<?xml version="1.0"?>
<rss version="2.0.">
  <channel>
    <title>Advogato blog for tmorgan</title>
    <link>http://www.advogato.org/person/tmorgan/</link>
    <description>Advogato blog for tmorgan</description>
    <language>en-us</language>
    <generator>mod_virgule</generator>
    <pubDate>Fri, 25 Jul 2008 21:50:03 GMT</pubDate>
    <item>
      <pubDate>Mon, 5 Jan 2004 18:20:09 GMT</pubDate>
      <title>5 Jan 2004</title>
      <link>http://www.advogato.org/person/tmorgan/diary.html?start=3</link>
      <guid>http://www.advogato.org/person/tmorgan/diary.html?start=3</guid>
      <description>&lt;a href="http://advogato.org/person/forrest/" &gt;forrest&lt;/a&gt; is exactly right.  I think the one addition you have to make is make key-signing be more than a binary concept, as in "I trust this person".  It needs to be more of a way of verifying attributes.  E.g. there should be a way of saying:&lt;ul&gt;&lt;li&gt;
"I trust that this email address belongs to this real-life person"
&lt;li&gt;
"I trust the owner of this email address not to send me spam"
&lt;li&gt;other attributes as need arises?
&lt;/ul&gt;
I'm not very familiar with GnuPG - is this a feasible idea?</description>
    </item>
    <item>
      <pubDate>Fri, 28 Nov 2003 15:08:52 GMT</pubDate>
      <title>28 Nov 2003</title>
      <link>http://www.advogato.org/person/tmorgan/diary.html?start=2</link>
      <guid>http://www.advogato.org/person/tmorgan/diary.html?start=2</guid>
      <description>&lt;a href="" 'http://advogato.org/person/mikehearn/'&gt;mikehearn&lt;/a&gt;, the &lt;a href="" 'http://boost.org/libs/python/doc/index.html'&gt; boost python&lt;/a&gt; library has pretty seamless C++ to Python exception translation that by default handles some of the standard C++ exception types, and allows you to provide custom exception translation functions.&lt;br&gt;
An example is given &lt;a href="" 'http://boost.org/libs/python/doc/tutorial/doc/exception_translation.html'&gt;
here&lt;/a&gt;.&lt;br&gt;&lt;br&gt;Do you think that it's better for exceptions to be defined by the OS or the runtime environment, given that  not all languages support exceptions, and the definition of an exception and exception handling semantics vary between languages?  </description>
    </item>
    <item>
      <pubDate>Fri, 21 Nov 2003 14:47:38 GMT</pubDate>
      <title>21 Nov 2003</title>
      <link>http://www.advogato.org/person/tmorgan/diary.html?start=1</link>
      <guid>http://www.advogato.org/person/tmorgan/diary.html?start=1</guid>
      <description>&lt;blockquote&gt;
&lt;b&gt; Functors in Haskell and C++ &lt;/b&gt;

&lt;p&gt; C++ is my language of choice for almost every project I do, not least for it's wide range of support language paradigms - OO, functional, generic etc.  But I'd be the last person to say that it's a &lt;em&gt;pretty&lt;/em&gt; language.  Example: creating an object that will add 2 to a number:-

&lt;p&gt; C++ 

&lt;p&gt; &lt;tt&gt;
std::binder1st&amp;lt;std::plus&amp;lt;int&amp;gt; &amp;gt; binder=
  std::bind1st(std::plus&amp;lt;int&amp;gt;(),2);
&lt;/tt&gt;

&lt;p&gt; Now compare this with &lt;a href="" 'http://www.haskell.org/'&gt;Haskell&lt;/a&gt;

&lt;p&gt; &lt;tt&gt;
 (+2)
&lt;/tt&gt;

&lt;p&gt; I haven't got very far with Haskell yet, I'm just trying to learn it to stretch my brain a bit, but this does highlight the hoops one has to jump to to make C++ behave like a functional language.  That said, the &lt;a href="" 'http://boost.org'&gt;boost&lt;/a&gt; people are making this easier with the boost::bind library, which I use a fair bit,  and maybe the lambda library, though I haven't tried that yet.

&lt;p&gt; I guess it's a testimony to the power of C++ that it can handle so many diverse paradigms, even if it isn't always pretty.  And it can be taken far beyond what the original creators intended - witness what's been done with template meta-programming. 

&lt;p&gt; It also seem that the C++ compile-time language (i.e. template meta-programming) is a purely functional language, (see &lt;a href="" 'http://boost.org/libs/mpl/doc/paper/mpl_paper.html'&gt; here&lt;/a&gt;, so we have this weird hybrid of a language that is OO and procedural at runtime, but functional at compile time.  No wonder it takes so long to become truly skilled at using it.
&lt;/blockquote&gt;

</description>
    </item>
    <item>
      <pubDate>Thu, 20 Nov 2003 20:15:59 GMT</pubDate>
      <title>20 Nov 2003</title>
      <link>http://www.advogato.org/person/tmorgan/diary.html?start=0</link>
      <guid>http://www.advogato.org/person/tmorgan/diary.html?start=0</guid>
      <description>&lt;a href="" 'http://advogato.com/person/slamb/'&gt;slamb, &lt;/a&gt;, I'm also very glad I stumbled across &lt;a href="" 'http://www.scons.org/'&gt;scons&lt;/a&gt;.  It's made maintaing cross platform library software vastly simpler than the ugly makefiles I was writing.  It definitely should be publicised more, it's a very nice piece of work.</description>
    </item>
  </channel>
</rss>
