I agree an email address is intrinsically easier to remember to a human, but it has a huge flaw I experience all the time. Ever try to give your email address out over the phone, or any combination of unfamiliar letters? 'V' gets confused with 'B' and so on, especially when you have a unique spelling for a a name. My email address and my first name have been malformed a number of times by humans over the phone, but my phone number not once. Numbers are just easier to convey and less ambiguous, I always wondered if by design?

That being said there are technological solutions to the problem, when I meet someone in person I should not be verbally relaying my address (phone number or email...) we should be doing some digital vcard exchange over Bluetooth or something between our phones. Over the phone I should not be verbally relaying information that is more clearly conveyed in text. When ordering my air plane tickets over the phone (sometimes the human operators can pull off things that the online interface is not letting me do in booking) I should simply be able to switch to instant/text messaging the operator and clearly relay any text as long as I do not hit the wrong key on my damn virtual keyboard...

Thanks, someday I hope grow up to be a real DBMS hot dog!
What about Robert Johnson's claim that there is no clear segregation? Is he wrong?
What is the problem, Facebook should be even easier with a real DBMS and scale very nicely.
The only problem I see is over time an individual's profile expanding, like saving all your emails form the last 10 years.
But computer hardware advances or archiving the older less accessed stuff off to another server will help.

I want to take a crack at this, I know enough about databases to make a PHP MySQL web pageis all! How can there be no clear segregation? The data is about me, about someone else, or about a group (a group being a single entity on Facebook similar to an individual) and the data either came from me, someone else, or a group.

My inbox in Facebook is no different than my email inbox, all the messages are to me, just like my email inbox it should reside on one search-able database restricted to a certain manageable size like every other email inbox in the world. My inbox can be on one server and my friend's on another, the whole thing should be segmented by user. Same with my outbox, just like email I retain a copy of sent messages in my personal Facebook database. Wall posts and pictures should work the same way, anything that shows up in my profile should be copied to my personal tables and database.

Now the trickiest part is when another user posts a picture and tags me in that picture. That picture reference should then be duplicated and placed in my picture database. This is in contrast to retaining one copy stored in a Facebook wide database and searching that database for pictures of me each time someone wants to bring up my pictures. The picture data storage can be spread across multiple servers and when someone views the my pictures section of my profile there is simply a dump done on my picture database of references to the pictures to present, links to the pictures and the pictures are retrieved from storage and displayed. When the owner of a picture deletes a picture or the owner untags me or I untag myself from a picture, the picture would simply be deleted from my picture reference database.

I am going to go watch Robert Johnson's ACM talk cause this seems easy and the segregation is as clear as good ole email. From what I can see each individual's profile is fairly small in database terms and never viewed all at once. A profile is viewed by wall, inbox, pictures (and not even all the pictures at once), info...

I am amateur at this at best, but with my hack mySQL skills I do not see it as a big deal to create an easily scalable Facebook segmented by users.

What type of chat room?

Yeah but I am the exception to the morons, I can actually drive down the interstate at 70 and text.

Yeah, after taking an AI course in college and building robots in graduate school I was no longer afraid of movies like the Terminator and the Matrix. Although my weed cutting robot did cut the crap out of the PhD's hand and my JAVA based GoMoku playing AI, using alpha beta pruning trees, still beats me most of the time. OK, I live on the second floor, with some liquid steel, in a dark place (that whole Matrix block out the sun, humans make good batteries for the AI, does not make sense to me).

I am glad you brought this up! I have been worried about this. I want my cable, DSL, fibre and cellular ISPs to be dumb pipes as much as the next Slashdotter, but I worried about the quality of service when I needed it for making a VoIP call while my neighbors P2P has the same priority. I do not work in networking so I am hoping for your expert feedback on my understanding and ideas.

I agree that "jitter-sensitive", like VoIP, data should have priority in high traffic situations over less sensitive traffic like Bittorrent traffic. But I am not convinced that automated network management systems are where I want the decision to be made. Plus I may feel my Bittorrent traffic or what appears to be less sensitive traffic to you or the system is important and do not want to leave it to the automated network management to decide what to prioritize. I would bet net neutrality is a pipe dream for wire based networks and essentially impossible for wireless networks. I do not see why we cannot achieve the overprovisioning fantasy-level of service on the ground, but agree it is most likely a waste of resources. But in the air, it seems today at least, there will always be very strong technical reasons for why people cannot just have all the bandwidth they are willing to pay for, it is not just artificially limited by politics and the lack of infrastructure spending, but by the natural limitation of spectrum real estate. That being said I think a tiered internet service will be necessary even if net neutrality is implemented. No, not the evil tiered plans we hear with service providers like Google paying, the other way around.

First off I will make the assumption that there are four major classes of internet traffic types: low-latency high-bandwidth (video conferencing, remote desktop today, 3D video and smell streaming tomorrow), low-latency low-bandwidth (VoIP, gaming, live chat and collaboration), high-latency high-bandwidth (bittorrent, one way video streaming), high-latency low-bandwidth (text, and simple web pages). Maybe the latency factor would be replaced with reliability or quality, some quantity that measures the jitter factor. Now this is not to unlike how I see the current mobile industry today: messaging (low-bandwidth, high reliability), voice (mid-bandwidth, mid-reliability), data (high-bandwidth, low-reliability). I like it this way to a certain extent. I know someone streaming YouTube videos will not delay my voice calls. But what is disappointing is that I am stuck with my carriers voice service, I cannot use my own VoIP application, service, or protocol. With net neutrality would all three tiers be smashed into one? Would the YouTube videos get streamed with the same priority as my VoIP? That is what it seems would happen, and you and I agree this is not a good idea either. But do I not want the dumb pipe analyzing my packets and deciding for itself what priority to give them? No, what if it judges my VoIP not as important, either intentionally and maliciously to push the carriers own voice service, or just unintentionally. Or what if I really want to stream that YouTube video down, NOW. Well then I let the carrier know what priority or tier to place the connection in, and I pay based on free market demand for that tier. The ISP then has no choice but to treat it like any other packet in that tier, whether it be voice or P2P or streaming video, I just pay accordingly. If the low-latency tiers get clogged up they boost the price for using them and hopefully get rid of the YouTube streamers or P2P traffic costumers that can wait. This also reduces the overhead on the ISP's side as well with packet inspecting equipment and cost.

Sorry I am not going to let you get away with raining on their parade so easily. I am hoping you have some BIOS and firmware expertise or insight to share and advance my knowledge and back up your claims.

Why are UEFI, OpenFirmware and the like exempt? Why would a unified or open standard for a firmware not make it easier, since it could possibly remove your first objection of requiring specific payloads?

Sure these more modern firmwares may be intrinsically more secure. But the UEFI group it self admits that the BIOS and firmware are not completely going away. Perhaps it will be enough to render such exploits void, but I would not be so quick to claim that.

Malware is all about being persistent and difficult to remove. We are talking about an exploit that gets as close to the holy grail of malware, infecting the hardware, as virtually possible. The only deeper exploit would be hard coded malware in the hardware.

I, myself, am quick to try and deflate someone else's party balloons in my own arrogance for sure. But come on, have some humility to admit this is at least interesting. Maybe you are not worried, but it is impressive. It at least has peaked my interest to take a low level look at BIOS and motherboard firmware again.

The new fragrance by Calvin Klein.

Hehe! I do not comment much sorry, did not realize I could use html tags...