Slashdot Mirror

We developed a motion capture system for a bit of a different application (6-DOF joint movement tracking for biomed research). Getting the cameras working is trivial compared to the processing required to actually get 3d motion capture working reliably. Of course, we were going for something that probably has to be much more accurate than what you need, but it's not a trivial manner to write the software for something like this. Of course, there may be stuff out there you can use. Anyway, here's a brief rundown of what we were using and what might help.

We used 3 point-grey cameras (the flea). They might be a bit expensive for what you want, but all their cameras are top-quality, and they are intended for use in computer vision applications, and thus come with some source code pre-written that you can use for the interface. http://www.ptgrey.com/

We actually used windows machines at the request of the lab, but we did originally look into linux. The cameras are firewire, and the best linux drivers we've found for these types of cameras are the the libdc1394 drivers on sourceforge: http://sourceforge.net/projects/libdc1394/

The Open computer vision library is also invaluable. It has a lot of pre-written functions to deal with the more basic processing problems. It's got most of the major filters and algorithms in there that you'll need to extract the info from the camera pictures. Here: http://sourceforge.net/projects/opencvlibrary

I'm not sure how you're planning on combining or calibrating the system, but we used a static set of known coordinates and used DLTs to actually give the real 3D coordinates. A good tutorial is here: http://www.kwon3d.com/theory/dlt/dlt.html

Lastly, good trackers can really help the processing a good deal. Our trackers used a 4-ball system because we needed the accuracy and refrences for the angular rotations, but even a 1-ball tracker can be well designed. If the ball is a significant bright point on the image, simple thresholding is all that is needed in terms of preprocessing before you extract location in the image. Reflective paint or another bright source is KEY. If you're going colour, a distinct colour is also a good option.

good luck.

shouldn't we link to VirtualDub, Sourceforge and its download links a few thousand times?

shouldn't we link to VirtualDub, Sourceforge and its download links a few thousand times?

(Hiya, I'm the author of TFA and Bingo Card Creator). Here's the closest OSS program to my software: http://sourceforge.net/projects/bingo-cards/ . Feel free to use it if it fits your needs better. (I'll be perfectly honest: I think I do a much better job. For example, I have features such as "actually runs on a Windows PC instead of crashing on install" and "prints without leaving the program". If I didn't think I could do a better job than what was available for free, I wouldn't have invested my time and money into the project.) If not, you can do things the traditional way by paying your educational publisher of choice $15 a bingo card set. If you plan on doing this activity twice, ever, I really do save you money.

To get some sense of scale Celestia is a very good help.

I would definitely be wary of anything too cheap or with a higher MP count.
Preferably, try not to have the pixel count too high.

Typically, webcams are fairly lousy in terms of noise and focus. Getting a
cam which is pushing the MPs is really just going to be asking for trouble.

As far as linux go, I believe that mtp cameras are pretty much all going to
work.

http://ptp.sourceforge.net/

...I think it's very difficult, if not impossible, to have an analysis of exactly where we are as a number with supporting or complying with CSS - given that there isn't an official test suite that exhaustively tests whether you comply with the standard or not.

...in terms of stating that we really do fully support the CSS 2.1 spec, it's hard to tell because there is a bias to any analysis. We're certainly somewhere between those two... I don't think we're at 90%, I think we're above 50% though - and again, it really depends on how you end up weighing things. The problem is, if I gave any number I'd really want to support how I came up with that number - and I don't have a great way to do that today.

What really gets me as a web developer is his "Standards? Define standards? We're just groping in the dark like everybody else." attitude. Safari, Opera and Firefox seem to be figuring it out okay. As a web developer, I can design a web page and have those three browsers look pretty much the same with only minor differences. Then I spend an inordinate amount of time figuring out how to hack it in IE.

I mean, what's so difficult about figuring out how compliant you are? Get a list of the CSS 2.1 spec and make a checklist out of it. "Child selectors, check. Pseudo selectors, check. Box model, check." Do the math - (IE7 CSS implementation/CSS spec)*100 = Percent compliant. Is it perfect? Probably not but it's a start for pete's sake.

Actually, I've given up wasting time on IE (including IE 7) and just run Dean Edward's IE7 script. It's just becoming less and less worth supporting IE down to the pixel. As web developers, we're tired and we've had enough. Don't expect us to jump for joy simply because you've begun making IE a little bit better toward standards. Nothing short of 90% compliance will be worthwhile. And if you get there, don't expect a big sloppy kiss from the web developer community either because it's been long overdue.

On the other hand, I welcome Microsoft's slide toward irrelevance and the inevitable renaissance of a non-microsoft controlled web. So...I guess keep up the good work Chris!

http://doublecommand.sourceforge.net/ will remap Caps Lock to Control, among other things. I'm trying it for a while, but I'm too used to using Caps Lock for small acronyms like OS X and URL and prolly will turn it off soon.

Artificial intelligence is a philosopher's stone of "arete" for the Web browser.

1. For Linux sync, check out the SynCE project. It doesn't support WM5 (the latest Pocket PC OS) yet, but the experience you describe suggests you have an older version of the OS.
2. ActiveSync is supposed to sync-on-connect. There must be some configuration error. You definitely should not need to "export" any data; at worst, you'd just have to start syncing manually.
3. Older Pocket PCs, like older Palms, stored data in RAM. Newer (WM5) devices store data in ROM, and are therefore not susceptible to losing data upon battery failure.
4. A copy of Outlook is (and has been) included with every Pocket PC sold. Did you got the device secondhand?

There are certainly legitimate gripes about Pocket PCs, but these aren't.

--bdj

Or, download Slacke17, the newest Enlightenment, packages for Slackware as a slackpack. It even makes a 700 Mhz PIII look cool and fast!

http://slacke17.sourceforge.net/

ttyl
          Farrell

To start a you tube you will need:
FFMPEG http://ffmpeg.mplayerhq.hu/: for video conversion
FlowPlayer http://flowplayer.sourceforge.net/howto.html: displaying flash video
or Flash Video Player http://jeroenwijering.com/?item=Flash_Video_Player
FFMPEG-PHP http://ffmpeg-php.sourceforge.net/: If php is used a nice extension for getting screen shots of videos, not necessary though
flvtool2 http://rubyforge.org/projects/flvtool2/: so you can seek though the created flash file
Then all you need is leverage framework or cms in php, or phython, or something and you are done. (well sort of!)

To start a you tube you will need:
FFMPEG http://ffmpeg.mplayerhq.hu/: for video conversion
FlowPlayer http://flowplayer.sourceforge.net/howto.html: displaying flash video
or Flash Video Player http://jeroenwijering.com/?item=Flash_Video_Player
FFMPEG-PHP http://ffmpeg-php.sourceforge.net/: If php is used a nice extension for getting screen shots of videos, not necessary though
flvtool2 http://rubyforge.org/projects/flvtool2/: so you can seek though the created flash file
Then all you need is leverage framework or cms in php, or phython, or something and you are done. (well sort of!)

Bingo Card Creator, software to help elementary school teachers make bingo cards. If you really want to spite me with an open source alternative, http://sourceforge.net/projects/bingo-cards would probably be the closest thing to my program. Oh no, I just spited myself! ;) I don't think its very responsive to the needs of my market, which you can verify in about 30 seconds if you've got an elementary school teacher handy. I actually passed it around to a few of my teacher friends and the response universally was "Um, it won't run". Have fun with it.

Tip #1: Use a Palm OS device.

I'm sure you realize that PalmOS devices store *ALL* of their data in cleartext, right? Marking those records private and protecting them with a password?

Futile, just fetch the records directly (and pilot-link is the de-facto tool for this) and open it in an editor, or run strings(1) across it to see everything in cleartext.

There are applications, such as GNU/Keyring and others that can help you secure your passwords, memos, data and whatever else you want on PalmOS devices.

In short, never trust the vendor's default application suite to do what you want, or be as secure as you need.

I have a few minutes before I rush off to work...

Yes I can arrange for administrative rights on all the target machines (I believe due to their set up anyone who can access the Administrator account on one will have the same rights on another). And yes they are all part of the same domain so if that is not workable a user with access to all of them can be created.

I've cast a quick eye over unattended and it certainly DOES look interesting (the points discussing other installation processes were enlightening too).

To the person who suggested frontmotion - that's interesting but I would like a en_GB build, it throws away the Mozilla branding (because it's not officially blessed) and has the disadvantage of not being a trusted source.

To the person who posted a bugzilla link - thanks, a quick skim is already proving informative.

My HP printer works fine, thanks to hplip (which, incidentally, is also a Debian package).

You'll have to spec your problem a bit more tightly, I'm afraid... Do you have administrative access on the machines you wish to install stuff onto, through a domain or something (if you've been added to the local administrator group on each of the target machines)? If so, you could use unattended, and make a service using perl to modify the registry to start unattended on next reboot, and go through and do all the installation tasks. I estimate about a week to do all the research, right now the part that isn't in place is the service that adds stuff to the registry, AFAICT....

Cheers,

Michael

First of all, I thought the GPL could not be modified. In fact, this is the very first sentence of the GPL:
  Everyone is permitted to copy and distribute verbatim copies
  of this license document, but changing it is not allowed.

Secondly, sourceforge.net clearly states that they are for hosting OSI-compliant projects ONLY. I wouldn't be suprised if sf.net terminates their hosting, only before asking them to revert the license.

This patch restricts the field of endeavour of the Program in such a way that this license collides with paragraph 6 of the Open Source Definition. Therefore, this modified version of the GPL is no more OSI compliant. ... You can find the full text of the license at: http://gpu.sourceforge.net/GPL_license_modified.tx t

To summarize: this is a project released under a non-Free license, and is also not Open Source according to the OSI. I hope that Sourceforge (which claims to provide hosting and resources only to Free software) will immediately stop hosting their project files.

This is a violation of Stallman's Freedom 0 - military users don't even have the freedom to run the software, let alone modify it! Sourceforge should take a stand on principle.

http://sourceforge.net/projects/gpu/

They need to update the shrink-wrap license description:

Project Admins: akoese, artee, dangermaus, delphifreak71, edsartori, paulatreides, ptea, redshift_eric, seeschloss, shakezilla, simbasabi, wppetersen
Operating System: WINE, All 32-bit MS Windows (95/98/NT/2000/XP), Linux
License: GNU General Public License (GPL)
Category: Gnutella, 3D Rendering, Scientific/Engineering, Frameworks, Clustering, Distributed Computing

From the 'No Military License':
"... The Program and its derivative work will neither be modified or executed to harm any human being nor through inaction permit any human being to be harmed. This is Asimov's first law of Robotics."

Aside from the obvious direct military applications, would this restrict its use by people who create laws which allow harm to come to humans? The "harm" clause is quite subjective in its definition. How they perceive harm may be different from how others percieve harm. This may be struck down in the courts as being "too broad" and find it is not a valid license as has been done many times in the past in regards to laws (both state and federal laws).

-- The software industry is driven by wanting more while using less; which is why Microsoft products provide successively more options and features while leaving you with less free drive space, less free memory, less stability and less peace of mind. --me
sTc

Below is what was added to the GPL:

                          PATCH FOR NO MILITARY USE

This patch restricts the field of endeavour of the Program in such a way that this
license collides with paragraph 6 of the Open Source Definition. Therefore, this
modified version of the GPL is no more OSI compliant.

The Program and its derivative work will neither be modified or executed to harm a
ny human being nor through inaction permit any human being to be harmed.
This is Asimov's first law of Robotics.

You can find the full text of the license at: http://gpu.sourceforge.net/GPL_license_modified.tx t

AFAIK, Matlab deals mostly in ordinary floats, and does none of the esoteric things you mention to develop perfectly exact numbers. That's mainly the province of symbolic mathematics packages such as Mathematica, Maple, and (for something just as powerful that's actually Free Software) Maxima. These programs are actually capable of performing exact arithmetic even on irrational and transcendental numbers, by using rational numbers where this is possible, and by representing irrational/transcendental numbers as functions involving only rational numbers that can be used calculate the number to arbitrary precision, e.g. they might represent pi by using one of the many available power series expansions for it, and use their symbolic mathematics system to take care of the manipulations. Needless to say, this is very expensive, and massive overkill for almost everything but mathematics research and some specialized work in theoretical physics.

"I welcome any example where it took 4 weeks for a fix for a main package."

Well offhand, here is one opened 3 years ago which still hasn't been fixed, though it would be difficult to exploit. Basically what happens is that that a machine with trust level 4 (the default is 3, so again this would be difficult to exploit) to gain level 5 access (meaning they can run arbitrary commands on computer running the service. No, STAF/STAX is not as big as Linux (which is why I was talking about open source in general, not just Linux, which isn't even the software this article was about), but it is used in many corporate environments as an automated testing tool.

info.edu.org - Speedy information and news from the Top 10 educational organisations.

• TWM, 1987
• FVWM, 1993 (Enlightenment puts it at 1992)
• Next Step publishes Open Step which is quickly followed by
• AfterStep, Window Maker and others much nicer than Windows 95. Most are still available and usable with the latest and greatest free software.
• Enlightenment, released 1996, still a leader.
• Gnome used Enlightenment until they moved to Sawfish. The history has just begun

Of course, everyone should see the first web browser from 1990 (actually a screen shot from 1993, but much the same) running on a Next.

It might be hard to dig up screenshots all of desktops, but not much harder than the ones they found. It's nice to see someone including KDE in the line up so people can see a little of what they have been missing, like Virtual desktops, since the early 90's.

Correction: it seems privoxy development is alive and well

http://equ.sourceforge.net/, XMMS plugin equalizer, works with any format.
The wide selection of plugins is the reason I still use XMMS, it's the only player that plays every audio file I have.

The only open-source tool I know of is called FindBugs developed by the University of Maryland and available at http://findbugs.sourceforge.net/.

However, it only performs static analysis on Java programs.

Yes.

http://mediacoder.sourceforge.net/

Besides, I'm hard pressed to think of GTK+ as dominant. You'll notice that KDE has begun its path into the Macintosh and Win32 markets, both places where Gnome doesn't tread.

Give me a break. Firstly, GTK+ is not the same thing as GNOME. Secondly, GTK+ is and has been available free for Win32 for years. I recently wrote a Python program that used GTK and Glade for a project I was working on. The program ran out of the box (with all the dependencies installed, of course) on Debian and Win32, and I chose this Python/Glade/GTK+ technology combination because I needed to get the entire program written in a weekend, I needed it to work on both Windows and Linux, and I knew that the technology was stable enough to Just Work (TM). It did.

You mean like DOSBox?

The cost of your space has nothing to do with how you organize that space.

Your eyes still need to look at the tail of that 5000-message-blob in your inbox, EVERY TIME YOU OPEN IT.

Your brain has to subconciously categorize each of those messages.. EVERY TIME!

No wonder some people I know never get anything done until I've reminded them 5 times. They've got finished projects, to-do items, calendar items, junk mail, forwarded crap, outdated crap, and who-knows-what in their inbox.

Do yourself, and the people you work with, a favor: keep your inbox reserved for new messages only.

do you keep your bills in your mailbox after the postman has delivered them? No, you probably put them in a "bills to pay" pile.

http://ficl.sourceforge.net/

If you liked neon you might like ficl; it has the same features you mentioned with choice of early and late binding.

Many forth object systems do.

I did a port of Ficl to MS Smartphone 2002/2003 but couldn't get permission to release it.
(By port I mean filling in some missing libc functions)

Sam

Take for example, expose's ability to show you every window you have open at the same time. This is trivial to do. But it's such an amazingly useful thing and it's implemented elegantly.

Those of us who've been using 3DDesktop under *nix for the past four or five years would agree.

While I'm not a regular OS X user these days I really haven't seen a single thing in any release that was a completely original idea from Apple, except for the bits they took from their previous OSes (System7-9, especially). Implementing other people's idea in a visually appealing fashion has been Apple's only strength. Now if they could just make it *efficient* as well...

A free program called Space.app ran on 10.2. My addiction to multiple desktops runs deep. It worked flawlessly until I upgraded to 10.4.

This is not always true. Some two-factor systems are actually well thought out and resolve both of your problems. Check out http://sourceforge.net/projects/wikid-twofactor for an example. It uses Public Key Crypto to allow you to easily register your single "token" at any number of totally autonomous sites. Your private key is never shared with anyone and the websites you use it with have no idea where else you bank/shop/etc. It's real two-factor, so you not only need the pub/priv keys but you need to know the PIN that you setup for that site. This isn't stored locally so a stolen token gets the thief exactly nothing. This system also uses mutual authentication to ensure the site you're logging into is really the site you think it is. Use your phone, PDA, laptop, etc. to generate the (single use) passcode and that's a really tight system.

Or a free, opensource two-factor solution like http://sourceforge.net/projects/wikid-twofactor

I gotta agree. There are real two-factor solutions that make this and other problems go away. Even open source solutions http://sourceforge.net/projects/wikid-twofactor/ that totally remove cost as an excuse for not being secure. Keylog a One-Time-Passcode (OTP) all you want, it's only good once. Add to that that you can generate the code on a separate device (like a phone or pda) for use logging in on your PC and that's pretty tight, even beats a race attack. Top that off with mutual authentication to prevent a man-in-the-middle or DNS attack. Why aren't companies doing this? Instead they roll out flawed systems like the HSBC junk or the pretty pictures tech that is becoming popular. This is a solved problem.

Keeping passwords in text files seems like a bad idea. If you want to keep passwords safe on a networked machine you should be using something like http://keepass.sourceforge.net/. Additionally if you are worried about key loggers never type in your password even after failed attempts, just copy and paste the letters of your password into the password field (getting the letters from a web page, or typing out the alphabet and numbers 0 to 9, etc...); this is the only way to be sure that your password will never show up on the keyboard logger.

0: dban
1: autonuke

Maybe the solution to untrainable filters, if there really is a way to do this, is to use other means. If you're concerned about all this check out the ASSP project at http://assp.sourceforge.net/ or http://www.magicvillage.de/~Fritz_Borgstedt/assp/ for the absolute latest stuff, and don't worry so much about the Bayes part of things. It uses bayesian filtering also, but it incorporates a wide array of other methods to block spam which happen before it even gets to the bayesian filters.

This seems like a good on-topic thread in which to mention the freedesktop.org (X.org folks) effort to write a 100% open source 3D driver for the NVidia cards -- nouveau

http://nouveau.freedesktop.org/wiki/

If you're an owner of an nVidia card, please do all you can to help contribute! They appear to be suprisingly far along.

--
Slashcode bug # 497457 - unfixed since December 2001 - Go look it up!

Yup: http://ipw2200.sourceforge.net/

I use assp as my spam filter: http://assp.sourceforge.net/ It always filtered spam very well for me but the latest version added an interesting technique that has reduced the amount of spam that's even hitting the filters to near zero. Since SMTP is considered "unreliable" a sending server will retry on failure. Apparently, spammers tend not to bother retrying. ASSP builds tables using an identity triplet (I can't remember the three message/source attributes it uses). On first view of a given triplet, ASSP responds with a SMTP error suggesting the source retry later. ASSP tables the triplet and allows that traffic to pass later on a retry. The triplet expires after some period. I'm not aware of any false rejections and the messages hitting the dump mailbox has dropped from around 10 a day to a couple a week. I suppose one might argue that it increases packet traffic and I assume spammers will workaround it but I suspect the extra packet traffic is far exceeded by the spam that I would otherwise handle and it handles the spammers for now. Sentience unnecessary perhaps.

Yeah, they're annoying, and doubly annoying for anyone joe jobbed, and poorly setup C/R systems annoying mailing lists, but there's one thing that can't be beat about them: You can guarantee a human at the other end (assuming it takes more than a just pressing reply) and you can track spammers down that bother to put the effort in. Oh, and you don't need to "upgrade" SMTP or get someone to adjust your DNS server (Here's looking at you, SPF!) to get them to work.

The net cost of getting humans to reply to C/R mails means spam becomes expensive.

Yes, it sucks, and yes, there's the people out there that refuse to work with C/R systems. But I don't care. I don't need to talk to everyone on the internet, and the 1% - 2% that won't deal with C/R can FOAD for all I care.

The issues of C/R systems having infinite loops, etc, have been worked out over the years. That doesn't happen anymore with the latest versions. I would reccomend looking at either TMDA for a server side solution, or ASK for a client side solution.

(Of course, there's specific instances where C/R systems are simply too annoying, like trying to get sales leads, etc, but for the average person, that's not an issue.)

The best design would be a SPAM filter with a C/R system for mail that isn't marked SPAM. Joe jobs become much less of an issue, and you still don't get any SPAM.

the Delphi grammer is supposedly proprietary and a trade secret, and that no one else can build a parser that handles all of the corner cases

Of course it's possible to build a parser which handles all corner cases, be it not in one shot since there is indeed no design document to base yourself on. But by now at least we (the Free Pascal Compiler) do support most of it.

I am kinda switching over to Java Swing for the GUI, C++ through the JNI for the hardcore numeric stuff -- a person sees the handwriting on the wall about Borland longevity. I kinda want to get off Windows by the time Vista, Aero, and whatever Windows-specific GUI gobbeldygook takes hold, and Java looks attractive to me. It is easy to get spoiled by GC and some other Java hacks, but I miss the fast compiles -- javac is dog slow.

You might want to have a look at Lazarus. Although the underlying Free Pascal Compiler is not as fast as Delphi, it's a lot faster than most other compilers (and javac). Have a look here for a small unscientific test. Can't compare to Delphi, because the compiler itself is not written in a Delphi-compilable way.

For Java, try using IBM's jikes compiler instead (not to be confused with the JikesRVM -- jikes is simply a Java compiler, not a whole virtual machine).

Eclipse, on the other hand, is still something I am struggling with -- it is IBMish in its weirdness about a whole bunch of stuff -- how do you just take a bunch of .java files in a directory and call it a project? It doesn't like a directory that is already there, and it wants to create some other directory in some standard location instead of where you want it. Still working on that one. So even Java needs "make files" of some kind of external metadata to organize multi-file source codes. Object Pascal famously has all of the dependencies specified in source code as a feature of the language.

Java does specify the dependencies in the source as part of the language. You have to "import" all packages you need, much like a uses clause (and it allows for both more fine-grained and coarse-grained control regarding what you want to import).

The only difference is indeed, as you mentioned, that packages should recide in a directory called the same (or in a jar file). This organisation however avoids problems you can have in Pascal if you e.g. have multiple units with the same name in different directories. It does not necessitate the use of "makefiles of some kind of external metadata" in any way though (unless you consider directories falling under that).

If there are two features, no, make that three features! that make Delphi problematic for the future, they are 1) Delphi has some real solid Windows lock-in, especially since Kylix didn't go anywhere -- the Delphi extension to Pascal are quite Windows-specific even if you are not doing GUI programming,

Most Delphi extensions can be perfectly ported to other platforms. FPC runs on Dos, Windows, Linux, Mac OS X, FreeBSD, Solaris, Gameboy Advance, ... and the supported basic features are the same everywhere (except for threading, which is not supported on Dos and some other lightweight platforms).

2) while Delphi is great and everything, it is behind the times with collection classes and other library richness of everything from Java to Python -- what collection classes are there are also tied into the VCL and bloat non-GUI or GUI-non-VCL programs,

I do agree with you here (in spite of the fact that there are some third party container collections for Delphi).

>I hope this is something which makes the closed source WLAN (it's working, it's ok) fans
>a bit quieter.

Wlan? Think NVIDIA + http://eckbox.sourceforge.net/