What I really like is when you get random J's in emails from people.

Took me a while to figure out this is some kind of autoreplace thing in Outlook that switches out emoticons with characters from the wingdings font, and J is a smiley face.

Visibility of the source code is a side-show in electronic voting systems. Even if the source code is published, there is no way to be sure that that is the code that is running on the hardware, or to be certain that the hardware itself has not been tampered with. Votes need to be printed out on paper, verified by the voter, and counted by hand.

Still, when we had the source code for the Irish system (now abandoned due to our efforts) analyzed by a commission, it was found it had actual counting errors.

http://www.stdlib.net/~colmmacc/www.cev.ie/htm/report/part4_...

Amazing!

(Note: I used to be employed by Mozilla, and in that capacity I was the owner of Mozilla's image decoders. I've been disconnected from all decisions for almost a year, though.)

The main take-home here is that while Google's numbers all show WebP as being objectively better, the metrics they chose for comparison were relatively bad (i.e., some of them didn't take into account colour or didn't model colour correctly), and once you accounted for that the numbers were not nearly as good a story for WebP; in some cases, JPEG outperformed it.

The facts that (1) WebP was not terribly compelling technically, (2) JPEG is already supported by everything on the web, not to mention devices and mobile phones etc, and (3) there's still headroom to improve JPEG in a backwards-compatible way, meant that WebP was (and, it seems, remains) a non-starter.

Personally, I wasn't too concerned with Google knowing my actual identity, but more with the prospect of tons of information leaking to anyone who I ever meet. For example, it's no longer possible to anonymously review or rate android apps. They retroactively changed all older app reviews to be by "An anonymous user", but now all reviews and ratings are done under your Google+ account, so I just stopped reviewing and rating apps.

If I were to rate and review apps as I would prefer to do, then anyone googling my name when they meet me would find my Google+ account, then it would be trivial for them to click around and find out essentially a list of all the apps I use on a regular basis, and how I use them. I don't know that there's anything too personal at the moment, but given that that information is essentially out there forever, I don't want to accidentally rate some app that might reveal my political, religious or cultural affliations (which could bias a future employer or landlord), or something else that might be very telling about my personal life.

For me, Google having a huge trove of information on me only matters to the extent that they'd ever show it to anyone else. I don't care what they do with the data internally, I've disclosed it to them, but it's clear after the Snowden documents that Google doesn't have the ability to fight government intrusion into their databases (making their huge troves a very tempting target), and after the fiasco with Google Buzz (and this nonsense with Google+ and the "real names" policy), it's become obvious that they don't seem to have the sort of mindset that is appropriate for a good steward of my private data. It's sad, because they provide many great products, but they are simply not trustworthy - as that becomes more obvious to more people, I imagine it will start to affect their bottom line.

Your obvious rival here is PCPartPicker. Having used that extensively, here's some things I don't like about yours:

Expert mode removes compatibility checks. This shouldn't be the case - as soon as I select an item, it should cull the results and show me only the intersection of items that are now compatible with my chosen piece of hardware.

The lack of an ability to filter. Maybe I want to splurge more on a graphics card, but I need one that has at least 4 mini-display port outputs to drive my monitors. Right now there doesn't seem to be a way to say "show me only gpus that are nvidia, above $500, and can drive at least 4 monitors".

One of the biggest things for me: lack of reviews. You have a pangoly score. That's great and all, but you're a brand I don't (yet) trust. If I'm buying a component that costs multiple hundreds of dollars, I want more than a single score from a company I haven't heard of. I want to know what the reliability is like, and I want to hear it from people who have had hands on experience with it. I want to know how loud the fans are, what temperature it runs at passively and actively, etc.

Visually your site is really nice, but your expert mode leaves a lot to be desired, and I think you have a ways to go before you catch up with the competition. If I were you I would try and find your niche - either target users who are switching over from console gaming, and want to build their first computer, or build out a bunch of more expert features and target regular hobbyists.

All asymptotically fast algorithms for integer multiplication basically reduce the problem to one or several polynomial multiplications.

Polynomial multiplication can be done in subquadratic time using an evaluation-interpolation strategy: namely, to multiply two polynomials of degree less than n, we evaluate both polynomials at 2n points, compute 2n pointwise products, and recover the polynomial product by interpolation.

When one chooses roots of unity as the evaluation points, the evaluation and interpolation steps reduce to discrete Fourier transforms (forward and inverse, respectively), which can be done in O(n log n) arithmetic operations by means of the fast Fourier transform (the pointwise products require O(n) arithmetic operations).

Now, there are a couple of difficulties. How do you decompose the integers into polynomials? Is it better to use short polynomials with large coefficients, or long polynomials with small coefficients? Since the coefficients asymptotically grow with the length, you have to account for the bit complexity of the arithmetic operations in the FFTs and pointwise products (here you get recursive integer multiplications).

Which FFT algorithm should you use (there are many to choose from)?

Also, the ring of integers does not have enough roots of unity to do FFTs, so you have to lift the computation to a larger ring, such as the field of complex numbers (with numerical approximations), an appropriate finite field, or (as in the Schönhage-Strassen algorithm) a polynomial extension ring. Which choice gives the best complexity?

The improvements to integer multiplication published in the last few decades have basically been tightened analyses of such tradeoffs. I have not read this paper in detail yet, but it seems to provide a clear description in the introduction. Quote:

The idea of the new algorithm is remarkably simple. Given two n-bit integers, we split them into chunks of exponentially smaller size, say around log n bits, and thus reduce to the problem of multiplying integer polynomials of degree O(n/log n) with coeffcients of bit size O(log n). We multiply the polynomials using discrete Fourier transforms (DFTs) over C, with a working precision of O(log n) bits. To compute the DFTs, we decompose them into "short transforms" of exponentially smaller length, say length around log n, using the Cooley-Tukey method. We then use Bluestein's chirp transform to convert each short transform into a polynomial multiplication problem over C, and finally convert back to integer multiplication via Kronecker substitution. These much smaller integer multiplications are handled recursively.

I've worked with IBM.

Your post all seems to be very consumerist. IBM sells servers. Apple does not sell servers. IBM sells their own server hardware. IBM dropped their x64/x86 servers recently.

Apple makes consumer hardware, IBM makes expensively marked up enterprise hardware.

IBM makes TERRIBLE software, apple makes shiny consumer stuff, the absolute most that will come out of this is sametime and or notes getting a small bit more attention from IBM on iOS products. Most of IBM's crap is a thinly veiled customisation on top of Eclipse, it makes me laugh thinking about how that will translate to an iOS device. IBM's server side software, while marginally better - what apple hardware is that going to run on?

IBM clients couldn't give less of a hoot about samba shares in OSX. This news will have approximately zero effect.

I've been using Tree Style Tabs[0] with Firefox for a long time. Being able to nest tabs and then collapse them is awesome. It is hard to live without it when using Chrome (I know there is a similar addon for Chrome[1], but it doesn't have the same feel since it just adds a new window off to the side and still has the tabs across the top).

[0] https://addons.mozilla.org/en-US/firefox/addon/tree-style-ta... [1] https://chrome.google.com/webstore/detail/sidewise-tree-styl...

Ha, what a disaster G+ has to be for Google. They must kind of be looking at it like "ew what do we do with this thing now?"

They should've taken a more holistic and serious approach to bringing your offline identity online. Instead they tried to copy FB. They should've just called it Google Identity(with the same level of seriousness as Youtube's copyright protection product: Content ID) and provided real-life perks with associating say your Driver's license number with it which maybe would allow me to renew my driver's license much more quickly. I don't know. G+ is just weak.

I mean at the end of the day, they DO have your identity. I'm sure their algorithms could within a certainly small margin of error determine who I am to a reasonably high degree of probability based on nothing other than lines in a bunch of log files(DoubleClick Ad requests on various sites, YouTube video watch times, Google Searches based on time of day/location, mobile usage, etc); again, regardless of where or when or how I use Google(i.e. using igonito browsers, using a friend's phone, using a computer an internet cafe, etc.). So why not throw all that information under one umbrella, Google Identity, and provide it to me once I claim it.

They're likely able to treat online "behavior" as essentially unique online fingerprints and can just associate that with individual users; that is, they don't really need uniquely identifying data(unique IDs, usernames, etc.), they just need to watch your online activity(regardless of if you're logged in or have identifying cookies, etc.) for a while in order to identify you.

I assume it's because consumers probably aren't ready to know just how much data Google actually has on them AND how much information can be extracted from said data. It's likely pretty accurate and pretty scary.