Irregular Webcomic!
http://www.irregularwebcomic.net/
Now neither irregular nor a webcomic! Geeky essays on Sunday and rerun comics other days.en-auTue, 28 Mar 2017 10:11:00 +0000Tue, 28 Mar 2017 10:11:00 +0000Custom codeIrregular Webcomic!
http://www.irregularwebcomic.net/
http://www.irregularwebcomic.net/logos/IWC_88x31.png8831Irregular Webcomic! #3632<img src="http://www.irregularwebcomic.net/comics/irreg3632.jpg" width="815" height="275" alt="Comic #3632" /><p>The other reason Minnesota Jones had to do this with a yardstick was that <a href="https://en.wikipedia.org/wiki/Smoot">Oliver R. Smoot</a> hasn't been born yet.Tue, 28 Mar 2017 10:11:00 +0000
http://www.irregularwebcomic.net/3632.html
http://www.irregularwebcomic.net/3632.htmlIrregular Webcomic! #3631<img src="http://www.irregularwebcomic.net/comics/irreg3631.jpg" width="815" height="275" alt="Comic #3631" /><p>I thought and thought and thought about how to show the unshrinking...
and eventually decided to brush it under the carpet.
Apparently the unshrinking device looks like a giant magnifying glass.
<hr>
<b>Quantum computers</b> are a cool new thing, and most people don't really know that much about them. It's difficult to know where to start explaining, for two reasons:
<ol>
<li>People's knowledge about how normal everyday digital computers work varies a lot, so using them as an analogy baseline might not always work so well, and</li>
<li>Before starting to write this I knew virtually nothing about quantum computers either.</li>
</ol>
So, right up front, I'll say that I might get some of this slightly wrong, but hopefully most of it will be accurate.
Another thing to get out of the way is that <a href="http://www.smbc-comics.com/comic/the-talk-3">this topic has also been tackled</a> by the webcomic <i>Saturday Morning Breakfast Cereal</i>. No doubt many of you will have seen this already, though I personally found it to be still rather cryptic and unenlightening, except to tell me what quantum computers <i>don't</i> do - namely, "trying all the answers in parallel". Okay, so if that's not what quantum computers are about, what <i>are</i> they doing?
Firstly, let's review how digital computers work. These are the computers we use every day, and you're using one right now to read this (unless you're reading a print-out). One of the most fundamental things a computer needs to do is to <i>represent data</i>, or information, in such a way that it can be stored, manipulated, and read out in a way that is meaningful for human users. Computers do this by rendering data into patterns of <i><a href="https://en.wikipedia.org/wiki/Bit">bits</a></i>, which is short for <i>binary digits</i>.
A bit is essentially either the digit 0 or the digit 1. A bit may be represented in other ways as well, such as heads or tails; or positive or negative; or on or off. The important thing is that there are exactly two choices of symbol or state to represent the bit. This is why it's called <i>binary</i> digit - because it represents one of two binary states. The word comes from the Latin <i>binarius</i>, meaning "consisting of two", and the <i>bi-</i> prefix can be seen in other English words relating to the number two, such as bicycle (two wheels), binocular (having two eyes), or biped (walking on two feet).
A binary digit is different from the usual decimal digits we use for counting: 0, 1, 2, 3, 4, 5, 6, 7, 8, 9. "Decimal" means related to the number 10, and you can see there are ten different digits. But with these ten digits, we can represent numbers of any magnitude by using a <a href="https://en.wikipedia.org/wiki/Positional_notation">positional or place-value system</a>. When we write 752, we mean the number formed by adding 2 (the rightmost digit), plus 5 multiplied by 10, plus 7 multiplied by 10<sup>2</sup>. As we move positional places to the left, each digit represents the number equal to itself multiplied by a power of the <a href="https://en.wikipedia.org/wiki/Radix"><i>base</i> (or <i>radix</i>)</a> of the number system, in this case 10. The first digit on the right is the number of ones, the next digit left of it is the number of tens, the next is the number of hundreds, followed the number of thousands and so on. We are all familiar with reading decimal numbers in this way.
In a binary representation, you do the same thing, except that the base of the system is 2, because there are two different digits. So let's read the binary number 1001101. The first binary digit (or bit) on the right represents the number of ones (1), the next number to the left represents the number of twos (0), then the number of fours (1), the number of eights (1), the number of sixteens (0), the number of 32s (0), and finally in this number the number of 64s (1). Adding these up: 1 + 4 + 8 + 64 = 77. So the binary number 1001101 is equal to the decimal number 77.
Why am I going into such detail on this point? Because it's important to understand that bits can <i>represent</i> much more than just the 0s and 1s they are made of. Imagine if we want to represent the entire contents of a book. What if we wanted to represent this in decimal notation, using just the digits 0 to 9? Well, we could convert each letter of the alphabet to a number, from A = 1 to Z = 26. We could even add some extra numbers to represent punctuation symbols, and lower case letters if we want. Then we could write out the entire book in numbers, by using the code numbers for each letter or symbol. You'd end up with a book full of digits, which a human couldn't read, but importantly the information in the book is <i>still there</i>, because you can decode the numbers back into letters and punctuation, reproducing it.
From here it's just a tiny step to encoding your book in bits, rather than decimal digits. Just assign each letter and punctuation symbol a binary number code, such as A = 1000001 to Z = 1011010, and record those<sup><a href="#note1" id="ref1">[1]</a></sup>. You end up with a whole bunch of 0s and 1s and nothing else, but again, all the contents of the book are there, and just need to be decoded back into English for someone to read.
From here it's a relatively small step from encoding English text to storing other sorts of information as well. Images can be represented by numbers which give the brightness levels of various colours in each pixel, and audio can be represented by numbers giving the amplitudes of sound waves at different times, and movies can be represented by combinations of both of those things. It's not quite this straightforward, as we've come up with cleverer ways to represent these sorts of data that reduce the quite large amounts of storage needed. But in general, this is how computers encode the data that we use and manipulate every day - all as sequences of bits.
So much for representing data. A computer also has to <i>physically store</i> the bits it uses somehow. There are several ways it can do this:
<ul>
<li>In the electrical circuitry of a computer's processor, bits can be represented by different voltage levels. Typically, a level above a certain threshold might represent a 1, while any voltage below the threshold represents a 0.</li>
<li>In a magnetic storage medium such as a hard drive (or a <a href="https://en.wikipedia.org/wiki/Floppy_disk">floppy disk</a>, back when people still used them), a small area of the magnetic medium can be magnetised with its field aligned in a certain direction to represent a 0, and in the opposite direction to represent a 1.</li>
<li>In a solid-state storage device, such as a USB stick or solid-state drive (SSD), a small region of the device's circuitry chip can contain a pool of excess electrons to represent a 1, and when empty of electrons it represents a 0.</li>
</ul>
The next thing a computer needs to do is <i>manipulate</i> the bits it is storing, to carry out calculations on them. This can be done electronically using <a href="https://en.wikipedia.org/wiki/Logic_gate">logic gates</a>. Conceptually, a logic gate takes one or two bits as input, and produces a bit as output which depends on the input bits. There's a lot that can be said about how these work at the logical level, but for now let's just say that by combining logic gates you can get them to perform arithmetic on binary encoded numbers, and at higher conceptual levels they can perform operations such as word processing, image processing, and so on - all the things we use computers to do. Jumping down a conceptual level, logic gates are implemented as electronic circuits which change in output voltage according to the voltages of their inputs.
Okay, that's a very quick overview of digital computers. Now let's talk about quantum computers, and where they differ from digital computers. And the differences start right at the very fundamental point of how they represent data.
A quantum computer doesn't use bits. It doesn't represent data using individual pieces of information that hold one of two possible symbols or states. The fundamental unit of data in a quantum computer is not a bit - it is a very different thing called a <i>quantum bit</i>, or <i><a href="https://en.wikipedia.org/wiki/Qubit">qubit</a></i> for short. The "bit" part of the name is carried over merely by analogy to the "fundamental unit of data" of digital computing, and the "bi-" part no longer indicates an etymological meaning of "two". Qubits do not have just two possible states; they are not really binary at all.
So what <i>is</i> a qubit? Let's imagine we want to encode a regular digital bit, which has two possible states, using some other physical system that, at least naively, has two possible states. I'm going to choose the polarisation of light. I've written a <a href="/3335.html">whole other annotation about polarisation</a>, and it might be useful to review that material before continuing if you don't remember much about polarisation.
Now, I'm going to encode my bit of data in the polarisation of a light beam, such that to represent a 0 the beam will be polarised with left-handed circular (LHC) polarisation, and to represent a 1 the beam will be polarised with right-handed circular (RHC) polarisation. Sound reasonable? But wait, some of you are saying, what if the beam is <i>linearly</i> polarised? Or it's <i>elliptically</i> polarised? Or it's <i>not polarised at all</i>? What value do any of those represent? Good question!
Let's simplify things by talking about a single <a href="https://en.wikipedia.org/wiki/Photon">photon</a> of light, rather than a beam of many photons. A photon, as it happens, can have two different measured polarisation states: LHC, or RHC. By no coincidence, these neatly align with the values I chose to represent 0 or 1. Okay. In a classical physics world, we'd be done, and we could go and design a digital computer that uses photons of light to encode bits in their polarisation.
But photons are not described by classical physics. Photons require the use of <a href="https://en.wikipedia.org/wiki/Quantum_mechanics">quantum physics</a> to fully describe their behaviour. This means that rather than merely being in one of either the LHC or RHC states, a photon can be in a quantum mechanical superposition of these states. Without delving into the precise mathematical intricacies, we can imagine a photon as potentially being "half LHC and half RHC", or "1/4 LHC and 3/4 RHC", or any other combination that adds up to 1. However, when we <i>measure</i> the polarisation of a photon, we only ever see it as either LHC or RHC. We don't get a polarisation measurement of "half LHC and half RHC". What is going on?
The fractions in "half LHC and half RHC" are manifested as <i>probabilities</i>. Imagine we have a continuous stream of individual photons, all in the polarisation state "half LHC and half RHC". If we measure the polarisations of these photons, each individual photon will be seen to be either LHC or RHC... but we will see that approximately half the photons we measure read as LHC and the other half will read as RHC. Similarly, if we measure a stream of photons that are all in the same "1/4 LHC and 3/4 RHC" polarisation state, we will measure roughly 1/4 of them as LHC and 3/4 of them as RHC. In other words, if we measure the polarisation of a "half LHC and half RHC" photon, there is a 50% chance we'll measure it as LHC, and a 50% chance we'll measure it as RHC. And if we measure the polarisation of a "1/4 LHC and 3/4 RHC" photon, there is a 25% chance we'll measure it as LHC, and a 75% chance we'll measure it as RHC.
This may sound weird and counter-intuitive, because everyday objects we interact with don't show this sort of superposition of states and probabilistic behaviour. This is in fact the motivation behind the <a href="https://en.wikipedia.org/wiki/Schrödinger%27s_cat">Schrödinger's cat</a> thought experiment, which is an attempt to show how absurd quantum behaviour is if you think about it on a macro level. But at the level of fundamental particles, it's what happens, and we have plenty of experiments that demonstrate this superposed-states/probabilistic behaviour.
Going one level deeper into the mathematics, the probabilities we can measure - namely the halves in "half LHC and half RHC", or the 1/4 and 3/4 in "1/4 LHC and 3/4 RHC" - are actually the magnitudes of the squares of numbers called <a href="https://en.wikipedia.org/wiki/Probability_amplitude">probability amplitudes</a> in quantum mechanics. In other words, mathematically, the superposed state we've called "half LHC and half RHC" can more properly be written using the square root of a half, as:
<div style="text-align:center"><img src="/annotations/annot3631_1.png" width="250" height="71" alt="1/sqrt(2)|LHC>+1/sqrt(2)|RHC>"> (equation 1)</div>
and the superposed state we've called "1/4 LHC and 3/4 RHC" can more properly be written as:
<div style="text-align:center;"><img src="/annotations/annot3631_2.png" width="250" height="78" alt="1/2|LHC>+sqrt(3)/2|RHC>"> (equation 2)</div>
In this form, the mathematical operations that govern the physics of the photons are more meaningful and useful. The thing to remember is that to produce the probabilities of measuring LHC or RHC, you need to <i>square</i> the given probability amplitudes and then take the magnitude of the resulting number.
But wait a second, you might say, if we're now squaring the probability amplitude <span style="white-space: nowrap;">1/√<span style="text-decoration:overline;"> 2 </span></span> to produce the observed probability 1/2, what's to say the amplitude isn't negative: <span style="white-space: nowrap;">-1/√<span style="text-decoration:overline;"> 2 </span></span>, because when you square that it also gives 1/2? In fact, nothing stops this; the probability amplitudes can indeed be negative. In fact, because we take the magnitude of the square, the probability will work out even if the amplitude is an imaginary or complex number!<sup><a href="#note2" id="ref2">[2]</a></sup>
At this point, it might be useful to visualise the difference between a bit and a qubit using a diagram. This can be done by looking at the diagram known as the <a href="https://en.wikipedia.org/wiki/Bloch_sphere">Bloch sphere</a>, reproduced here as Figure 1.
<table style="margin-left:auto; margin-right:auto;" border="1" cellpadding="5" align="center" hspace="10" width="310">
<tr><td>
<img src="/annotations/annot3631_3.png" width="300" height="341" alt="Bloch sphere" caption="Bloch sphere" hspace="10"><br><br>
<b>Figure 1:</b> Bloch sphere. (<a href="https://commons.wikimedia.org/wiki/File:Bloch_Sphere.svg">Creative Commons BY-SA by Glosser.ca from Wikimedia Commons</a>.)
</td></tr></table>
The Bloch sphere shows graphically the difference between possible values of a bit and a qubit. The two possible values of a bit are written as |0⟩ and |1⟩, which can just be thought of as fancy quantum mechanics notation for 0 and 1. The value |0⟩ lies at the top of the sphere - the north pole if you will. The value |1⟩ lies directly opposite it at the bottom of the sphere - the south pole. For a bit, these are the only two possible data values it can contain.
For a qubit, the situation is very different. As we've seen above, a qubit can hold any value given by a sum of probability amplitudes of |0⟩ and |1⟩. If the squares of the amplitudes add to give 1, as constrained by the probabilities of normal measurement, this restricts the location of the qubit values to the surface of the Bloch sphere. So:
<ul>
<li>A bit can only hold values at the north and south pole of the Bloch sphere.</li>
<li>A qubit can hold <i>any value on the surface of the entire sphere</i>.</li>
</ul>
As you can see, this is in some ways a much, much richer amount of information than a single bit can hold. However, what happens when you read out a qubit? Let's assume you have a qubit in a quantum computer memory, and you read out its value. Let's take the qubit specified by equation 1 above. It's an equal superposition of probability amplitudes of |0⟩ and |1⟩. On the Bloch sphere, it would be a point somewhere on the equator of the sphere, exactly midway between the |0⟩ and |1⟩ poles. But if we read out the value, we are measuring the state of a quantum system - in fact we might physically be measuring the polarisation of a photon, if that's how our quantum computer stores its qubits. We saw above that measuring a photon in the state given by equation 1 results in a measurement of LHC (or 0) with probability 50%, and a measurement of RHC (or 1) with probability 50%.
In other words, we can't <i>measure</i> the superposed state of the qubit directly - we still only read out a 0 or a 1. But the superposed state is <i>there</i> before we measure it, and it can affect other things, and be affected by other things - and <i>this</i> is the basis of quantum computing.
In a nutshell: Quantum computing is operating on and manipulating the potentially superposed states of qubits, and then reading out the probabilistic values of the qubits at the end of the calculation.
How do you operate on qubits? You can't just use digital logic gates, as you would with regular bits. Instead you use physical interactions that affect the states of the physical components you are using to represent the qubits. Photon polarisations are only one possible way to do this, but you can also represent qubits with things such as the quantum spin of electrons, the nuclear spin of atomic nuclei, or other more esoteric quantum systems.<sup><a href="#note3" id="ref3">[3]</a></sup> Whatever you use, such quantum systems exhibit a property known as <i>quantum interference</i> - which I don't want to go into in any detail here, but which essentially means that their probability amplitudes interact in ways which can affect the outcomes of measuring the system.<sup><a href="#note4" id="ref4">[4]</a></sup>
So for example, if you get a bunch of electrons with their spins encoding data as qubits, then operating on some of the electrons by magnetic fields that could change their spin states can also affect the states of some of the other electrons not being directly manipulated. This is where my understanding, even after some extensive reading, peters out, so I'll stop there and take a few steps back to try to bring all this together.
The important thing about operating on qubits is that you can affect the probability amplitudes of all of your qubits, essentially in unison. If you're clever, you can hopefully design an algorithm that works to push the probability amplitudes of all of the qubits towards the correct answer that you're trying to compute, while reducing the probability amplitude of incorrect solutions. Such algorithms exist for some relatively simple problems - possibly the most interesting one is <a href="https://en.wikipedia.org/wiki/Shor%27s_algorithm">Shor's algorithm</a>, which is a quantum algorithm designed for factorising integers. (I won't even pretend to understand how Shor's algorithm works - despite reading a description, I don't really grok it.)
Some quantum computing algorithms are deterministic, meaning they always give the correct answer. But others, which will usually include more interesting and complex algorithms - may only be <i>probabilistic</i>. This can come about by the nature of qubits. Imagine that in the correct answer, a particular qubit should read out as 1. But the best algorithm available may only be capable of setting that qubit to the superposed state given by equation 2. This means that 75% of the time, the bit will give the correct value when read out, but 25% of the time it will produce the incorrect value of 0.
This means quantum computers <i>don't necessarily always give the correct answer</i>! What use is that?
Well, imagine you have an algorithm that manipulates a bunch of qubits, and when the qubits are read out, they give you the correct answer just 50% of the time. For another 25% of the time, it gives incorrect answer A, and the other 25% of the time it gives you incorrect answer B. Well, you can run the algorithm on your quantum computer say 100 times. On 48 of those runs it produces the number 1234, on 25 of the runs it gives the number 1238, and on the other 27 runs it gives the answer 1247. What do you think is the correct answer? Right, it's almost certainly 1234. You may not be 100% convinced, but for many purposes you only need to be 99.999% convinced anyway. If you need more convincing, you can run it more times to build up the statistics.
Okay, so we've very roughly covered what a quantum computer is and how it works. Why would you want to use a quantum computer? The answer is that some quantum algorithms can take advantage of the interference relations between qubits to significantly speed up some calculations that take a prohibitively long time in digital computers.
Factorising integers (what Shor's algorithm does), is a notoriously time-intensive problem for standard digital computers. Once the intergers get large enough - and we're talking a few hundred digits here, not billions or anything - even the fastest digital supercomputers we have would take thousands of years of calculating to factorise them. This fact is used in many cryptography applications, where you can generate a public cryptographic key by multiplying two large numbers (which is easy), but to break the encryption without the original numbers you need to factorise the key, which is effectively impossible with current technology.
However, a quantum algorithm, successfully implemented, could factorise large numbers like those used in modern cryptography in a very short amount of time. This is the power of quantum computing. A working quantum computer with enough qubits could break every encryption currently in use. Besides rendering encryption obsolete, there are many other practical uses for quantum algorithms, particularly in fields with many combinatorial interactions, such as DNA analysis, or indeed in studying quantum physics itself.
So what's the state of the art? Well, quantum computers are <i>hard to build</i>. Even building a single qubit is difficult. It needs to be an isolated quantum mechanical system, which basically means it has to be something like an elementary particle or an atomic nucleus, or something of that scale. To build such a thing, you often need large machines that generate precise magnetic fields, or to do the whole thing at cryogenic temperatures, or to isolate the thing in a vibration-proof room, or all of the above. You can't just stick a qubit into a silicon chip that plugs into your computer - not yet anyway.
So far, researchers have built arrays of just a few qubits, 4 or 8 of them. Shor's algorithm has been demonstrated to work on such a qubit array, but so far the highest number successfully factorised is just 21. The Canadian company <a href="https://en.wikipedia.org/wiki/D-Wave_Systems">D-Wave Systems</a> manufactures and sells what it claims are quantum computers with around 1000 qubits, though there has been much scepticism in quantum computing circles that their machines are actually quantum computers as claimed. There seems to be <i>some</i> sort of quantum behaviour, and they show <i>some</i> speed-up over normal computers for some algorithms, but it's not yet entirely clear what is going on with them.
And that's where we are with quantum computing. It's at about the stage of technology that aeroplanes were in the years immediately after the Wright Brothers flew at Kitty Hawk. The very first crude attempts have been shown to work, and now we're trying to improve the technology in terms of capability and reliability. How fast it takes off from here... is anyone's guess at the moment.
<hr>
Note: This annotation was inspired by reader Sean M., who requested an essay on the topic of quantum computing as part of his Patreon supporter reward. If you'd like me to write an extended annotation on any topic you care to name, or if you just want to show some support for the comics and other creative work I share, please consider <a href="https://www.patreon.com/dmmaus">becoming a patron</a>.
<hr>
<a href="#ref1" id="note1">[1]</a> In fact, these are the binary codes assigned to A and Z in the <a href="https://en.wikipedia.org/wiki/ASCII">ASCII character encoding standard</a>. ASCII also specifies binary codes for lower case letters, decimal digits, and several punctuation characters. The newer <a href="https://en.wikipedia.org/wiki/Unicode">Unicode</a> standard specifies binary encoding for thousands of other symbols, including characters for non-English languages, musical and mathematical notation, and emojis.
<a href="#ref2" id="note2">[2]</a> I talk about imaginary and complex numbers a bit <a href="/1960.html">here</a>, if you need a refresher.
<a href="#ref3" id="note3">[3]</a> You can see a larger list of possible physical qubit systems listed <a href="https://en.wikipedia.org/wiki/Qubit#Physical_representation">near the bottom of Wikipedia's qubit article</a>.
<a href="#ref4" id="note4">[4]</a> In terms of the Bloch sphere, I think this means that some of the qubits can even end up in states represented by points <i>inside</i> the sphere, rather than just on the surface - though honestly I'm a little fuzzy myself on this point, so don't take that as gospel.Mon, 27 Mar 2017 10:11:00 +0000
http://www.irregularwebcomic.net/3631.html
http://www.irregularwebcomic.net/3631.htmlIrregular Webcomic! #1544 Rerun<img src="http://www.irregularwebcomic.net/comics/irreg1544.jpg" width="815" height="275" alt="Comic #1544" /><p>Of course, the Allosaurus has been President since he <a href="/polls/poll0021.html">won the election in 2004</a>.
<p>
The flower garden outside the window is a photo I took of the gardens of the California State Capitol in Sacramento in 2004.
Not that you can see much detail in it.
<hr>
<b>2017-03-26 Rerun commentary:</b> And the Allosaurus won his <a href="/polls/poll0409.html">fourth</a> <a href="/polls/poll0611.html">consecutive</a> term <a href="/polls/poll0666.html">in 2016</a>.
Yeah, I don't think I'll say anything more about that at this point in time.
It's interesting that the Allosaurus installed a <i><a href="https://en.wikipedia.org/wiki/Dutch_door">stable door</a></i> between his secretary's office and the Oval Office.Sun, 26 Mar 2017 10:11:00 +0000
http://www.irregularwebcomic.net/1544.html
http://www.irregularwebcomic.net/1544.htmlIrregular Webcomic! #1543 Rerun<img src="http://www.irregularwebcomic.net/comics/irreg1543.jpg" width="815" height="275" alt="Comic #1543" /><p>This is the one bit of the film <i>Dr No</i> that never really made any sense to me.
<p>
James Bond, placed in a cell by Dr No, removes a grate and crawls into what is apparently an air conditioning duct. Fine.
<p>
But as he crawls through it, the duct becomes scorchingly hot, forcing him to tear his shirt into strips of cloth to insulate his
hands. Then a wave of water washes through the duct, drenching him. And then he negotiates a long vertical shaft, leading down, and finally
emerges in another room.
<p>
Now I'm no expert at air conditioning, but I'm of the general opinion that having your ducts get nearly red hot is not ideal for cooling a
building down. And what's with the water? Where does it come from? Where does it go? And why doesn't it fall down the vertical shaft and
out the vent and into the room at the end? Again, not an expert, but I imagine pumping water through your air conditioning ducts is not
exactly standard practice.
<p>
If you read Ian Fleming's original novel, however, it has a similar sequence. Except that it's not Bond crawling through an air conditioning duct
in an attempt to escape from his cell. It's Bond crawling through a deliberate obstacle course designed by Dr No, with viewports through which
he can observe the suffering Bond has to go through. All of a sudden the heat and the water and the vertical shafts make sense. Well, as much sense
as an evil villain's torture-course-cum-deathtrap can.
<hr>
<b>2017-03-25 Rerun commentary:</b> I remember one time I was watching this film, when very young - it may have been the first time I ever saw it. I asked my mother what the deal was with the water rushing through the ducts, and she said it had something to do with the air conditioning. At the time I accepted that answer, but in the fullness of time I came to realise how odd that was. I'm not sure if my mother was similarly confused and just made up something to satisfy my childish curiosity, or if she genuinely thought that air conditioning might actually work that way. I suppose I could ask her one day, though I don't know if she'd remember the specific occasion.Sat, 25 Mar 2017 10:11:00 +0000
http://www.irregularwebcomic.net/1543.html
http://www.irregularwebcomic.net/1543.htmlIrregular Webcomic! #1542 Rerun<img src="http://www.irregularwebcomic.net/comics/irreg1542.jpg" width="815" height="275" alt="Comic #1542" /><p>Good sales tactic.
<hr>
<b>2017-03-22 Rerun commentary:</b> I'm quite pleased with the night time street lighting outside the windows. I believe that I had to darken the windows in post-production, and add the yellow cast to the window across the street. I can just see some artefacts around some of the edges.
If I did this now, I'd probably also add a pool of light directly under the street lamp.Fri, 24 Mar 2017 10:11:00 +0000
http://www.irregularwebcomic.net/1542.html
http://www.irregularwebcomic.net/1542.htmlIrregular Webcomic! #3630<img src="http://www.irregularwebcomic.net/comics/irreg3630.jpg" width="815" height="275" alt="Comic #3630" /><p>I'd be curious to hear how Serron could rephrase that in a way that didn't result in him essentially demanding to be thrown out the airlock. Though given it's Serron, I'm sure he could manage to turn it into throwing everyone <i>else</i> out the airlock somehow.Thu, 23 Mar 2017 10:11:00 +0000
http://www.irregularwebcomic.net/3630.html
http://www.irregularwebcomic.net/3630.htmlIrregular Webcomic! #1541 Rerun<img src="http://www.irregularwebcomic.net/comics/irreg1541.jpg" width="815" height="275" alt="Comic #1541" /><p>This is probably funnier if you've seen <i>Indiana Jones and The Last Crusade</i>, as it references Indiana Jones' line
that "Seventy percent of all archaeology is done in the library."
<p>
If you haven't seen it, I won't spoil why that line is amusing in the film.
<hr>
<b>2017-03-19 Rerun commentary:</b> Honestly, I wonder about the accuracy of that 70% figure. I suspect that in reality it may actually be higher, and that <i>substantially</i> less than 30% of the work is done in the field.
I haven't formally studied archaeology, but I know from experience with astronomy that you're lucky if more than about 1% of the work in astronomy is done in an observatory.
Oh, and yes, that barrier that just happens to be sitting there in St Mark's Square does seem to be conveniently handy for them to hide behind, doesn't it?Wed, 22 Mar 2017 10:11:00 +0000
http://www.irregularwebcomic.net/1541.html
http://www.irregularwebcomic.net/1541.html