mjschultz

Google, NYT, Amazon, et al. have data caching and it conceptually works in the same way as the idea of named data networking (on of the names of the idea above).

However, named data networking would push a cache to every network device between the requester and the provider for every piece of data. Unless an attacker is very close to the provider, they won't be able to take down the resource. The best they could do is take down the resource for themselves and their nearest neighbors.

Special Sandboxing that, I think, is implemented in Java. So this is an OS in the sense that it only depends on having a JVM and Lua interpreter installed.

Not that this is a great metric (especially when the source languages are taken into account), but here are the LOC counts in the project:

    $ find . -name '*.java' | xargs cat | wc -l
    28877
    $ find . -name '*.lua' | xargs cat | wc -l
    6826
    $ find . -name '*.slua' | xargs cat | wc -l
    1725

This is precisely the reason I stuck with GoDaddy for so long. My first domain was registered on 07-Sep-01 and, at the time, GoDaddy was the cheapest registrar I knew about/could find. I think this was even before they really started pushing their "edgy"/sexist advertisements.

Since then I haven't really had a compelling reason to leave, though earlier this year I decided to--as my domains expire--move to name.com. Of course as of today I initiated the transfer processes on the remaining 12 domains.

It's a win-win for me.

I don't have a Kindle or other ebook reader, so I can't speak from experience but are some publishers better than others?

I mean, there are many very bad physical book publishers out there that don't (or minimally) edit the copy sent to them before sending it to the presses. That translates into the digital realm too, but there is a much lower barrier for entry for bad publishers and it's hard to differentiate between bad and good publishers in these early days.

If no publisher out there is doing a good job, it might be something to invest time into and maybe make some money.

I think Dropbox is considered SaaS (and doesn't precisely fit your definition of "web application"). I would put S3 into PaaS as it boils down to a storage platform.

Right now I agree with saying IaaS is just a VM service, but I'd hesitate because there may be future changes that invalidate that connection.

Keep in mind these terms are provider-centric not customer-centric. For example, Amazon provides their networking and computer infrastructure as a service to you the customer. You realize this service as VMs. Similarly, GAE provides a platform for you to use, and you realize this as a place to run your cloud application. And to finish the triad off, SaaS provides customers with software and they realize it as an application. So it somewhat depends on if you are looking at these things from a customer perspective or a provider perspective.

I believe this NIST document provides fairly clear terminology: http://www.nist.gov/manuscript-publication-search.cfm?pub_id...

> "Cloud" is what marketing and non-technical people use. I'm ok with that.

I would assume this means he'll use the general "cloud" term when talking with non-geeks.

Technical people should know that [SPI]aaS are specific parts of the "cloud." Within technical communities, the correct terms should be used to avoid ambiguity. It's about knowing your audience and making sure they can relate to you.

> I'm not saying that you wouldn't be able to do this with a web-based mail client.

In fact, using gmail should take fewer strokes than that. Assuming you have a keyboard shortcut to spawn a gmail window, enter will take you to the first message, and 'e' will archive that message. Command-Q/Alt-F4 to quit.

Labels can be applied with an 'l' and the label name, next/last message is 'n'/'p'.

The full list of keyboard shortcuts is available here: https://mail.google.com/support/bin/answer.py?answer=6594

It would, of course, be better if they let you customize the shortcuts and sped up the loading of the initial page.

My curiosity side agrees with you, but my practical side does not.

> He says passwords were "hashed and salted." This could be anything from the naïve MD5(pass+salt) to the more secure bcrypt or PBKDF2. Now, I have every reason to believe that Valve is smart enough to not use methods like the first, but information is always welcome in a scenario like this.

Do these details actually matter to you, a technically savvy user? If they told you they used bcrypt or PBKDF2 for their password hashed and salted passwords would you think to yourself: "oh well, in that case I don't need to change my password" or are you going to take the few moments and change you password anyway? I'd probably just change my password.

For even less savvy users, they're getting technical details that they don't really care about now. Depending on where those details are in the message they might miss important bits of useful information.

I suppose if they said it was MD5(salt:pass) and you used the same password for steam and something else you might have reason to be concerned, but probably not unless they are targeting you specifically.

That said, I think any company should provide a link to a blog that does dig into the important technical details for people that want to know. And keep updating it as new information is found.

In what way is this table layout not that accessible? I just fired up Voice Over and it read the page in the expected order. (Heading, left text, right text, chapter list.) The only complaint I would have is the bad alt text on images in the chapters (but every site has that problem it seems).

The font size can be increased without much issue (again, no more than most sites).

* I'm not blind or hard of sight, so I'm not an authority on these things. Just curious.

Indeed, my first thought on seeing all the options was: how could I game this?

For example, if I live on the east coast (UTC-5) and select the lowest-cost option (3 hours of use from 1am until 9am for $20), but claim my timezone is the Alaska (UTC-9) I believe I could actually use the service from 5am until 1pm in the east coast.

I don't know if that's how it would actually work, but it certainly is confusing.

That said, if you look at the academic pricing plans the presentation is much clearer and easier to understand.

Does anyone know if there are tools out there that look at a code base and find "similar" snippets of code?

I would expect such a tool to parse the language into AST form and find branches that are the same except some identifiers and a few other details. It is probably intractable in general, but I think it is feasible for most code bases.

That doesn't say that Amazon is their CDN at all. They probably use AWS for pages and user-oriented stuff, but content delivery is done through Level 3 [1].

[1] http://www.reuters.com/article/2010/11/11/us-levelthree-idUS...

Their own site has the same pictures that are on the original post: http://www.raspberrypi.org/?p=78

Looking through the media section says that the picture you have is a prototype board, not the manufactured alpha board they're talking about here.

I'm inclined to agree with this fraction. I just used POP3 to LIST the first 360 messages in my mailbox and the graph shows an distinct upshot around 95%.

Around 95% of my (360) emails consume 6.5% of the space (360 messages take).

It's been a while since I used an IDE (instead of vim), but don't most of them support code-folding that provides a light gray bracket for a block of code (usually with a collapsible indicator at the top)?

Or is the effect you want slightly different?

You don't need to give them your password or phone number, just the email address associated with your account(s). Adding a phone number would be optional.

Unfortunately, this is very x86 device dependent and most embedded devices are not x86.

Yeah, I was planning on writing something more about NAPI and GRO at the beginning of the next part, though I'm not sure there is terribly much to GRO. I hadn't thought about zero-copy yet though, but it is certainly important so I'll probably try to fit something like that in.

Thanks!

I'm sure it isn't 100% accurate (Linux has some complexities to it after all), but I welcome any constructive criticism. I want to improve my low-level understanding before forging ahead to the higher level details that I'll write about in the next few weeks.

I would not be surprised if Maurice Herlihy managed to get a Turing Award in the next few years.

I would be surprised if Stallman got a Turing Award however. I can't think of anything beyond his philosophy that makes him stand out in any technically innovative way.

I can't explain why double the logarithm of the number of passwords stored would be the target. But it does make sense to have at least log_2(n) salts for n passwords. That would guarantee each password, even if they were all identical, would have a unique salt making every password hash differently (at least theoretically).

Now perhaps you would double that as a just-in-case scenario or because there is some fundamental property of every (existing or future) hash algorithm that causes a collision half the time (i.e. both salt 10001 and salt 10000 generate the same hash value).

Precisely what I was thinking. Does anyone know if there is a reason Intel hasn't just acquired ARM? I assume Intel believes that they might still be able to compete and win in the market, but that seems to be less true every day.

(I assume the answer is that it would raise a bunch of anti-trust flags everywhere, but I'm not positive about that.)

Perhaps, right now I'm in a condo so I still don't have to worry about most of those things. But I still think just having a choice between hiring someone competent and doing it myself is a nice one.

I stress competent because my previous landlord(s) were less than able to fix things beyond the superficial problems. Their goal is to strike a balance somewhere between quality and ability to stop complaints. Usually the choice fell one one side of the ability to stop complaints side. (Fix it so it works, not so it works well, i.e. replace broken washing machine with cheap almost broken washing machine.)

Hmmm, my coffee must not have kicked in at that point. I skipped right over the "mouse pilots" part and just assumed "mice."

It seems entirely plausible that they have exposed mice to backscatter scanners for extended periods and studied the reactions/effects. That study I would be interested in, but it seems rdtsc was making a funny. I understand now.

I absolutely agree with the satisfaction of repairing something yourself. In fact, I think even having the choice of either repairing it myself or hiring someone I deem competent to repair it is a worthwhile difference.

When I was an apartment dweller, I signed a 1 year contract and they kept telling me that if anything breaks call them. That included windows, pipes, stoves, refrigerators, and light bulbs. I wasn't even supposed to change a light bulb if it burned out on me. Admittedly this is only one experience with what I hope was a very poor renting operation, but they owned all the buildings in the vicinity.

I was so happy when I replaced my kitchen sink faucet on my own and in a way that satisfied what I wanted.

Can you provide a reference for this study please?

A few months ago I read: "The Impact of Virtualization on Network Performance of Amazon EC2 Data Center" (http://www.cs.rice.edu/~eugeneng/papers/INFOCOM10-ec2.pdf) which has some performance numbers (latency and throughput) for small and medium instances.

Unfortunately, I can't recall enough of the paper right now to give you the nickle-and-dime overview, but it has graphs you can look at!

I'm want to agree with you, but I have a question.

Binaries tested against what? Obviously the testsuite that upstream uses, but every program will have dynamic dependencies into the distro. In that case wouldn't the upstream maintainers have to test every version or every distro to make sure it works for you? That seems time intensive (not infeasible, but it would become chaotic very quickly when we multiply the number of distros by the number of apps people want).

I'd also presume that the upstream maintainers include their testsuite with their source, so nothing should prevent the downstream distro maintainers from running the testsuite before shipping (I have always assumed they do, I hope I'm not wrong).

Just to let you know, I experienced the same thing with editing as GP described using Firefox 4.0b5 on OS X 10.6.

I also used Xinu in my O/S course. I'll mention that one of Doug Comer's (http://www.wikipedia.org/wiki/Douglas_Comer) intentions with Xinu was to purposefully make a different design decision than Unix when pedagogically relevant. For example, instead of treating everything like a file (Unix/Linux) everything is treated like a device.

Its various iterations over the year do not necessarily remain true to that principle, but it is something different about the O/S.