sync changes:

- M  site/index.html

- M  site/systems-research/hints-for-computer-system-design.html

- M  systems-research/hints-for-computer-system-design.md

site/index.html

@@ -202,13 +202,13 @@ <h1>My Notes</h1>
                     <h2>Recent Notes</h2>
 <ul class='recent-posts'>
                 <li>
-                    <a href="/natural-language-processing/reading/information-retrieval.html">Information Retrieval</a>
-                    <span class="date">2025-01-07</span>
+                    <a href="/systems-research/hints-for-computer-system-design.html">Hints For Computer System Design</a>
+                    <span class="date">2025-01-08</span>
 
                 </li>
                 <li>
-                    <a href="/systems-research/hints-for-computer-system-design.html">Hints For Computer System Design</a>
-                    <span class="date">2025-01-06</span>
+                    <a href="/natural-language-processing/reading/information-retrieval.html">Information Retrieval</a>
+                    <span class="date">2025-01-07</span>
 
                 </li>
                 <li>

site/systems-research/hints-for-computer-system-design.html

@@ -179,7 +179,7 @@
         </div>
         <h1>Hints For Computer System Design</h1>
         <div class="meta">
-            <span>Last modified: 2025-01-06</span>
+            <span>Last modified: 2025-01-08</span>
 
         </div>
         <div class="content">
@@ -197,8 +197,32 @@ <h4 id="caching">Caching</h4>
 <p>Caches should ideally have adaptive sizes.</p>
 <p>A classic example is the caching in hardware that uses $[Fetch, \text{address}, \text{content of address}]$ tuples. Similarly, virtual memory uses $[Page, \text{address}, \text{content of address}]$ tuples.</p>
 <p>However, more complicated applications of caching exist. In real-time systems, you're often trying to cache the state of a system given small changes corresponding to events. The key here is to try and invalidate as few entries as possible in response to events.</p>
+<h3 id="lecture-review-notes">Lecture Review Notes</h3>
+<h4 id="why-is-system-design-hard">Why is system design hard?</h4>
+<ul>
+<li>external interface isn't well defined <ul>
+<li>requirements aren't clear</li>
+<li>Things are often not well-designed</li>
+</ul>
+</li>
+<li>The measure of success is not very clear<ul>
+<li>Many different ways to interact with a system</li>
+<li>Many systems, even in production, have bugs</li>
+</ul>
+</li>
+</ul>
+<h4 id="throw-one-away">Throw one away</h4>
+<ul>
+<li>Always be prepared to discard your prototype</li>
+<li>Throw ideas at the wall and go with what sticks</li>
+</ul>
 <h3 id="open-questions">Open Questions</h3>
 <h2 id="-">-</h2>
+<h3 id="further-reading">Further Reading</h3>
+<ul>
+<li>MicroLog</li>
+<li>https://github.com/DPDK/dpdk</li>
+</ul>
         </div>
 
     </main>

systems-research/hints-for-computer-system-design.md

@@ -24,7 +24,30 @@ A classic example is the caching in hardware that uses $\[Fetch, \text{address},
 
 However, more complicated applications of caching exist. In real-time systems, you're often trying to cache the state of a system given small changes corresponding to events. The key here is to try and invalidate as few entries as possible in response to events.
 
+
+### Lecture Review Notes
+
+#### Why is system design hard?
+
+- external interface isn't well defined 
+    - requirements aren't clear
+    - Things are often not well-designed
+- The measure of success is not very clear
+    - Many different ways to interact with a system
+    - Many systems, even in production, have bugs
+
+#### Throw one away
+
+- Always be prepared to discard your prototype
+- Throw ideas at the wall and go with what sticks
+
+
 ### Open Questions
 
 -
 -
+
+### Further Reading
+
+- MicroLog
+- https://github.com/DPDK/dpdk