<?xml version="1.0" encoding="UTF-8"?><rss version="2.0"
	xmlns:content="http://purl.org/rss/1.0/modules/content/"
	xmlns:dc="http://purl.org/dc/elements/1.1/"
	xmlns:atom="http://www.w3.org/2005/Atom"
	xmlns:sy="http://purl.org/rss/1.0/modules/syndication/"
		>
<channel>
	<title>Comments on: Accessible Celebrities</title>
	<atom:link href="http://uwmike.com/articles/2005/04/21/accessible-celebrities/feed/" rel="self" type="application/rss+xml" />
	<link>http://uwmike.com/articles/2005/04/21/accessible-celebrities/</link>
	<description>Hacker, Engineer, Dancer, Gentleman</description>
	<lastBuildDate>Mon, 05 Sep 2011 13:32:22 +0000</lastBuildDate>
	<generator>http://wordpress.org/?v=2.9.2</generator>
	<sy:updatePeriod>hourly</sy:updatePeriod>
	<sy:updateFrequency>1</sy:updateFrequency>
		<item>
		<title>By: Mike Purvis</title>
		<link>http://uwmike.com/articles/2005/04/21/accessible-celebrities/comment-page-1/#comment-3439</link>
		<dc:creator>Mike Purvis</dc:creator>
		<pubDate>Sat, 23 Apr 2005 18:53:33 +0000</pubDate>
		<guid isPermaLink="false">http://dev.uwmike.com/articles/2005/04/21/accessible-celebrities#comment-3439</guid>
		<description>In &#039;Beating the Averages&#039; and a few of his other essays, he&#039;d talked about how the Viaweb Store used flat-files and directories instead of databases. He even said that when a competitor advertised for hackers with Oracle experience, they immediately considered them no threat at all.

I asked him if he felt that relational databases had *no* application at all, or if it&#039;s just that they&#039;re over-deployed in situations where 99% of the queries look like &quot;*SELECT * FROM table WHERE id = &#039;x&#039;*&quot; (sites like this one, for example)

He said that he felt they are not necessary at all, since they were originally devised to a) not lose your files and b) allow quick retrieval. Since modern OSes and filesystems don&#039;t lose your files anyways, particularly if you&#039;re running something like RAID-5 on your server, plus they do a lot of caching and so on for you, he said that that&#039;s adequate on its own.

Of course, you need to know what you&#039;re doing as far as creating indexes and so on, but that&#039;s not rocket science... it&#039;s just that a relational DB normally does a lot of the heavy lifting for you.

I&#039;m not sure I completely agree, since there&#039;s a lot of sophisticated queries easily possible with SQL that could take a fair bit hassle with files and directories. Also, if you&#039;ve got multiple applications all trying to access the same information, DBs like PostgreSQL give you extremely fine-grain control over what operations may be performed on which tables and indeed columns. Yes, you could have a desktop client tunnel in through SSH and then set file permissions, but I&#039;m not persuaded that it&#039;s quite the same thing.

Anyhow, as an experiment, I&#039;ve been fiddling with creating a simple forum structure with flat files and directories. All the current file-based forums I&#039;ve seen are called &#039;slow&#039; in reviews, so I&#039;ve been trying to figure out how to make on that&#039;s blistering fast.

We&#039;ll see.</description>
		<content:encoded><![CDATA[<p>In &#8216;Beating the Averages&#8217; and a few of his other essays, he&#8217;d talked about how the Viaweb Store used flat-files and directories instead of databases. He even said that when a competitor advertised for hackers with Oracle experience, they immediately considered them no threat at all.</p>
<p>I asked him if he felt that relational databases had *no* application at all, or if it&#8217;s just that they&#8217;re over-deployed in situations where 99% of the queries look like &#8220;*SELECT * FROM table WHERE id = &#8216;x&#8217;*&#8221; (sites like this one, for example)</p>
<p>He said that he felt they are not necessary at all, since they were originally devised to a) not lose your files and b) allow quick retrieval. Since modern OSes and filesystems don&#8217;t lose your files anyways, particularly if you&#8217;re running something like <acronym title="Redundant Array of Independent Disks">RAID</acronym>-5 on your server, plus they do a lot of caching and so on for you, he said that that&#8217;s adequate on its own.</p>
<p>Of course, you need to know what you&#8217;re doing as far as creating indexes and so on, but that&#8217;s not rocket science&#8230; it&#8217;s just that a relational DB normally does a lot of the heavy lifting for you.</p>
<p>I&#8217;m not sure I completely agree, since there&#8217;s a lot of sophisticated queries easily possible with <acronym title="Structured Query Language">SQL</acronym> that could take a fair bit hassle with files and directories. Also, if you&#8217;ve got multiple applications all trying to access the same information, DBs like PostgreSQL give you extremely fine-grain control over what operations may be performed on which tables and indeed columns. Yes, you could have a desktop client tunnel in through <acronym title="Secure SHell (encrypted protocol replaces telnet and FTP)">SSH</acronym> and then set file permissions, but I&#8217;m not persuaded that it&#8217;s quite the same thing.</p>
<p>Anyhow, as an experiment, I&#8217;ve been fiddling with creating a simple forum structure with flat files and directories. All the current file-based forums I&#8217;ve seen are called &#8217;slow&#8217; in reviews, so I&#8217;ve been trying to figure out how to make on that&#8217;s blistering fast.</p>
<p>We&#8217;ll see.</p>
]]></content:encoded>
	</item>
	<item>
		<title>By: Jeffrey Aho</title>
		<link>http://uwmike.com/articles/2005/04/21/accessible-celebrities/comment-page-1/#comment-3438</link>
		<dc:creator>Jeffrey Aho</dc:creator>
		<pubDate>Sat, 23 Apr 2005 18:27:25 +0000</pubDate>
		<guid isPermaLink="false">http://dev.uwmike.com/articles/2005/04/21/accessible-celebrities#comment-3438</guid>
		<description>As a matter of curiosity, what was the question you sent Paul Graham and what was the statement and response?</description>
		<content:encoded><![CDATA[<p>As a matter of curiosity, what was the question you sent Paul Graham and what was the statement and response?</p>
]]></content:encoded>
	</item>
</channel>
</rss>
<!-- uwMike.com will return...-->

