcurl Java bindings you've pointed out don't make any sense to me, because the task they achieve can be done in pure Java. Maybe I'll try to put one or two developers to the task of putting together a system to auto-download missing libraries, based on property files that have the URLs for platform-specific libraries (where they're needed).
Yaz.
Well then, use the libcurl bindings for Java, rather than the curl command line tool. You can still suck down any needed libraries for them. You might have to whip up dialogs to ask them for password(s) during the install.
Unfortunately, this code relies upon JNI to access libcurl. This is fine for Linux, where this library is present, but once again, this is _not_ a library that ships on many other Java-enabled platforms.
So now, in order to get the libraries using such a theoretical automated tool, we'd either have to port libcurl to every platform a user _might_ possibly want to install onto, or we'll have to have them go out and download a libcurl implementation for their platform -- which defeats the purpose of having such a tol in the first place.
Yaz.