Initial Application 3s: Fill & Download for Free

I will try explaining this in a way both tech/non technically inclined people can understand. I am also using the current Quora page as an example. Assuming all other factors (platform, bandwidth, processing power, disk speeds, memory and the page being viewed) to be equal(they should be for fair comparison), the factors that determine the browsers speed in terms of what is perceived /experienced by the user are:1. The Layout engine/Browser engine:Each browser has a layout engine under the hood. For example, the different browser engines used in browsers are:WebKit: Chrome and SafariGecko: FirefoxTrident: IEPresto: OperaThe layout engine is what decides when, what part and how HTML tags, images and CSS needs to be arranged/rendered. Basically everything seen by the user is processed by the Layout engine.The parts of this page handled by the layout engine are circled out:In the picture, the "main wrapper" in the left corner that you see is responsible for rendering everything in the page other than the search bar at the top.A layout engine affects speeds because:After the HTML text you see above has been downloaded, the order of display of different components and rendering speeds are decided by the layout engine. This varies between the engines, hence the difference in when the user starts to see the page.This engine also results in the difference in how different images, SVG files and visual elements are handled. For example, IE has limited/no SVG support, while Opera has the best support for interactive SVGs.The apparent differences in HTML5 support are also handled by this part of the browser. That is the reason this same HTML above shows up differently in older browsers like IE8.All the engines listed above are written in C++.2. The JavaScript engine:Every browser has a JavaScript engine which handles how the user interacts with the visual elements. For example, the different JavaScript engines used by browsers are:V8: ChromeSpiderMonkey/JägerMonkey: FirefoxNitro:SafariCarakan: OperaChakra: IE9The Java script engine processes how a user interacts with the page (everything happening in the background process). It executes all the JavaScript that tags along in the page.A JavaScript engine affects speeds because:When the user interacts with the page, (e.g. presses a button, types some text in the search bar etc.), the JavaScript engine processes the request, and fills the search information back into the text box.All AJAX calls, jQuery, DOM manipulation etc done in programs are also handled by this part of the browser.The way the JavaScript is handled by each of these engines is different. Due to the javascript compiler and handling of JSON data. Hence the difference in speed.For example, Apple reserves the Nitro javascript engine for Safari alone, and other developers are forced to use a different, slower engine. This is the reason why Safari will always be the fastest browser in iOS.Some engines like V8 have some critical parts written in assembly, and non critical parts in C++. All the other engines are written in C++3. Caching/compression/processing in the cloud:In scenarios where bandwidth is limited, the two factors listed above become less important. This is a game changer that makes the other two factors listed above kind of irrelevant (the code to be fetched is slower, so they are in wait for the processing to start) up to a certain level. This technology was introduced by Opera - Opera Turbo.Caching affects speeds because:The page is fetched by a server, compressed/optimized in terms of images, text, scripts etc and then sent to the user after the processing is handled by the server. This significantly speeds up the browsing experience.This scenario almost completely eliminates the need for a java script engine, while reducing the size of the page to be downloaded.Hallvord R. M. Steen pointed out some details I did not know about in the comments, so I am adding them verbatim:Opera Mini relies on server-side *rendering* - all JS execution actually happens on the server. The client gets instructions in the limited and efficient OBML language. (OBML stands for Opera Binary Markup Language). Opera Turbo's server-side *caching* is more about compressing resources, lowering image quality and so on to save data transfer.4. Local Caching:Browsers also cache some of the content like scripts, frequently accessed pages, etc for later use. This leads to faster load times, as the content is fetched from the disk instead of over an internet medium. This is the reason go is faster than reload/refresh.-Suggested by Kevin Ernest Long5. Plugin management:A typical user opens pages containing RIA (rich internet applications such as Flash, Silverlight, Java FX, Quicktime etc) which handle video, audio, games etc in the browser. The way that these plugins are handled by the browser window and how it interacts with it are different. Chrome has built in Flash, PDF support. The other browsers handle this differently. For example, I am currently using Firefox. The process associated with firefox are:Plugin management affects speeds because:Not to a large extent, but the sluggishness that results when too many of these applications are opened is because of poor plugin management.6. Memory management:Each tab is also to be given equal priority, script execution and everything else mentioned above. The memory address of these has to be kept separate so that security breaches do not occur and one tab crashing does not case the browser to crash.Firefox has 40+ tabs, yet runs as a single process with as many threads. IE and Chrome open a separate process for each new tab. This affects the speed of tabs, stability and the overall experience for the user.<update>7. Network related Optimization:Mike Belshe (who, BTW worked on Chrome and the SPDY protocol) was kind enough to point out that network related optimizations can significantly affect the speed of a browser. I had assumed that browsers are equals when it comes to networks, but as it turns out, I was wrong to make that assumption. I am quoting his comment, as it has relevance here.Oh - yes, very wrong. The HTTP stack is generally embedded in each browser (IE is the only exception). Here are some innovations that came in just the last few years with Chrome:a) DNS prefetchingThe browser learns the subdomains that are associated with a site and pre-warms the DNS lookups in a single shotb) Dynamic connection bindingIf no idle connections are available, the easy thing to do is to create a new connection and wait. The harder (and better approach is to schedule a new connection, but take either a freed up connection or the new connection, whichever comes first.c) TCP timeout workaroundsPacket loss is a fact of life. The initial timeout on a TCP connection, per standard, is 3s. Chrome saw significant wins by opening a second connection after 250ms rather than waiting a full 3s before creating a second one.d) Connection managementWhen it needs a new connection, IE now just opens two instead of one, assuming it will most likely be able to leverage the second.e) SSL OCSP cachingEach time you connect via SSL, you authenticate the server's identity. The verification can be cached in RAM or on disk for up to (generally) 7 days. Prior to about 2 years ago, most browsers didn't cache on disk, leading to 1+sec delays when opening the SSL site after restarting your browser.f) SPDYA whole new protocol.g) PipeliningGenerally pipelining has been a failure, so most browsers don't implement it, but it does make a difference if you can make it work.h) PrerenderingSome browsers are now pre-rendering entire pages in the background, assuming you'll be loading them based on learning algorithms.Google Chrome (Also Chromium) uses prerendering. More info here http://support.google.com/chrome/bin/answer.py?hl=en&answer=1385029Anyway, these are just a few that have been worked on in the past few years. Network stacks are absolutely not created equal!</update>There are many other parameters such as plugins/extensions, platform dependencies, rendering modes (open GL/DirectX) etc. however I have covered the most important ones. Browsers try to display a standard HTML page, but the reasons why different browsers exist (corporate/community), how they are developed (open/closed) and platforms they are meant for (mobile/PC/tablet) place differences in how they work. That is the ultimate difference in why different browsers have different speeds.

This may be more information than you want. ;-)HistoryThe inertial navigation system arose from the need for self-contained navigation system that did not require reference to external navigation aids. The first systems developed were analog, and used electro-mechanical integrators to perform navigational computations. This also provided for pitch, roll, and heading information to be provided to flight station instruments (in the case of aircraft), and other avionics systems.For navigation, they computed north/south and east/west distance traveled and provided a dead-reckoning position to be calculated from initial position. Early systems were not particularly accurate. This lack of accuracy was a problem for aircraft, since the small discrepancies were magnified due to the speeds and distances involved. Therefore, early systems usually were integrated with an additional sensor, such as a doppler radar, to improve accuracy.Operating PrinciplesWhen I flew in P-3 Orion aircraft, they were equipped with two LTN-72 INS, which was initially introduced in 1977. The LTN-72 is a semi-analytic gimbaled platform system, which used an internal computer.An INS has several basic components:AccelerometerStable platformGyroscopes (used to maintain the platform in a stable position)Computer (which not only calculates velocity and distance but also provides various corrections, which I will discuss below)The basic operation of an accelerometer is pretty simple - the inertia of a mass causes a displacement when the mass is moved (relative to the stable platform). The movement of the mass is detected and measured, and this translates into an acceleration. Basically, this is just a practical application of Newton’s Laws.In the LTN-72 accelerometers, torque rebalancers are then employed to move the inertial mass back to the neutral position.Three accelerometers are used (x-,y-, and z-axis) which are converted into pitch, roll and azimuth (heading) readouts.Now things get more interesting. The INS needs an initial position to enable dead-reckoning navigation. It turns out that the INS can figure out its own latitude (it needs help with longitude though). As the earth rotates, the gyroscopes (and therefore the stable platform) will detect the rate of angular rotation, and can therefore determine the latitude. Longitude doesn’t work, as there is no similar reference. However, as the earth rotates, and the unit drives the platform compensation to zero, the INS also calculates the heading to true north.There are several factors required to allow the INS to calibrate. The system needs to be at rest, and be at a stable temperature (and power available, obviously).Okay. Now that we have a system that knows its current position, off we go. And we immediately introduce additional problems.Fortunately, the system can compensate for attitude changes of the aircraft. However, as the aircraft transits over the surface of the earth, the stable platform must compensate for the curvature of the earth. No problem - the navigation computer can compensate for this (and also for altitude changes).The last major consideration is the accuracy of the platform, and the tendency to “drift”. This means that the accumulation of small errors over time can add up and become progressively worse. The immediate solution to this problem is to periodically update the system with navigational fixes (external references). This can be done in numerous ways (radar, visual, TACAN, celestial). Nowadays, it is usually done using GPS coordinates.Our LTN-72 also used an internal compensation for the natural tendency of the stable platform to oscillate like a pendulum. Using Schuler tuning, a compensation which simulates a pendulum with a length equal to the radius of the earth, the oscillation is reduced (or eliminated), which further increases the accuracy of the INS. (This is the part that may be TMI).TodayNowadays, INS systems generally use ring-laser gyros, or fiber-optic systems, which essentially measure the phase shift of the laser beams as a result of acceleration.Accelerometers have also been miniaturized, and are routinely incorporated into smartphones. This allows the smartphones to measure movement around three axes (handy for playing games and controlling stuff).Oops … forgot to add a sea story or two.The inertial alignment on our aircraft generally took about 20 minutes, so that was usually one of the first things we did on our preflight, as soon as the APU was started and we had power on the plane.I recall one time, when we were deployed to Iceland, it was so windy that the nose of the plane was being lifted, and the inertials would not align. We had to have the linemen cinch the tie-down chains tighter, to keep the plane from moving.Keflavik, Iceland is an interesting place. In midsummer, since it is so near to the Arctic Circle, it was basically light around the clock. Of course, in midwinter it was mostly dark. And, being an island in the North Atlantic, it generally experienced horrendous windy conditions. In fact, the winds are so consistently strong, that the major aircraft manufacturers send their aircraft their for crosswind landing tests.Occasionally, the aircraft don’t make it.I recall one snowy, windy night I was relaxing at our club in the Officer’s quarters. One of our P-3s was still out on a mission. When one of the USAF F-4 pilots learned that our plane was scheduled to land soon, he commented that we should not hold our breaths, since no one could land in that weather, and they would have to divert to Scotland.Moments later, we heard the drone of the Allison engines, and the familiar hum of the props being thrown into reverse pitch. They had to send a tow truck out to pull them back in - the crew was unable to taxi in those conditions.Mavi Yesil asked a question in a message, and I thought it appropriate to answer within the body of this response.As far as I understand, a Schuler adjustment is made to keep the platform perpendicular to the earth's surface. But isn't the platform perpendicular to the surface like an aeroplane during flight? Or does the platform have a stability like Gyro?The mounting surface for the accelerometers (in this case, of the LTN-72) must provide a reference plane so that the effects of gravitational acceleration can either be isolated from the accelerometers, or be calculated and subtracted from the acceleration output signal. The accelerometer mounting surface is called a “stable platform”.In other words, the stable platform must allow measurement (and isolation) of the acceleration effects due to the rotation of the earth; and then the acceleration due to the movement of the aircraft in three axes.The stable platform is controlled to define a level (and an azimuth) relative to the earth. The INS uses gyroscopes to maintain the platform level with reference to the earth. As the individual gyros detect a change in attitude, the gimbal torque motors drive the platform to the required attitude (and azimuth).Inertial systems designed with a gimbaled stable platform are subject to errors that are produced by the acceleration of the platform. A natural oscillation develops as a result of the “pendulum” effects of the platform’s gimbal joints and the stable element. A pendulum breaks into oscillation about its rotation axis when it is accelerated as a result of its mass resisting the movement of its axis.Schuler Tuning In the early 1900’s, a German engineer named Max Schuler proved that the pendulous oscillation of gyroscopes could be eliminated if the device was designed with mechanical characteristics that simulated a “pendulum” whose length is equivalent to the radius of the earth. This would ensure that the “pendulum’s” center of mass would effectively be at the center of the earth, rendering the gyroscope insensitive to acceleration. The errors produced by the natural oscillation of the inertial platform also can be eliminated by designing the platform control response using this technique.Primary reference: LTN-72 Inertial Navigation System, Orion Service Digest Issue 45, September 1987, Lockheed Aeronautical Systems Company.

The easiest change that is likely to improve cognitive performance is in the area of diet. In particular, a proper ratio of omega 3 to omega 6 fatty acids in the diet is needed for proper cortical brain function. While these benefits have been proven in clinical trials only in young children - incidentally likely accounting for part of the 7 IQ point advantage of breastfeeding, which has also been proven - they likely apply to adults as well.The ideal ratio of omega 3s to omega 6s is about 1:1 or 1:2. These days, most people get most of their dietary fat from vegetable oils which have an unbalanced ratio in the range of 1:20 or 1:30. Getting rid of the vegetable oils in favor of animal fat from pastured animals is ideal. Failing that, getting some omega 3s from fatty fish, the best source, can partially offset the poor fatty acid balance of vegetable oils.For people older than the onset of metabolic syndrome, typically in the 40s, excessive blood sugar levels and swings tends to become an issue. A reduction in dietary carbohydrate intake - less sugar and starch - can help avoid this issue. There may be some additional benefit to going to a ketogenic diet, which provides an alternative source of energy to the brain - ketones instead of glucose - especially as ketones provide more usable energy per unit of oxygen.The next easiest change is exercise. For older adults - again, roughly those of us past the onset of metabolic syndrome - exercise is useful, probably because it helps burn off excessive blood sugar levels and stores. Anaerobic exercise is especially effective since it burns blood sugar less efficiently, and thus uses more of it.Then, there are changes in intellectual habits. In particular, it’s likely worthwhile to do a lot of reading. This is partly because reading builds vocabulary, which contributes to most peoples’ definition of intelligence and also is an element in many IQ tests. In addition, though, critical reading - thinking about what is read and its implications - helps one practice logical and rational analysis. Don’t worry about reading things that you think are good for you; just read what you enjoy enough to think about analytically. Find quiet, undisturbed time to do your reading.Learning math through calculus is probably also useful. Math at these levels - arithmetic through algebra to calculus - is applicable to all sorts of everyday issues. Do you have enough money to make it through payday? Arithmetic. What combination of potluck dishes do you need to make sure everyone has enough main course and dessert at that party? Algebra. Can you make it through the yellow light before it turns red? Calculus. I’m not saying you should break out the pencil and paper and write down equations for all these things, but if you’ve fully learned the math, you should be able to apply the concepts to these and many other everyday problems, even if your goal is not an exact numerical answer.Edit: Finally, there are other people. The more time you spend in the company of people smarter than you are - but not so much smarter you can’t understand them - the more opportunity you have to learn things from them, including thinking habits. Perhaps this one should have been first.Don’t expect an overnight change. Even babies take years to demonstrate the benefits from that initial year of breastfeeding. Keep up the good habits year after year, though, and you’ll likely see some benefits.