<?xml version="1.0" encoding="UTF-8"?>
<!--Generated by Squarespace V5 Site Server v5.13.159 (http://www.squarespace.com) on Sun, 26 May 2013 06:06:07 GMT--><?xml-stylesheet type="text/css" href="/universal/styles/feed.css"?><rss version="2.0"><channel><title>High Scalability - Comments</title><link>http://highscalability.com/blog/</link><description></description><copyright></copyright><language>en-US</language><generator>Squarespace V5 Site Server v5.13.159 (http://www.squarespace.com)</generator><item><title>Charl comments on Stuff The Internet Says On Scalability For May 24, 2013</title><author>Charl</author><pubDate>Fri, 24 May 2013 20:37:47 +0000</pubDate><link>http://highscalability.com/blog/2013/5/24/stuff-the-internet-says-on-scalability-for-may-24-2013.html#comments</link><guid isPermaLink="false">426227:4867632:comment/20042869</guid><description><![CDATA[<p>When I follow the link about 100 million hours of uploads per minute on youtube, the article says only 100 hours per minute. Still very impressive, but a massive difference...</p>]]></description></item><item><title>Uli comments on Stuff The Internet Says On Scalability For May 24, 2013</title><author>Uli</author><pubDate>Fri, 24 May 2013 18:50:56 +0000</pubDate><link>http://highscalability.com/blog/2013/5/24/stuff-the-internet-says-on-scalability-for-may-24-2013.html#comments</link><guid isPermaLink="false">426227:4867632:comment/20042670</guid><description><![CDATA[<p>I think it is &quot;just&quot; 100 Hours uploaded on Youtube each minute :) Still crazy ;)</p><p>Best Regards,<br/>Uli</p>]]></description></item><item><title>Lawrence Kesteloot comments on Stuff The Internet Says On Scalability For May 24, 2013</title><author>Lawrence Kesteloot</author><pubDate>Fri, 24 May 2013 18:12:00 +0000</pubDate><link>http://highscalability.com/blog/2013/5/24/stuff-the-internet-says-on-scalability-for-may-24-2013.html#comments</link><guid isPermaLink="false">426227:4867632:comment/20042591</guid><description><![CDATA[<p>That&#39;s 100 hours, not 100 million hours (YouTube). What does it say that I originally believed the numbers?!</p>]]></description></item><item><title>vdp comments on Stuff The Internet Says On Scalability For May 24, 2013</title><author>vdp</author><pubDate>Fri, 24 May 2013 16:05:30 +0000</pubDate><link>http://highscalability.com/blog/2013/5/24/stuff-the-internet-says-on-scalability-for-may-24-2013.html#comments</link><guid isPermaLink="false">426227:4867632:comment/20042260</guid><description><![CDATA[<p>s/100 million hours/100 hours/</p>]]></description></item><item><title>Jason Gordon comments on The Tumblr Architecture Yahoo Bought for a Cool Billion Dollars</title><author>Jason Gordon</author><pubDate>Fri, 24 May 2013 15:06:13 +0000</pubDate><link>http://highscalability.com/blog/2013/5/20/the-tumblr-architecture-yahoo-bought-for-a-cool-billion-doll.html#comments</link><guid isPermaLink="false">426227:4867632:comment/20042031</guid><description><![CDATA[<p>Just want to say thank you to Blake and Todd for taking the time to deliver this article. I got some validation out of the tumblr interviewing process. I believe irrelevant brain teaser gauntlets can intimidate and discourage potentially solid developers. Good to know other people doing real work think that way too!</p>]]></description></item><item><title>DB comments on Strategy: Stop Using Linked-Lists</title><author>DB</author><pubDate>Thu, 23 May 2013 17:57:37 +0000</pubDate><link>http://highscalability.com/blog/2013/5/22/strategy-stop-using-linked-lists.html#comments</link><guid isPermaLink="false">426227:4867632:comment/20039898</guid><description><![CDATA[<p>It&#39;s funny that there&#39;s so many people butthurt over the suggestion that Linked Lists aren&#39;t performant. Probably because none of them have ever designed a system at scale. </p><p>Age and pimple (or pimple free) has nothing to do with how good the idea is, and in this case, the author is correct. </p><p>LInked lists have a ton of pointers, and a ton of them are bad for the branching lookahead units. ArrayLists are always in contiguous memory, and pay an amortized growth penalty. </p><p>It&#39;s not hard to code up an example that thrashes both array types from multiple threads on multiple cores. Spend the 5 minutes and look for yourself.</p>]]></description></item><item><title>Isuru comments on Start Here</title><author>Isuru</author><pubDate>Thu, 23 May 2013 12:21:11 +0000</pubDate><link>http://highscalability.com/blog/2007/7/6/start-here.html#comments</link><guid isPermaLink="false">426227:4867632:comment/20039045</guid><description><![CDATA[<p>Almost 3 years of experienced Software guy here. But I&#39;m fairly new to all things web but I&#39;ve always been amazed and intrigued by how huge sites operate and everything. Although I don&#39;t understand (yet) half of the information in these articles, I love reading them. Hopefully one day, I&#39;ll be able to make my own contributions as well.</p>]]></description></item><item><title>JB comments on Strategy: Stop Using Linked-Lists</title><author>JB</author><pubDate>Thu, 23 May 2013 12:19:07 +0000</pubDate><link>http://highscalability.com/blog/2013/5/22/strategy-stop-using-linked-lists.html#comments</link><guid isPermaLink="false">426227:4867632:comment/20039042</guid><description><![CDATA[<p>It&#39;s good advice that many people don&#39;t seem to have yet got and is worth repeating.<br/>There are few situations where the performance of linked lists is better than a simple vector. Even the ones that it&#39;s supposed to be good for like inserting at the start a vector will often outperform a list.</p><p>The main time I use them is when I have to splice together multiple lists quickly. It still seems to have some advantage there,.</p>]]></description></item><item><title>Andreas comments on Grace Hopper to Programmers: Mind Your Nanoseconds!</title><author>Andreas</author><pubDate>Thu, 23 May 2013 11:22:29 +0000</pubDate><link>http://highscalability.com/blog/2012/3/1/grace-hopper-to-programmers-mind-your-nanoseconds.html#comments</link><guid isPermaLink="false">426227:4867632:comment/20038958</guid><description><![CDATA[<p>|Please correct this sentence:<br/>|&quot;... a nanosecond, a billionth of a second, which was the speed of their new computer circuits.&quot;<br/>|<br/>|(because, of course, the nanosecond is a unit of time not speed)</p><p>Why? &quot;one billionth of a second&quot; equals 1 GHz, which is the clock speed of (still some) CPUs</p>]]></description></item><item><title>Vomitorium comments on Strategy: Stop Using Linked-Lists</title><author>Vomitorium</author><pubDate>Thu, 23 May 2013 10:08:12 +0000</pubDate><link>http://highscalability.com/blog/2013/5/22/strategy-stop-using-linked-lists.html#comments</link><guid isPermaLink="false">426227:4867632:comment/20038866</guid><description><![CDATA[<p>OK, interesting, didn even thought about that.<br/>But, all modern frameworks &amp; APIs (and yes, IDEs) and laguages make it *fairly* easy to use linked-lists, because all functions are returing usually some type of linked list :-(</p>]]></description></item></channel></rss>