Many people have often tried to hack web stuff into doing shared 3D. For example, the Open Source Metaverse Project uses an SQL database to store stuff and web servers to serve it.
It really doesn't make a ton of sense once you start trying to do it, other than the fact that web technology is already deployed. Just getting the initial data to people is not hard. Its how to implement changes, interaction, etc.
SL is build both technically and socially on the idea of real estate. The scalability of that is one aspect in which it doesn't work right. But this kind of thing is a hard problem. It's something we have thought about in designing another open source shared virtual environment based on the Virtual Object System but have not yet implemented anything to solve it in our servers. (Though the tools are there for it to be added in the future.)
I've just started playing with Bazaar-NG (http://www.bazaar-vcs.org) for my own personal local branches. Even though the official repository may be in CVS or Subversion, you can then use bzr to seperate out into branch directories various features still in progress, while keeping your CVS/Svn checkout "pristine". This is very helpful at keeping potentially disruptive changes seperate for your own sake, as well as your team members'. You don't want to be the poor fool who broke everything by a careless or mistaken commit; and you certainly don't want to waste time re-writing code you accidentally obliterated when mucking about moving files around, or trying to re-seperate cross-cutting changes to create seperate patches, revise, or debug stuff.
Bazaar-NG is still very new, and there are a few bugs though.
For a non-technical group, I highly recommend using CVS or Subversion with TortoiseCVS (http://www.tortoisecvs.org/) on the clients. But you need techies to manage the repository. If a non-techie needs to be managing the repository, I've used CS-RCS (http://www.componentsoftware.com/products/rcs/). It's not free, and it's clunky and dumb, but it works without needing anything other than a network and Windows.
Also, note that this is research into *autonomous* mobile robots (or semiautonomous). NOT remote controlled like all current unmanned ground vehicles used by the military. (UAVs tend to have some autonomy these days, I think.)
What is not shown in the video that I'd like to have seen is a more dynamic environment, even hostile. Navigating around stationary objects is not in itself impressive, though their forrest of PVC pipes is more (spatially) constrained than most tests. Navigating in a dense hostile environment is a less explored problem area.
The robot hardware platforms are a customized ATRV (four wheels) by RWI (now iRobot) and a Pioneer DX (two wheels) by ActivMedia (now MobileRobots Inc.). This hardware would not be used in the field (Since RWI no longer exists and the DX is about 2 feet tall), but the navigation principles they developed might be (along with many others).
If you're doing it on the cheap and only have three or four projectors, you don't need much of a cluster, just a three or four networked computers. Or, use two dualhead computers.
You'll have a small amount of lag in the syncronization (network + OS + application software) but with some tweaking of the OS network configuration, or using some insanely fast system rather than a network (shared memory backplane?), you might get it to a few ms?
If you find one that's not wireless, it might be a whole lot cheaper. Also I used to have a finger mouse I got for like 2 bucks that had a little trackball on top for the thumb with the mouse button as trigger, but lost it.
If you have some time and expertise, you can do some motion tracking with webcams. The lower the resolution, the faster, actually!
For software though I have no choice but to selfishly invite people to join the interreality project (http://interreality.org) which can't do a CAVE out of the box but could if you synced up several clients (one for each projector) -- not hard, we did it with an older version of our software.
1. The fact that it only displays one emblem in list view mode is unfortunate -- if in list view there was a column for each emblem (or a "subcolumn" for an "emblems" main column), which you could use as a sort criteria, then you could very easily find files with certain emblems.
2. Automatic and dynamic emblems based on combinations of things like current age, original directory of creation, current directory, file type, size, patterns in the filename or grepped from the contents, etc.
3. Ability to create new emblems on the fly, even without an icon (just text), right from a particular files "properties" window or the sidebar. Really they are the same idea as "tags" and you should be able to invent new ones as needed without going through the "miscellaneous file properties" catchall bin that is "Backgrounds and Emblems" in the edit menu.
4. Using emblems when doing a full filesystem search; a seperate catalogue for emblemized items could be kept to make it very fast. If the actual filesystem supports "tags" or "keywords" as metadata for files, then add emblem tags to the files, so non-nautilus aware tools could use these.
The nautilus-scripts thing is useful however. There is a script to upload photos to flickr at http://nozell.com/blog/archives/2004/09/04/flickr- upload-for-gnomes-nautilus/ though the progress bar doesn't update right. I also made some shell scripts that resize images using 'convert' from ImageMagick to thumbnail size and webpage size (e.g. max 700 px wide).
One thing it shows though is that there is still a lot of confusing inconsistency on where Gnome-related applications store preferences and other data. IMO it should *All* be in ~/.nautilus, not scattered between there, ~/.gnome2, ~/.gtk, etc. You probably also have a ~/.gnome too for non-Gnome2 apps.
The global settings for Gnome are also scattered everywhere.
Welcome to the world of working on large projects. Figuring out how they work when you've only just started peeking at a little piece of it is a skill you have to work on over time. Almost all large projects are like this to one degree or another.
Everyone who says about VIM that "it's hard to learn" don't really get VIM.
It's not supposed to be easy to learn. That's just not a consideration when it comes to VIM. In fact, the only consideration of VIM is to provide good tools for editing a file of source code.
That's pretty much it.
And it does that one job very well and efficiently.
(Of course, I don't really see how spellcheck has anythnig to do with editing source code, but I guess it was easy enough to bring in a spellchecker. Does it have its own dictionaries or does it use e.g. ispell's dictionary?)
I know Linux doesn't really "do" *cohesive documentation* or anything like that but a table of { kernel versions vs. hardware drivers (with both module name and an actual descriptive name, and author/contact person/website) -> status (not supported; present but experimental/untested; available from a third party patch/module; present but bugs outstanding; stable) } is one of the most basic and most useful things that could be placed on http://www.kernel.org/ and actively maintaned by the driver authors and other developers.
The company I work for (http://www.mobilerobots.com) sells both hardware and software.
Some of the software is GPL, namely all the development libraries, and the robot platform usually runs Linux (though you can get Windows on it if you really want). We also use a lot of free software in development, or depend on free libraries. This lets a software development staff of four turn out a lot of useful stuff quickly, and using Linux of course brings down the customer's final cost by a few hundred dollars, as well as being a bit more technically suited to the application. Having free/open source development libraries is of course also a great benefit to the customer, who can hack it however they want.
In our case it really is the hardware that brings in the dough; the software just makes the hardware a *lot* more usable!
Bob Kahn can be given most of the TCP/IP credit, along with ARPANET researchers, Vint Cerf, and other colleagues at places like SRI, BBN, PARC, various universities. Bob Metcalfe at PARC for Ethernet (which got some inspiration from MAC and other projects at MIT I think, along with various contemporary telecom industry protocols).
If the close box is the right size, it's not a big problem. For example, in Safari the close button is a small circle, you have to be pretty precice to hit it.
Fitt's law can actually be exploited for positive effect sometimes.
The advantage to pen and paper is you don't have to worry about the mechanics of typing, syntax, whatever. I know that most programmers have evolved into a keyboard-hand cyborg and we all have atrocious handwriting as a result!
But on a piece of paper or whiteboard you can draw arrows and make boxes around things and put things in parallel columns, etc. You can draw subscripts on variables, put exponents and fraction components in the right places, use greek letters....... the possabilities are endless. You're trying to express your thoughts precicely and do a mental check on whether it's correct.
I certainly use a pen and yellow legal pad for the actual tricky bits. I also tend to do a lot of geometry calculations in the kinds of programs I write, so that helps.
My wife uses MS speech recognition to make notes and journal entries while driving. It requires a bit of cleaning up later but since the text is for her use only it's not a big deal. She "trained" it in the car while we were on a long trip and it's quite accurate for common words and phrases.
-1 redundant is a filter. moderation's goal is to improve the *output* -- what posts you see when you browse at 3, 4, 5. redundant posts in that output is a waste of readers' time. it's not personal.
Think about what you're saying, and what "conservative" really means -- Lieberman is quite conservative. Kerry is pretty moderate. Even Dean's policy would probably have been only somewhat liberal, even if for some odd reason everyone thought he was some kind of Marxist simply because he actually opposed Bush strongly.
"I'm not sure why this is instantly regarded as some sort of conspiracy rather than either hardware problems or incompetent voting machine vendors."
Either way, the voting machines are unsuitable and must be replaced. Furthermore, whatever process the county used to choose a voting machine vendor is unsuitable, and must be replaced.
Re:Time to get into the habit of read-only filesys
on
First Mac OS X Virus?
·
· Score: 1
I meant to keep the files at mode 555, or on a read-only filesystem. Then, to modify the file, you would need to either chmod them or remount it first, which could be made to loudly notify the user that someone is trying to screw with some improtant system files.
Obviously some people would still click OK, but, maybe less.
Slashdot Mirror
By your logic, rock is always better -- worst case is a tie with your opponent's rock.
Many people have often tried to hack web stuff into doing shared 3D. For example, the Open Source Metaverse Project uses an SQL database to store stuff and web servers to serve it.
It really doesn't make a ton of sense once you start trying to do it, other than the fact that web technology is already deployed. Just getting the initial data to people is not hard. Its how to implement changes, interaction, etc.
SL is build both technically and socially on the idea of real estate. The scalability of that is one aspect in which it doesn't work right. But this kind of thing is a hard problem. It's something we have thought about in designing another open source shared virtual environment based on the Virtual Object System but have not yet implemented anything to solve it in our servers. (Though the tools are there for it to be added in the future.)
I've just started playing with Bazaar-NG (http://www.bazaar-vcs.org) for my own personal local branches. Even though the official repository may be in CVS or Subversion, you can then use bzr to seperate out into branch directories various features still in progress, while keeping your CVS/Svn checkout "pristine". This is very helpful at keeping potentially disruptive changes seperate for your own sake, as well as your team members'. You don't want to be the poor fool who broke everything by a careless or mistaken commit; and you certainly don't want to waste time re-writing code you accidentally obliterated when mucking about moving files around, or trying to re-seperate cross-cutting changes to create seperate patches, revise, or debug stuff.
Bazaar-NG is still very new, and there are a few bugs though.
For a non-technical group, I highly recommend using CVS or Subversion with TortoiseCVS (http://www.tortoisecvs.org/) on the clients. But you need techies to manage the repository. If a non-techie needs to be managing the repository, I've used CS-RCS (http://www.componentsoftware.com/products/rcs/). It's not free, and it's clunky and dumb, but it works without needing anything other than a network and Windows.
For technical documentation, I high
Also, note that this is research into *autonomous* mobile robots (or semiautonomous). NOT remote controlled like all current unmanned ground vehicles used by the military. (UAVs tend to have some autonomy these days, I think.)
What is not shown in the video that I'd like to have seen is a more dynamic environment, even hostile. Navigating around stationary objects is not in itself impressive, though their forrest of PVC pipes is more (spatially) constrained than most tests. Navigating in a dense hostile environment is a less explored problem area.
Reed
Here are papers and more videos from the lab's web page: http://www.eng.fsu.edu/ciscor/research_highlights. htm
Lab's web page: http://www.eng.fsu.edu/ciscor/
The robot hardware platforms are a customized ATRV (four wheels) by RWI (now iRobot) and a Pioneer DX (two wheels) by ActivMedia (now MobileRobots Inc.). This hardware would not be used in the field (Since RWI no longer exists and the DX is about 2 feet tall), but the navigation principles they developed might be (along with many others).
foreach page in wikipedia:
baidupedia.createPage(page)
See how long it takes for their censorship bureau to get out from under that...
If you're doing it on the cheap and only have three or four projectors, you don't need much of a cluster, just a three or four networked computers. Or, use two dualhead computers.
. asp
You'll have a small amount of lag in the syncronization (network + OS + application software) but with some tweaking of the OS network configuration, or using some insanely fast system rather than a network (shared memory backplane?), you might get it to a few ms?
If you want frame-by-frame synchronization you need some specialized equipment driving the projectors, stuff like this: http://www.es.com/products/image+generators/index
(Anyone making a homebrew CAVE want to try using http://interreality.org/ VOS software in it?)
wireless mice designed for people giving powerpoint presentations are a nice cheap solution. E.g: http://www.logitech.com/index.cfm/products/product listns/US/EN,crid=1999,categoryid=371
If you find one that's not wireless, it might be a whole lot cheaper.
Also I used to have a finger mouse I got for like 2 bucks that had a little trackball on top for the thumb with the mouse button as trigger, but lost it.
If you have some time and expertise, you can do some motion tracking with webcams. The lower the resolution, the faster, actually!
For software though I have no choice but to selfishly invite people to join the interreality project (http://interreality.org) which can't do a CAVE out of the box but could if you synced up several clients (one for each projector) -- not hard, we did it with an older version of our software.
There is nothing this room can do that a decent set of VR goggles can achieve.
More than one person can be in it?
Besides have you ever actually used VR goggles? For any length of time?
Does anyone actually use these?
Four things would make them actually useful:
1. The fact that it only displays one emblem in list view mode is unfortunate -- if in list view there was a column for each emblem (or a "subcolumn" for an "emblems" main column), which you could use as a sort criteria, then you could very easily find files with certain emblems.
2. Automatic and dynamic emblems based on combinations of things like current age, original directory of creation, current directory, file type, size, patterns in the filename or grepped from the contents, etc.
3. Ability to create new emblems on the fly, even without an icon (just text), right from a particular files "properties" window or the sidebar. Really they are the same idea as "tags" and you should be able to invent new ones as needed without going through the "miscellaneous file properties" catchall bin that is "Backgrounds and Emblems" in the edit menu.
4. Using emblems when doing a full filesystem search; a seperate catalogue for emblemized items could be kept to make it very fast. If the actual filesystem supports "tags" or "keywords" as metadata for files, then add emblem tags to the files, so non-nautilus aware tools could use these.
OK, that was a completely useless article.
- upload-for-gnomes-nautilus/ though the progress bar doesn't update right. I also made some shell scripts that resize images using 'convert' from ImageMagick to thumbnail size and webpage size (e.g. max 700 px wide).
The nautilus-scripts thing is useful however. There is a script to upload photos to flickr at http://nozell.com/blog/archives/2004/09/04/flickr
One thing it shows though is that there is still a lot of confusing inconsistency on where Gnome-related applications store preferences and other data. IMO it should *All* be in ~/.nautilus, not scattered between there, ~/.gnome2, ~/.gtk, etc. You probably also have a ~/.gnome too for non-Gnome2 apps.
The global settings for Gnome are also scattered everywhere.
I wish they'd fix that.
Welcome to the world of working on large projects. Figuring out how they work when you've only just started peeking at a little piece of it is a skill you have to work on over time. Almost all large projects are like this to one degree or another.
Beware, SGI's STL is very close, but is not exactly like that in the C++ standard or in GCC or Visual C++.
Everyone who says about VIM that "it's hard to learn" don't really get VIM.
It's not supposed to be easy to learn. That's just not a consideration when it comes to VIM. In fact, the only consideration of VIM is to provide good tools for editing a file of source code.
That's pretty much it.
And it does that one job very well and efficiently.
(Of course, I don't really see how spellcheck has anythnig to do with editing source code, but I guess it was easy enough to bring in a spellchecker. Does it have its own dictionaries or does it use e.g. ispell's dictionary?)
I know Linux doesn't really "do" *cohesive documentation* or anything like that but a table of { kernel versions vs. hardware drivers (with both module name and an actual descriptive name, and author/contact person/website) -> status (not supported; present but experimental/untested; available from a third party patch/module; present but bugs outstanding; stable) } is one of the most basic and most useful things that could be placed on http://www.kernel.org/ and actively maintaned by the driver authors and other developers.
The company I work for (http://www.mobilerobots.com) sells both hardware and software.
Some of the software is GPL, namely all the development libraries, and the robot platform usually runs Linux (though you can get Windows on it if you really want). We also use a lot of free software in development, or depend on free libraries. This lets a software development staff of four turn out a lot of useful stuff quickly, and using Linux of course brings down the customer's final cost by a few hundred dollars, as well as being a bit more technically suited to the application. Having free/open source development libraries is of course also a great benefit to the customer, who can hack it however they want.
In our case it really is the hardware that brings in the dough; the software just makes the hardware a *lot* more usable!
Bob Kahn can be given most of the TCP/IP credit, along with ARPANET researchers, Vint Cerf, and other colleagues at places like SRI, BBN, PARC, various universities. Bob Metcalfe at PARC for Ethernet (which got some inspiration from MAC and other projects at MIT I think, along with various contemporary telecom industry protocols).
It's not as bad as middle-click closing a tab.
If the close box is the right size, it's not a big problem. For example, in Safari the close button is a small circle, you have to be pretty precice to hit it.
Fitt's law can actually be exploited for positive effect sometimes.
The advantage to pen and paper is you don't have to worry about the mechanics of typing, syntax, whatever. I know that most programmers have evolved into a keyboard-hand cyborg and we all have atrocious handwriting as a result!
But on a piece of paper or whiteboard you can draw arrows and make boxes around things and put things in parallel columns, etc. You can draw subscripts on variables, put exponents and fraction components in the right places, use greek letters....... the possabilities are endless. You're trying to express your thoughts precicely and do a mental check on whether it's correct.
Most code? Hell no.
I certainly use a pen and yellow legal pad for the actual tricky bits. I also tend to do a lot of geometry calculations in the kinds of programs I write, so that helps.
My wife uses MS speech recognition to make notes and journal entries while driving. It requires a bit of cleaning up later but since the text is for her use only it's not a big deal. She "trained" it in the car while we were on a long trip and it's quite accurate for common words and phrases.
-1 redundant is a filter. moderation's goal is to improve the *output* -- what posts you see when you browse at 3, 4, 5. redundant posts in that output is a waste of readers' time. it's not personal.
Think about what you're saying, and what "conservative" really means -- Lieberman is quite conservative. Kerry is pretty moderate. Even Dean's policy would probably have been only somewhat liberal, even if for some odd reason everyone thought he was some kind of Marxist simply because he actually opposed Bush strongly.
Bush is the biggest radical in Washington.
"I'm not sure why this is instantly regarded as some sort of conspiracy rather than either hardware problems or incompetent voting machine vendors."
Either way, the voting machines are unsuitable and must be replaced. Furthermore, whatever process the county used to choose a voting machine vendor is unsuitable, and must be replaced.
I meant to keep the files at mode 555, or on a read-only filesystem. Then, to modify the file, you would need to either chmod them or remount it first, which could be made to loudly notify the user that someone is trying to screw with some improtant system files.
Obviously some people would still click OK, but, maybe less.
Just an idea.