>> ^Drachen_Jager:

>> ^mxxcon:
>> ^Drachen_Jager:
No scanner would see a program in an image, realize it's an executable and execute the program all without the user's knowledge.
It's barely possible that you could transfer a workable program that way through fractals, and the people on the other end would have to spend weeks or months decoding the program before they could make it run IF they could figure out what it was in the first place.
pattern might be constructed in such a way that when scanned it can confuse any piece of software in the chain to cause a buffer overrun, at which point they could craft the following data to actually be executable and contain malware payload.
This is kinda what happened in a few instances of MP3-based viruses. MP3 file's metadata was malformed in such a way that it would crash the player and execute trojan payload that was embedded in that MP3 file.
Look up information about MP3Concept(MP3Virus.Gen)
There was another incident that involved MP3 file played in Winamp player. I can't find link about it now.
So while realistically unlikely, it's possible.

The main difference there being that the MP3 was a digital file which had been manipulated, not a sound recording. The file was corrupted in a very clever way. In the show the bone 'picture' was taken by the forensics team. They did not bring in a file the guy had created, they took a photograph. You can't insert corrupted data, because the data is coming from the camera they own.
you've never had a legitimate program crash while working with a legitimate file? this is when malware gets to do its stuff.

>> ^mxxcon:

>> ^Drachen_Jager:
No scanner would see a program in an image, realize it's an executable and execute the program all without the user's knowledge.
It's barely possible that you could transfer a workable program that way through fractals, and the people on the other end would have to spend weeks or months decoding the program before they could make it run IF they could figure out what it was in the first place.
pattern might be constructed in such a way that when scanned it can confuse any piece of software in the chain to cause a buffer overrun, at which point they could craft the following data to actually be executable and contain malware payload.
This is kinda what happened in a few instances of MP3-based viruses. MP3 file's metadata was malformed in such a way that it would crash the player and execute trojan payload that was embedded in that MP3 file.
Look up information about MP3Concept(MP3Virus.Gen)
There was another incident that involved MP3 file played in Winamp player. I can't find link about it now.
So while realistically unlikely, it's possible.


The main difference there being that the MP3 was a digital file which had been manipulated, not a sound recording. The file was corrupted in a very clever way. In the show the bone 'picture' was taken by the forensics team. They did not bring in a file the guy had created, they took a photograph. You can't insert corrupted data, because the data is coming from the camera they own.

>> ^Drachen_Jager:

No scanner would see a program in an image, realize it's an executable and execute the program all without the user's knowledge.
It's barely possible that you could transfer a workable program that way through fractals, and the people on the other end would have to spend weeks or months decoding the program before they could make it run IF they could figure out what it was in the first place.
pattern might be constructed in such a way that when scanned it can confuse any piece of software in the chain to cause a buffer overrun, at which point they could craft the following data to actually be executable and contain malware payload.

This is kinda what happened in a few instances of MP3-based viruses. MP3 file's metadata was malformed in such a way that it would crash the player and execute trojan payload that was embedded in that MP3 file.
Look up information about MP3Concept(MP3Virus.Gen)
There was another incident that involved MP3 file played in Winamp player. I can't find link about it now.

So while realistically unlikely, it's possible.

What about metadata, metagroups, co-browsing, the opportunities for extensibility to create new browsing models?

This is definitely going to bring something different to the table, assuming they meet their goals.

And the things you can already do? This makes tab and bookmark management far less clunky. For example, new browser instances are nice but ideally something to be avoided. They occupy space and if I close Firefox I'm not sure about which tabs have been saved - yes there are extensions that help but Tab Candy (yes I'm not crazy about the name either) is less about fixing and more about starting afresh and providing a seamless, useful experience.

BTW my title was obviously phrased to get the clicks I know everyone has their preferred way of working with tabs. I'm just saying we shouldn't be afraid of change, particularly when browsers could always do with improving.


>> ^jimnms:

Does this guy just browse in one window? I start off with Blue's News in the morning, opening the different articles I want in separate tabs. Then when I click a link from one if the articles it opens a new window, I read it, then close it when I'm done. If it has a link to something else I want to check out, I open it in a new tab on that window. Sometimes I'm reading something and come across don't know about or want to know about, so I highlight it and chose search with... That opens the search results in a new tab which I tear off into a new window. I usually end up clicking the Wikipedia link, and I can never visit Wikipedia without clicking related things. Before I know it I've opened a dozen links in Wikipedia, then realize I've gotten off track, so I minimize that window to come back to later, finish reading the article and go back to Blue's.
So I'm already doing what that addon does, but instead of trying to keep it all in one browser window, I'm opening my tab groups in separate browser windows to keep them organized.

Seems like a cool idea with a certain amount of utility, but nothing that is going to take over the world by storm.

Metadata and data sharing are over-hyped ideas in my mind. Right now, I have a set of a few videosift tabs open, some "to-do" tabs with Blue's News and Fark, and some tabs looking up information on new versions of some software I use (Winamp).

I could apply metadata tags to those groups or sets, but doing so would be a completely pointless waste of time. I am the only person that cares, and if I can't simply remember why I opened those tabs in the first place without tagging them, then they couldn't have been that important to begin with.

Still, the Tab Candy idea seems to have been properly designed with the user in mind, and will have enough organizational utility to be worth trying out. Possibly even the meta-tagging and set-sharing features, in a few very specialized situations.

***Edit: cut an overly negative paragraph out.

It could be big if you think about the whole metadata thing and where that might lead.

Having something that works visually and allows grouping/notes takes it to another level. It's a more engaging way of working and a helluva lot more useful than clicking on bookmarks.

Currently I mark sites with Delicious and I know I might need to reference them if I'm working on a project or have an idea. It's useful and reassuring. But it's not really great to work with and I can't do anything of the things posited here. I wonder if adding Mind Maps and visual links to files on my system could work - you could capture all your online and offline stuff in one place.

So after a while, with all the coming extensions, Firefox is, perhaps, becoming an OS. if you add search and working with contacts you've got something that could be very, very powerful. Isn't that what Chrome and the Chrome OS is aiming for?

The other consideration is how to back this up and do it securely. I know about Firefox Synch but haven't tried it yet.


>> ^blankfist:

Cool tool. Not sure if creating "groups" of tabs is necessarily groundbreaking, but certainly useful.

Wow, did I just sound like an indignant internet nerd or what? Sigh. I've got to get off the internet.

If you had extra metadata about each video like running time the dupeof invocation could take this into account and fire a warning. Or maybe just pop up a warning anyway with the chance to cancel.

It was definitely an anti-Google rant, how it could be taken as anything else?

When they have enough detectors aimed at the real world, everything can and will be classified by metadata and stored in massive indexed databases.

My response is: so what? Draw your blinds.

i fully state that it is impossible as shown...no argument here.

the fact remains one day it could be possible somehow with a different approach... and the links i gave on sensor and recording technology are to that end.

that's the only approach i see besides some other really wacky nonsense idea about metadata/gps/cctv i have... but ive said too much!

In reply to this comment by rychan:
vairetube: Yes, there's a lot of cool research in computer vision, computer graphics, and computational photography. But it's pretty simple to say, in an information theoretical sense, that this CSI stuff is impossible.

Yes you could hallucinate plausible image content (super-resolution), but clearly that's not appropriate for a forensic setting.

But in this case we have a quantized, noisy, low resolution signal and the reflection in the eye could have been generated by any number of incident signals and it wouldn't make a difference in the recorded video signal. The information just isn't there.

I noticed something while posting vids on facebook. They have a search for YouTube videos exclusively if you want to post on someone's wall. Could someone make an app for facebook that posts through the Sift? Currently I find the good videos on VideoSift, and then switch over to Facebook, then search for the video again. It would be sweet if there was a "one-click post" or something. Also, an app like the scrolling top-15 gadget or something could be cool.

The Google-Feed-scrolling-top15 gadget: When I click on it, it directs me to videosift away from my google page, instead of opening up a new tab.

>> ^Deano:
Power up playlists. Make them searchable, give them metadata so I can find one on a given subject (there must be zillions). This would also prevent me making one that's already covered.

It would be so cool to have a running random playlist of videos. Like the video version of Pandora or Deezer.

Power up playlists. Make them searchable, give them metadata so I can find one on a given subject (there must be zillions). This would also prevent me making one that's already covered.

Honestly, the more I think about it, the more it seems gimmicky more than anything else. Much of what it shows you can already do, just with a wireless internet or WiFi phone. Yeah okay, it's nifty you can do it fully with gestures and without pulling a device out of your pocket, but I'd rather have that over having to stand in front of a wall for extended periods of times. Although it's still a nice idea, some things will definitely be more suited to this kind of interface.

The rest is predicated on the idea that it can take sensory inputs from where you are, correctly identify them and feed you back relevant information. If they can create a low cost, compact device that can do this and provide you will lots of valuable metadata, then great but I haven't seen a usable implementation of this yet and their demonstrated isn't exactly convincing enough by itself because it's obviously a simulation of what would be possible.

Lucky,

Is it possible on at least YouTube videos, to scrape the length from the metadata on the video page and get the length? And auto-tag them????????????????????????????????????????????????????????????????????

^I think it's useful because it presents complex information in a pleasing way that is transparent to one and all. The benefit is both that (dealing with the dupe process and gathering the metadata in one place) and making the sift a better place. In terms of comments alone it seems neat to seperate the dupe discussion from comments on the video.
If I see a dupe and leave a comment it might not register with the right person. I've got to start PMing people, then get a diamond involved and we have the dupe playlist. All of this made me feel we can surely instigate a process that helps flag dupes swiftly, resolves the problem elegantly, provides context for action and a history.
Overall it should be of benefit to *everyone* much like the dead pool or any other feature on the site (even though most of us don't visit that section).

Tags for this video have been changed from 'travel, map, overlay, metadata, android, cell phone, tmobile, gphone, g1' to 'travel, map, overlay, metadata, android, cell phone, tmobile, gphone, g1, google' - edited by doogle