17 KiB
Compañeros de viaje
Programadores famosos
A través de la historia de la computación ha habido personas que han demostrado increíbles capacidades a la hora de crear código. Residen en el panteón de los grandes programadores informáticos: Ada Lovelace (la primera programadora de computadoras), Dennis Ritchie (el creador del lenguaje de programación C), Rear Admiral Grace Murray Hopper (creador de COBOL y acreditado por encontrar el primer "error" (N. del traductor: un error informático en inglés es llamado "bug" que también se puede traducir como bicho o insecto) informático documentado) y muchos más. También tenemos desarrolladores en nuestras propias comunidades que tienen un cierto grado de notoriedad, ya sean las personas que escribieron el lenguaje de programación que usamos actualmente, las personas que mantienen el sistema operativo que usamos o alguien que saltó a la fama en nuestra comunidad preferida. Puede ser intimidante cuando nos comparamos con estas grandes personas. Después de todo, cualquier cosa en la que estemos trabajando actualmente podría no estar a la altura de lo que ellos hayan hecho. Peor aún, podemos estar trabajando en algo similar y sentir que cualquier cosa en la que estemos trabajando nunca estará a la altura de lo que estas personas han logrado. Incluso podemos ser amigos de programadores que parecen resolver las cosas mucho más rápido y de una manera más elegante que nosotros y maravillarnos de cómo parecen tener este cuerpo de conocimiento al alcance de la mano que posiblemente no podamos entender.
Backstage vs. performance
One of the best pieces of advice I have received about comparing ourselves to others is realizing that the comparison is between our backstage versus their performance. The metaphor draws from the theater, where performers know every thing that isn't right about their own theater-group's performance and unfairly compare it with another theater-group's finished performance. This metaphor is useful to us because we tend to compare all of the things that we know (the long hours of unproductive coding, the struggles with learning, and so forth) with the finished product of someone else. We don't see their struggle in getting things to work, or their countless hours making crappy prototypes and unfinished projects before making the thing we admire. Allow yourself to have a messy back-stage and do plenty of rehearsals and understand that it takes effort and practice to put on a good performance.
The lure of the post-mortem
There's a tradition in some programming projects (especially in game development projects where there is a clear end to the project when the product ships) of doing a post-mortem on the project. What the post-mortem does is allow the developers of a project to state what went right and what didn't go right. The better ones tend to be frank accounts of the successes and failures of a project.
The post-mortem can be a fascinating look into the development of a project. I've found myself reading a lot of these looking for insights into the development process. But there's a subtle trap in the post-mortem: they're a recollection of events from the vantage point of a successful (or unsuccessful) project. They're a recollection of someone who worked on a project that was successful enough for you to spend time reading about that project's ups and downs. They're written from a perspective where the success of the project is a foregone conclusion (or they're written about projects that were noteworthy in how they failed or didn't meet the expectations of those involved). It can lead to the belief that what you're working on is not as important as the things that other people are working on. But we don't know the importance of our project in real-time. Even the folks in the post-mortem didn't know if their project would work or be successful as they were working on it. Our projects may never see the light of day, or they might be something that changes the world. We can't know the value of what we're working on while we're working on it (though we can have a sense of whether or not we feel our work is important).
A post-mortem also has the benefit of hindsight. Decisions that were clear and definite at the time might not make much sense when viewed with data obtained later in the project's lifespan. There's also an issue with "selective memory" where something might not be remembered with the same clarity, or may be conflated with other events. Confident statements like "We knew this one thing wouldn't have worked" might actually have been "We weren't sure if this would work so we tried several things. They all didn't work.". Consider anyone writing about their past as an unreliable narrator. True, they may be the best and most knowledgeable narrator we have, but they do not have an objective perspective on whatever they were creating. They have their own biases and reasons for the stories they present in a post-mortem. Treat a post-mortem like you would treat an auto-biography of a famous person: a primary source with an agenda to show the subject in the best way possible.
There's nothing wrong with reading a post-mortem about a project --- we can learn a great deal about how a project is run (or shouldn't be run) and what pitfalls to be aware of if we go down a similar path, but understand that you're reading one account (whether by one person or one team of people). They have the perspective of someone deep in the conflict. You're looking at their recollections of tactics, not the overall strategy that brought that them to the place.
Ranking programmers
There are many metrics which folks use to rank programmers. You've likely seen these metrics manifest themselves in different ways: competition sites, numbers of commits to projects, productivity measurements, time to turn-around code, and good ol' fashioned gut feelings. We do it to ourselves and others. We compare our work against the work of our peers and folks that we admire, but that can lead us to make comparisons that aren't objective or based on all of the data. I can compare myself against folks who do low-level programming and find that I don't measure up in that realm. Never mind that I haven't done a whole lot of low-level programming; the comparison is valid. Or, I can compare myself against folks who were mentored by programmers whose names are legendary in the field. I will find gaps between my knowledge and their knowledge, because I didn't have access to those mentors (or worse: I didn't take advantage of the mentors I could have accessed. Whoops!). Comparisons like these are not helpful and lead us into punishing ourselves for not being someone else. Our assessment of our projects and history give us the conclusion that we're not that other person, nor could we ever be that other person.
The major problem with ranking programmers (or ranking anything for that matter) is that ranking systems are based on one set of criteria. There is no real standard for ranking programmers. Sites that rank programmers based on numbers of problems solved or difficulty of problems solved have only determined that there are a set of programmers who really enjoy solving these types of problems. They've also collected a set of programmers who will spend the time and effort to solve these problems and will be competitive while solving them. It tells us little about the programmer's abilities outside of that domain.
There are also other metrics to rank programmers. One classic metric is reviewing how many lines of code a programmer used in order to solve a problem (this is sometimes referred to as "code golf", where the fewer number of lines of code the better the solution). We can argue how "clean" the solution is (clean being another nebulous term). We can determine the "Big O notation", a notation used to describe the performance or complexity of the algorithms that a programmer used in their code. We can stress test the code to determine how well the code adapts to various circumstances. We can count the number of cycles a particular piece of code takes in order to run and benchmark it against similar code. Very little of this tells us anything about a particular programmer. What it does tell us is that the programmer has experience that lead them to that particular solution. It tells us that the programmer has seen these sorts of problems before and cared deeply enough about the problem to think hard enough about how to make a better solution. We learn that the programmer devoted time and energy to practice these sorts of problems. What it doesn't show us is an overall measurement of the programmer's skills or abilities. It's similar to the apocryphal tale of a brilliant professor. This professor was an absolute genius in his field and was one of the go-to people for answers about his subject, but despite his brilliance he was unable to understand how to change a tire on a car. Does that mean the professor was not as brilliant as folks claimed him to be? Hardly. It means the professor spent more time thinking about his profession than he spent thinking about changing tires. The same is true for programmers. If a programmer spends most of their time solving a particular set of problems they will eventually become skilled at those sorts of problems. But if that programmer struggles with a different sort of problem it doesn't discount their overall skills; it just points out areas they might want to work on.
Measuring programmer output
There's also a tendency to measure programmer productivity by how many contributions the programmer can make to a project.Under certain version control systems these are called "commits". They list out a set of changes that the programmer wishes to make to the code. In an era of social coding sites like Github and Gitlab we can easily review what other programmers are committing. Since we can measure the number of commits, we can use this measurement to feel that we're not generating the same number and frequency of commits as other programmers. And unlike measurements of old (lines of code in particular, which measures how many lines of code a programmer adds to a program) we can review the quality of their commits to a project. It can be daunting to see a lot of quality work done by our peers. It can also be source of frustration and feelings of inadequacy. "Why can't I be as productive or contribute as this other person?" we ask ourselves.
Even more frustrating is when others use these metrics to judge productivity and code contributions. We may find ourselves being criticized for our output (or lack thereof).
Commits and lines of code are the most visible measurement of coding productivity, but they don't show much about the actual practice of programming. We can't measure the amount of time thinking about the problem just by looking at a commit. We don't see the mounds of reference material the programmer used in order to figure out a solution, and we certainly don't know if this commit is the result of one afternoon of work or many days of work (unless they commit more often). We might even find out that this person is acting as the focal point of an organization and is folding the work of multiple folks into their commits.
Measuring ourselves on the quantity of others output is easy and seductive but it isn't useful for figuring out how to improve ourselves in relation to the other programmer (other than "generate more commits"). That way of thinking can lead us to believing that we're not spending enough time doing "actual programming" and lead to overwork, stress, and burn-out.
Traveling Companions
There are times when it is useful to compare ourselves with other programmers. Sometimes we can learn about new technologies or new methodologies by looking at the work of others, but it's easy to fall into the trap of thinking that because we're not at the level of other programmers we're somehow inferior to them. Rather than looking at other programmers as competition we should look at them as companions. We're all in this profession working to make our respective projects better. With Free and Open Source Software we have a unique opportunity to see how other folks do their work in public. We can learn from the code of others much in the same way that scientists can look at the papers of other scientists to see what worked (and can improve the validity of the paper with replication and repetition). No coder is completely isolated from the work of others. (It's a rare programmer that has coded their entire programming environment from scratch without the work of others.) We all learn from each other, but rather than be intimidated by the work of others we can instead take it apart and learn from it. If we're lucky we can take the opportunity to ask them how the code works and why they chose to write the code in that way.
There's value in asking questions of our fellow programmers. We tend to overlook asking questions for fear that we're going to ask something obvious or ask a question that will make us feel inadequate for asking. Asking questions is very useful when we don't understand what is going on with an idea or a particular piece of code. There are programmers out there who don't mind answering questions, and my hope is that you find them. Granted there are some programmers who are very busy and might not have the time or inclination to answer questions, but if we are truly stuck and have exhausted all other avenues perhaps we can ask questions of them that don't require much of their time and effort. They may even be grateful for the question because it gives them insights into a perspective they might not otherwise have. When we ask questions we initiate a sharing of ideas in both directions.
There is an art to asking questions and it can be frustrating when folks don't answer our questions or come back with other questions and suggestions that are less than helpful. This manifests itself in exchanges where person A asks: "I'd like to know how to do X" and persons B and C respond "I would do Y instead". It's frustrating when folks won't answer our questions directly. It's also easy to get embroiled in exchanges with folks about the merits of doing Y where Y was suggested by someone else that has nothing to do with the original question about X. But if we re-frame the experience as "this person is trying to help me; perhaps there is something in this recommendation that might be helpful" then we can have a better conversation. Perhaps what we're asking isn't the best way to do something and pausing to listen may help us better understand why they suggested what they did.
Pulling our egos out of the question allows us to be more open to the answers we receive. When people don't understand our question it becomes easy to take it personally and frame it as "they're not understanding me" or "they're not listening to me". Pulling ourselves out of the question allows us to take the merits of the answer provided "as-is" and gives us the ability to change the question as needed to get better answers.
Of course there are folks who won't respond with your best interests at heart and are only interested in imposing their own world-view upon you. Instead of answering your question they question why you're doing that at all and suggest that you should be using their methodology instead. It can take a lot of energy to engage with these folks to tell them "no, I really, really intended to learn more about X." I wish I had good answers for how to handle these folks. There are plenty of them that feel that whatever they're doing is the only right path and those that stray from their chosen path are anathema to their world. My best suggestion is to thank them for their time and ask someone else for help. Perhaps they may be useful in the future when you have questions about whatever is part of their agenda, but for now be as kind as possible and wish them well on their programming journey. Technology spaces have a lot of folks who have been working with computers for a long time and have formed strong opinions about their tools and technologies. My hope is that you can find the ones that are also kind and willing to share what they know and not badger you with their strongly held beliefs. Over time you too will form your own beliefs on what works and what doesn't work and pass that knowledge on to others. Recognizing folks who are there to help educate and those who are there to proselytize is part of our growth process.
If we look at other programmers as our traveling companions on this journey; as peers in our coding practice, then we can realize that we're all in this together. Even someone with many more years of experience than you have is your peer. You have knowledge and experience they won't have, and they have experience and knowledge you don't have. If we strip away the barriers of perceived rank and meritocracy we can better engage with and learn from each other.
The journey to becoming a better programmer is long and hard. We need the best companions we can find to help us along the way. We need more than just the technically-skilled companions; we also need companions we can talk to when the day is done. We need companions we can sit with around the proverbial campfire where we can laugh and commiserate about our struggles together.