The compiler as we know it is generally attributed to Grace Hopper, who also popularized the notion of machine-independent programming languages and served as technical consultant in 1959 in the project that would become the COBOL programming language. The second part is not important for today's post, but not enough people know how awesome Grace Hopper was and that's unfair.

It's been at least 60 years since we moved from assembly-only code into what we now call "good software engineering practices". Sure, punching assembly code into perforated cards was a lot of fun, and you could always add comments with a pen, right there on the cardboard like well-educated cavemen and cavewomen (cavepeople?). Or, and hear me out, we could use a well-designed programming language instead with fancy features like comments, functions, modules, and even a type system if you're feeling fancy.

None of these things will make our code run faster. But I'm going to let you into a tiny secret: the time programmers spend actually coding pales in comparison to the time programmers spend thinking about what their code should do. And that time is dwarfed by the time programmers spend cursing other people who couldn't add a comment to save their life, using variables named var and cramming lines of code as tightly as possible because they think it's good for the environment.

The type of code that keeps other people from strangling you is what we call "good code". And we can't talk about "good code" without it's antithesis: "write-only" code. The term is used to describe languages whose syntax is, according to Wikipedia, "sufficiently dense and bizarre that any routine of significant size is too difficult to understand by other programmers and cannot be safely edited". Perl was heralded for a long time as the most popular "write-only" language, and it's hard to argue against it:

open my $fh, '<', $filename or die "error opening $filename: $!";
my $data = do { local $/; <$fh> };

This is not by far the worse when it comes to Perl, but it highlights the type of code you get when readability is put aside in favor of shorter, tighter code.

Some languages are more propense to this problem than others. The International Obfuscated C Code Contest is a prime example of the type of code that can be written when you really, really want to write something badly. And yet, I am willing to give C a pass (and even to Perl, sometimes) for a couple reasons:

• C was always supposed to be a thin layer on top of assembly, and was designed to run in computers with limited capabilities. It is a language for people who really, really need to save a couple CPU cycles, readability be damned.
• We do have good practices for writing C code. It is possible to write okay code in C, and it will run reasonably fast.
• All modern C compilers have to remain backwards compatible. While some edge cases tend to go away with newer releases, C wouldn't be C without its wildest, foot-meet-gun features, and old code still needs to work.

Modern programming languages, on the other hand, don't get such an easy pass: if they are allowed to have as many abstraction layers and RAM as they want, have no backwards compatibility to worry about, and are free to follow 60+ years of research in good practices, then it's unforgivable to introduce the type of features that lead to write-only code.

Which takes us to our first stop: Rust. Take a look at the following code:

let f = File::open("hello.txt");
let mut f = match f {
    Ok(file) => file,
    Err(e) => return Err(e),
};

This code is relatively simple to understand: the variable f contains a file descriptor to the hello.txt file. The operation can either succeed or fail. If it succeeded, you can read the file's contents by extracting the file descriptor from Ok(file), and if it failed you can either do something with the error e or further propagate Err(e). If you have seen functional programming before, this concept may sound familiar to you. But more important: this code makes sense even if you have never programmed with Rust before.

But once we introduce the ? operator, all that clarity is thrown off the window:

let mut f = File::open("hello.txt")?;

All the explicit error handling that we saw before is now hidden from you. In order to save 3 lines of code, we have now put our error handling logic behind an easy-to-overlook, hard-to-google ? symbol. It's literally there to make the code easier to write, even if it makes it harder to read.

And let's not forget that the operator also facilitates the "hot potato" style of catching exceptions¹, in which you simply... don't:

File::open("hello.txt")?.read_to_string(&mut s)?;

Python is perhaps the poster child of "readability over conciseness". The Zen of Python explicitly states, among others, that "readability counts" and that "sparser is better than dense". The Zen of Python is not only a great programming language design document, it is a great design document, period.

Which is why I'm still completely puzzled that both f-strings and the infamous walrus operator have made it into Python 3.6 and 3.8 respectively.

I can probably be convinced of adopting f-strings. At its core, they are designed to bring variables closer to where they are used, which makes sense:

"Hello, {}. You are {}.".format(name, age)
f"Hello, {name}. You are {age}."

This seems to me like a perfectly sane idea, although not one without drawbacks. For instance, the fact that the f is both important and easy to overlook. Or that there's no way to know what the = here does:

some_string = "Test"
print(f"{some_string=}")

(for the record: it will print some_string='Test'). I also hate that you can now mix variables, functions, and formatting in a way that's almost designed to introduce subtle bugs:

print(f"Diameter {2 * r:.2f}")

But this all pales in comparison to the walrus operator, an operator designed to save one line of code²:

# Before
myvar = some_value
if my_var > 3:
    print("my_var is larger than 3")

# After
if (myvar := some_value) > 3:
    print("my_var is larger than 3)

And what an expensive line of code it was! In order to save one or two variables, you need a new operator that behaves unexpectedly if you forget parenthesis, has enough edge cases that even the official documentation brings them up, and led to an infamous dispute that ended up with Python's creator taking a "permanent vacation" from his role. As a bonus, it also opens the door to questions like this one, which is answered with (paraphrasing) "those two cases behave differently, but in ways you wouldn't notice".

I think software development is hard enough as it is. I cannot convince the Rust community that explicit error handling is a good thing, but I hope I can at least persuade you to really, really use these type of constructions only when they are the only alternative that makes sense.

Source code is not for machines - they are machines, and therefore they couldn't care less whether we use tabs, spaces, one operator, or ten. So let your code breath. Make the purpose of your code obvious. Life is too short to figure out whatever it is that the K programming language is trying to do.

Footnotes

• 1: Or rather "exceptions", as mentioned in the RFC
• 2: If you're not familiar with the walrus operator, this link gives a comprehensive list of reasons both for and against.

I have been not-so-recently tasked with the job of putting together my family's family tree. It has occurred to me that other people might want to give the task a try, so here is my experience as a programmer.

First steps

The first important step was collecting my family information. I did this the old-fashioned way, sitting down with my mom and writing down in a spreadsheet names and relations. We didn't go beyond "married to" and "son/daughter of" because it was not important for us, but you could extend this information with birthdays, marriage dates, and so on.

The spreadsheet was good for collecting data fast, but getting the graph done required better tools: like all good programmers I'm lazy, and I'm not going to sit down and hand-place over 100 nodes if I can avoid it. I therefore turned to Gramps, an open-source software for Genealogical research. This software offers a lot of functionality and may seem daunting at times, but fear not: the software is clearly done by people who understand the problem very well, and it lets you record as much or as little information as you want. My favorite example: you can add a person with no name and no birth date, and the software will not even complain. I think it also auto-fills a person's gender based on their name.

Gramps has plenty of visualization options, and if you're not a software developer you can use one of them and stop reading here. Luckily I am a software developer, and un-luckily I couldn't find any visualization that did what I wanted. I was looking for a compact graph that I could fit in an A4 page, but all I could generate in Gramps' GUI were trees with very generous white space between nodes. This is in no way a point against Gramps, by the way: no software package can be everything for everyone, and I don't think I've ever had a graphic design project that I didn't need to hand-tweak one way or another. But one way or another, it was now time for massaging some data.

Tweaking the graph

After exporting the Gramps database to a CSV file, I decided to use Graphviz for automatic node placement. Graphviz is my go-to tool for drawing complex graphs, and I knew it could get the job done, but first I would have to convert my CSV file to Graphviz' DOT format. So I wrote the following script, which generates a DOT version of the input file.

#!/usr/bin/python3

import csv

# Counter with my own internal IDs, for fake nodes
counter = 1000
# Maps people to their intermediate relation nodes
relation = dict()
# Maps between Gramps IDs and my own internal ones
old2new = dict()

print("digraph G {")

# This CSV file should contain the following columns:
# 1: Name, 2: Last name, 4: Father ID, 5: Mother ID, 6: Partner
# A default Gramps export should be already structured like this.
with open('gramps_database.csv') as csvfile:
    datareader = csv.reader(csvfile)
    for row in datareader:
        id = row[0]
        # By default, you are in a relation with yourself
        # Useful for single parents
        relation[id] = counter
        counter += 1
        name = row[1] + " " + row[2]
        parents = None
        partner = None
        if row[4]:
            father = row[4]
            parents = relation[father]
        if row[5]:
            mother = row[5]
            parents = relation[mother]
        if row[6]:
            partner = row[6]
            if partner not in relation:
                relation[partner] = relation[id]
            else:
                relation[id] = relation[partner] 
        # Note that we have all information we need for this person,
        # we are not dependent on a future iteration.
        # Therefore, we can print this section of the tree.
        # First, the name
        print("{}[label=\"{}\"];".format(id,name))
        # Now, relations
        if parents is not None:
            print("{} -> {};".format(id, parents))
        if partner is not None:
            print("{} -> {} [dir=none];".format(id, relation[partner]))
print("}")

I then fed this output (tree.dot) to Graphviz. After playing with different visualization options, I settled for the following options:

sfdp -Tpdf -Gsplines=true -Goverlap=false tree.dot -o tree_output.pdf

which ended up looking like this:

This is still not perfect, but it's 85% there and it's something I can easily work with. I imported this graph into Inkscape, moved nodes around until they fit as tightly as I wanted, and added some colors.

Border

One of my known weaknesses is that I'm very picky on the graphic design of my projects: why stop at doing something right, when you can do it right and good looking? For this reason, I am constantly collecting pictures of cool posters, certificates, book covers, t-shirts, and pretty much everything that can be designed for print. You never know when something might come useful as inspiration!

Interesting enough, my collection of museum family trees was not as helpful as I expected: while very nice to look at, I found most trees difficult to understand, too simple, or too much illustration and not enough content (most artists are really invested into the "tree" metaphor). What did help me, however, was my collection of old-fashioned documents: the border of this tree is based on a 1923 Share Certificate from the "Charlottenburger Wasser- und Industriewerke AG", as seen in Berlin's Museum in Alten Wasserwerk.

Using it as a template, I created a vector version in Inkscape and placed it around my tree, which gives it a nicer "feel" and leads us to this version (with names removed intentionally):

The space on the right was originally reserved for legends, references, and whatever else, but so far I haven't come up with anything. And of course, I also need some space for future family members.

Final steps

The final step was to share it with my family, which was done with a combination of e-mail and telephone calls. About once a month I get a list of corrections from family members, which I do directly in Inkscape because at this point it's faster. I still keep the Gramps database updated, though, as I never know which future projects might require it.

The final tree can be printed in A4, as desired, but I wouldn't recommend anything smaller than A3. As a service for my readers, and painfully aware of how difficult it is to find good frames, you can download the empty template in .SVG format following this link to use in your own projects.

April 21: see the "Update" section at the end for a couple extra details.

A very common operation when programming is iterating over elements with a nested loop: iterate over all entities in a collection and, for each element, perform a second iteration. A simple example in a bank setting would be a job where, for each customer in our database, we want to sum the balance of each individual operation that the user performed. In pseudocode, it could be understood as:

for each customer in database:
    customer_balance=0
    for each operation in that user:
        customer_balance = customer_balance + operation.value
    # At this point, we have the balance for this one customer

Of all the things that Python does well, this is the one at which Python makes it very easy for users to do it wrong. But for new users, it might not be entirely clear why. In this article we'll explore what the right way is and why, by following a simple experiment: we create an n-by-n list of random values, measure how long it takes to sum all elements three times, and display the average time in seconds to see which algorithm is the fastest.

values = [[random.random() for _ in range(n)] for _ in range(n)]

We will try several algorithms for the sum, and see how they improve over each other. Method 1 is the naive one, in which we implement a nested for-loop using variables as indices:

for i in range(n):
    for j in range(n):
        acum += values[i][j]

This method takes 42.9 seconds for n=20000, which is very bad. The main problem here is the use of the i and j variables. Python's dynamic types and duck typing means that, at every iteration, the interpreter needs to check...

• ... what the type of i is
• ... what the type of j is
• ... whether values is a list
• ... whether values[i][j] is a valid list entry, and what its type is
• ... what the type of acum is
• ... whether values[i][j] and acum can be summed and, if so, how - summing two strings is different from summing two integers, which is also different from summing an integer and a float.

All of these checks make Python easy to use, but it also makes it slow. If we want to get a reasonable performance, we need to get rid of as many variables as possible.

Method 2 still uses a nested loop, but now we got rid of the indices and replaced them with list comprehension

for row in values:
    for cell in row:
        acum += cell

This method takes 17.2 seconds, which is a lot better but still kind of bad. We have reduced the number of type checks (from 4 to 3), we removed two unnecesary objects (by getting rid of range), and acum += cell only needs one type check. Given that we still need checking for cell and row, we should consider getting rid of them too. Method 3 and Method 4 are alternatives to using even less variables:

# Method 3
for i in range(n):
    acum += sum(values[i])

# Method 4
for row in values:
    acum += sum(row)

Method 3 takes 1.31 seconds, and Method 4 pushes it even further with 1.27 seconds. Once again, removing the i variable speed things up, but it's the "sum" function where the performance gain comes from.

Method 5 replaces the first loop entirely with the map function.

acum = sum(map(lambda x: sum(x), values))

This doesn't really do much, but it's still good: at 1.30 seconds, it is faster than Method 3 (although barely). We also don't have much code left to optimize, which means it's time for the big guns.

NumPy is a Python library for scientific applications. NumPy has a stronger type check (goodbye duck typing!), which makes it not as easy to use as "regular" Python. In exchange, you get to extract a lot of performance out of your hardware.

NumPy is not magic, though. Method 6 replaces the nested list values defined above with a NumPy array, but uses it in a dumb way.

array_values = np.random.rand(n,n)
for i in range(n):
    for j in range(n):
        acum += array_values[i][j]

This method takes an astonishing 108 seconds, making it by far the worst performing of all. But fear not! If we make it just slightly smarter, the results will definitely pay off. Take a look at Method 7, which looks a lot like Method 5:

acum = sum(sum(array_values))

This method takes 0.29 seconds, comfortably taking the first place. And even then, Method 8 can do better with even less code:

acum = numpy.sum(array_values)

This brings the code down to 0.16 seconds, which is as good as it gets without any esoteric optimizations.

As a baseline, I've also measured the same code in single-threaded C code. Method 9 implements the naive method:

float **values;
// Initialization of 'values' skipped

for(i=0; i<n; i++)
{
    for(j=0; j<n; j++)
    {
    acum += values[i][j];
    }
}

Method 9 takes 0.9 seconds, which the compiler can optimize to 0.4 seconds if we compile with the -O3 flag (listed in the results as Method 9b).

All of these results are listed in the following table, along with all the values of n I've tried. While results can jump a bit depending on circumstances (memory usage, initialization, etc), I'd say they look fairly stable.

Final thoughts

I honestly don't know how to convey to Python beginners what the right way to do loops in Python is. With Python being beginner-friendly and Method 1 being the most natural way to write a loop, running into this problem is not a matter of if, but when. And any discussion of Python that includes terms like "type inference" is likely to go poorly with the crowd that needs it the most. I've also seen advice of the type "you have to do it like this because I say so and I'm right" which is technically correct but still unconvincing.

Until I figure that out, I hope at least this short article will be useful for intermediate programmers like me who stare at their blank screen and wonder "two minutes to sum a simple array? There has to be a better way!".

Further reading

If you're a seasoned programmer, Why Python is slow answers the points presented here with a deep dive into what's going on under the hood.

April 21 Update

A couple good points brought up by my friend Baco:

• The results between Methods 3, 4, and 5 are not really statistically significant. I've measured them against each other and the best I got was a marginal statistical difference between Methods 3 and 5, also known as "not worth it".
• Given that they are effectively the same, you should probably go for Method 4, which is the easiest one to read out of those three.
• If you really want to benchmark Python, you should try something more challenging than a simple sum. Matrix multiplication alone will give you different times depending on whether you use liblapack3 or libopenblas as a dependency. Feel free to give it a try!

If you are one of the lucky ones who didn't lose their jobs during the current pandemic, you may also be one of the involuntary volunteers taking part in the world's largest work-from-home experiment.

There are several well-written guides on how different companies and individuals do it. Software developers, who have been more or less working from home for the last decade, have particularly good tips - if you haven't yet read any of those guides, you can find a great thread here with practical suggestions.

A less-talked point, and the topic of this article, is learning how to work asynchronously with your team. That means: adapting your team's work style not to rely on other people being available to you right now and, conversely, not being "on call" for everyone all the time.

Synchronous work in the time of Cholera

In a regular, sane office setting the rules are clear: if my colleagues need something from me, they can come by and ask. If I'm busy they'll notice it, either because I'm with someone or because my door is closed. But otherwise, answering their question is literally my job, and when they ask me they get an immediate response. This is a synchronous situation.

Remote work, on the other hand, does not lend itself well to this. There can be a multitude of invisible reasons why I can't answer your question right now: I could be on the phone, stretching from hours sitting on my uncomfortable chair, discussing lunch with my (also at home) partner, answering the door, or multiple other situations that rarely happen in an office setting and cannot be conveyed with a status icon.

For the synchronous worker, this situation is unacceptable. They will send you an e-mail and five minutes later call to ask whether you received it, or are the type of bosses who film their employees to check whether they are working. Working synchronously while remote means being constantly dragged to pointless calls ("hey, quick question!"). You dread leaving your desk because you know there will be endless redundant notifications waiting for you. And you end the day exhausted from switching tasks all the time.

A new paradigm

For a term that's so difficult to write, working asynchronously is rather simple.

As part of a team, you need to embrace the idea that you cannot know whether your coworkers are busy. For all you know, they are in a different time zone and asleep. If you need someone's input you send them an e-mail or an instant message (see below) and then... you move on to something else. Your coworker will reply when it's convenient for them to do so. And meetings must be agreed beforehand - no more calling out of the blue!

As an individual, you need to mute your notifications. I recommend that you turn off all notifications except for when someone "tags" you (that is, they write your username and you get notified) for important topics that require your input. That doesn't mean that you can go fishing during work hours - you will still work, reply within a reasonable time frame, and be reachable when time is of the essence. But you no longer have to reply right now.

Working like this presents many advantages. Important points that would previously be discussed and forgotten are now being communicated on a need-to-know basis via e-mail; workers can easily step away from their desks knowing that nothing critical will be missed, and will stop spending valuable time catching up on notifications. And did you know that longer periods of interruption-free work can save up to 40% of your productivity?

"Asynchronous" does not mean "hermit"

The most common argument against asynchronous work is that you lose "human contact". And, while misguided, that's not a bad point: loneliness is a big factor for lots of people right now, and talking to their coworkers helps them cope in these uncertain times.

Luckily, there are solutions. A classic one is the "random" channel: a virtual board where people are encouraged to post non-work-related topics and discuss about their day. Another one is the "virtual office lunch" where everyone gathers for an hour with their cameras on and just talk.

And this is perfectly fine! I'm not saying "do not use Skype ever again". But I hope you'll agree that there's a difference between "I'm in this meeting because we planned it", "I'm in this meeting to waste some time because it's fun", and "I'm in this meeting to waste my time because someone thinks my time is worthless".

Working asynchronously is a team effort, where everyone needs to be on the same page. But it's not hard to pull off, and it can make your day more manageable. It might not be as helpful as toilet paper or flour, but it's a good start nonetheless.

If you're a software developer, you know the rules: new year, new programming language. For 2020 I chose Rust because it has a lot going for it:

• Memory safe and high performance
• Designed for low-level tasks
• Backed by Mozilla, of which I'm a fan
• Named "most loved programming language" for the fourth year in a row by the Stack Overflow Annual Survey

With all of these advantages in mind, I set to build something concrete: an implementation of the Aho-Corasick algorithm. This algorithm, at its most basic, builds a Trie and then converts it into an automaton, with the final result being efficient search of sub-strings (why? I hope I can write about why in the near future). It also seemed like the type of problem you'd like to tackle with Rust: implementing a Trie in C requires some liberal use of pointers, a task for which I had expected Rust to be the right tool (memory safety!). And since I need to run a lot of text through it, I need it to be as fast as possible.

So how did I fare? Two weeks into this project, Rust and I have... issues. More specifically, I'm having real trouble figuring out what is Rust good for.

Part I: Pointers and Strings are too complicated

Dealing with pointers is straight up painful, because allocating a piece of memory and linking it to something else gets very difficult very fast. I followed this book, titled "Learn Rust With Entirely Too Many Linked Lists", and the opening alone warns me that programming a linked list requires learning "the following pointer types: &, &mut, Box, Rc, Arc, *const, and *mut". A Reddit thread, on the other hand, suggests that a doubly-linked list is straightforward - all you need to do is declare your type as Option<Weak<RefCell<Node<T>>>>. Please note that neither Option, weak, nor RefCell are mentioned in the previous implementation...

So, pointers are out as killer feature. If optimizing memory usage is not its strong point, then maybe "regular" programming is? Could I do the rest of my text handling with Rust? Sadly, dealing with Strings is not great either. Sure, I get it, Unicode is weird. And I can understand why the difference between characters and graphemes is there. But if the Rust developers thought long and hard about this, why is "get me the first grapheme of this String" so difficult? And why isn't such a common operation part of the standard library?

For the record, this is a rhetorical question - the answer to "how do I iterate over graphemes" (found here) teaches us that...

• ... the developers don't want to commit to a specific method of doing this, because Unicode is complicated and they don't want to have to support it forever. If you want to do it, you have to pick an external library. But it won't be part of the standard library anytime soon. At the same time, ...
• ... they don't want to "play favorites" with any specific library over any other, meaning that no trace of a specific method is to be found in the official documentation.

The result, then, is puzzling: the experts who designed the system don't want to take care of it, the official doc won't tell you who is doing it right (or, more critical, who is doing it wrong and should be avoided), and you are essentially on your own.

Part II: the community

If we've learn anything from the String case, is that "just Google it" is a valid development strategy when dealing with Rust. This leads us inevitably to A sad day for Rust, an event that took place earlier this year and highlighted how bad the Reddit side of the community can be. To quote the previous article,

the Rust subreddit has ~87,000 subscribers (...) while Reddit is not official, and so not linked to by any official resources, it’s still a very large group of people, and so to suggest it’s "not the Rust community" in some way is both true and very not true.

So, why did I bring this up? Because the Reddit thread I mentioned above displays two hallmarks of the type of community I don't want to be a member of:

• the attitude of "it's very simple, all you need to create a new node is self.last.as_ref().unwrap().borrow().next.as_ref().unwrap().clone()"
• the other attitude, where the highest rated comment is the one that includes nice bits like "The only surprising thing about this blog post is that even though it's 2018 and there are many Rust resources available for beginners, people are still try to learn the language by implementing a high-performance pointer chasing trie data-structure". The fact that people may come to Rust because that's the type of projects a systems language is supposedly good for seems to escape them.

If you're a beginner like me, now you know: there is a good community out there. And it would be unfair for me to ignore that other forums, both official and not, are much more welcoming and understanding. But you need to double check.

Part III: minor annoyances

I really, really wish Rust would stop using new terms for concepts that already exist: abstract methods are "traits", static methods are "associated functions", "variables" are by default not-variable (and not to be confused with constants), and any non-trivial data type is actually a struct with implementation blocks.

And down to the very, very end of the scale, the trailing commas at the end of match expressions, the 4-spaces indentation, and the official endorsement of 1TBS instead of K&R (namely, 4-spaces instead of Tabs) are just plain ugly. Unlike Python, however, Rust does get extra points for allowing other, less-wrong styles.

Part IV: not all is hopeless

Unlike previous rants, I want to point out something very important: I want to like Rust. I'm sure it's very good at something, and I really, really want to find what it is. It would be very sad if the answer to "what is Rust good for?" ended up being "writing Rust compilers".

Now, the official documentation (namely, the book) closes with a tutorial on how to build a multi-threaded web server, which is probably the key point: if Rust claims that error handling and memory safety are its main priorities are true, then multi-threaded code could be the main use case. So there's hope that everything will get easier once I manage to get my strings inside my Trie, and iterating over Gb of text will be amazingly easy.

I'll keep you updated.