Edsger Dijkstra carried computer science on his shoulders (2020)
inference-review.com
inference-review.com
> "He also never purchased a computer. Eventually, in the late 1980s, he was given one as a gift by a computer company, but never used it for word processing. Dijkstra did not own a TV, a VCR, or even a mobile phone. He preferred to avoid the cinema, citing an oversensitivity to visual input. By contrast, he enjoyed attending classical music concerts."
This is very fascinating to me, as a generally unsure person, I can't imagine writing anything this way let alone scientific papers. It certainly requires a deep focus and great knowledge on what you are writing about. I will try to give it a go next time I write out my thoughts.
There is interesting interview where he goes deeper into this. https://www.youtube.com/watch?v=P0w1MJHxStg
Writing with a pen I tried, but no-one, including me, can read it after...
I don't know about anybody else, but I spend 8-18 hours a day attached to a mouse and keyboard. I'm not going to use one that doesn't feel good to me.
Buying nice peripherals is like buying nice clothes.
Perhaps some people just really need to take that to extremes, getting out of their head all their fragmentary or preliminary thoughts, and filtering and working them down on paper or the screen?
It generally seems (to me) that a good-heavy amount of thinking first works best and/or is most productive, but overall, we all have different brains, training, and experiences, so I'd go with: "whatever works for you" (but try multiple methods).
Human thought speed seems to be a reflection on how fast you need to think for normal conversations. Thus orders like "run" or "jump" are likely to be fast as if you need to tell someone to escape a lion speed is important. While if you are planning a hunt you are allowed a lot more time to get details right. Engineering often is a very slow activity as you need to think about a lot of details but you have time. Typing a message like this I want to be fast enough to follow my thought speed which is much faster (I won't claim to type this fast, but I get close and can believe with more practice I would be this fast)
When corresponding, I type at around 120 wpm. When coding, it's got to be closer to 5 or 10. But I really like this idea that it would be useful to push that number lower (while still focusing on the task at hand).
Thing is, I got into this field because I really enjoy interacting with computers via keyboard. It's super satisfying to punch some buttons and make things happen. This is what I love about vim: right off the bat, a single button press feels like it does a lot. Then you punch a whole bunch of buttons to make a new button make something happen. Nevermind that it doesn't achieve product goals. I'm here for the button pressing, and for the creating of new buttons to be pressed.
I can type at >100 WPM (just tested). I recently had to write a 3-page document for work, the summary of weeks of research and interviewing. It wound up 1959 words (just checked). I wrote it over the course of one 8 hour day, after having procrastinated on it for the previous 2 days.
I came home at the end of that day exhausted because I felt like I finally did weeks of work in one day.
But, this 8 hour day had 2 hours of meetings and 1 hour for lunch. Say another 1 hour for random slacking. So pure concentrated writing was 4 hours or 240 minutes, which winds up to about 8 words per minute. That's it!
And I would consider this fairly fast, and again I'm not including the 2 days where I tried to write the doc at various times but failed to get started. SOMETHING was spinning in my brain in the background getting the ideas clarity.
And we're talking journalistic opinion piece, which are a dime a dozen, not deeply nuanced scientific writing. If anything Dijkstra was writing his notes too fast and loose!
I spend 90% of my time formulating descriptions of the problem and the desired end state
Hallucinating futures where the state of the world is in a state that I either wanted to be or that somebody’s asking me to build
Once you know your final end state, then you need to evaluate the current state of the things that need to change in order to transition to the final state
Once you have your S’ and S respectively then the rest of the time is choosing between hallucinations based on sub-component likelihood of being able to move from S to S’ within the time window
So the process is to basically trying to derive the transition function and sequencing of creating systems and components that are required, to successfully transition from state S to state S'
So the more granular and precise you can define the systems at S and S' then the easier it is to discover the likelihood pathway for transitional variables and also discover gaps, where systems don't exist, that would be required for S'
Said another way: treat everything - both existing and potential futures- as though they are or within an existing state machine that can be modeled. Your task is to understand the markov process that would result in such a state and then implement the things required to realize it.
The religious call this "Prayer"
Others call it "Manifesting"
So you can basically hand me any problem and I will implement this process, and I have a high success rate for delivering desired outcomes.
And again, this isn’t really like my process I invented. It boils down into a practical implementation of a Markoff process in planning as applied to any set of tasks such they could be discretely described as a state machine.
The key challenge IMO is in describing the state machine, and that is what takes a lot of elucidation.
In many cases we don’t have the ability to precisely describe a process as a state machine because we haven’t defined the boundaries of the system, and then measured it enough, in enough different dimensions, across enough time to be able to give that level of understanding to input an outputs.
You use a problem solving process built on structured analysis by first defining the problem in terms of an Undesired Result (R1), Desired Result (R2) — the S and S’ in Andrew’s process. Then, you determine the Starting Point in terms of the logical structures that generate your R1; these structures can be a sequence of cause-effect, a structural decomposition (e.g. of organization, geography, etc.), a classification, or some combination of the three. From this structure you can hypothesize experiments to confirm/disconfirm where the causes are. With these causes in hand, you can generate possible solutions or corrective actions. Finally, you’d evaluate your alternatives and arrive at your solution to move from R1->R2.
MPP’s problem solving process has the additional advantage of structuring your actions/results in a way that makes writing a document or presentation simple and straightforward, to convince others for example.
Check out the book if you’re interested in improving your problem solving and analysis skills.
However you might want to start with The Thinker's Toolkit by Morgan Jones. This gives you a catalog of models which teaches you to structure your requirements born from problem analysis in various ways to aid problem-solving.
Clearly there is something novel about the way Dijistra thought and worked. Most of us don't do things that way - formulate our thinking, then work towards one perfect first draft.
AndrewKemendo saw that and said "I identify, I'm the same way", and tried to describe what his way of thinking sounds like.
If to you it sounds like the exact way of thinking and problem solving as yours or mine, well, then perhaps AndrewKemendo did not describe it well enough. Perhaps it's impossible to describe it to someone else. But the necessary context to his description is that his brain works different from yours or mine. So the short textual description also means something different than what it does for you and me.
Is there the possibility that AndrewKemendo is self-grandiose and not aware that he is not special and his verbose descriptions of his unique way of thinking are nothing but? Sure. But even if we had no other evidence that this is not the case, it would cost you nothing to be curious and assume best intention. But we do have other evidence - and it's in the first line - "There’s no other way to do it for this type of a brain. I know because I have the same type of brain."
The comment just comes off as mere posturing (eg. "I know because I have the same type of brain") and if you actually think about what is written down all i see is empty verbiage and something which could have been said simpler and more directly (for example there is no need to bring in "Markov Processes" here).
Dijkstra's (and Floyd/Hoare's) programming techniques are hard enough to learn that such comments merely obfuscate the essential ideas and pushes people away from trying to study it because it "appears too hard". Things should be made as simple as possible to motivate people to study and learn.
The cliff notes summary is not the same as what the parent tries to convey. The nuance in what the parent wrote was the thing that went wooosh.
I’m not making any bold claims about inventing anything or anything like that in fact, there are multiple times where I say I’m not doing anything special.
I’m just describing that my mental process resembles and steps the same way a markov process does (or OODA or SPA if you prefer) and that not everybody thinks like that and that and I found that certain types of scientific thinkers also have that same type of thinking. The Dykstra explanation there lined up with my experience
I cant control how people react to that so :shrug:
No, but you should be aware of how you come across so that people don't dismiss the substance of what you are trying to say because of the form in which you say it.
It's fine to dislike something - and just move on.
So did the "In this moment I am euphoric" guy.
I am with biggestbrain here. The post reads like "I'm just like Dijkstra, I solve problems".
Doesn’t make sense. Your next step depends on much more than just knowing where you are at S. One needs to account for the history of where you were before.
Or maybe you’re just using technical words with precise meanings to describe a vague imprecise heuristic?
Future reward trajectories are THE core focus of multi-step MDP, see Sutton [1]
"Now we consider transitions from state-action pair to state-action pair, and learn the value of state-action pairs. Formally these cases are identical: they are both Markov chains with a reward process. The theorems assuring the convergence of state values under TD(0) also apply to the corresponding algorithm for action values: "
I wasn't going to differentiate in my original post between sub-types of "cycles" within increasingly complex MDP's for long sequence reward estimation:
Markov processes are nice because they are simple objects and therefore have nice properties and solid mathematical proofs.
Many mathematical models are studied because they have nice theoretical properties and one can prove theorems about them. This should not be mistaken with an actual mechanistic explanation for complex emergent phenomena like human decisions.
This is definitely a learnable skill. However, it might be a lost art with younger generations. I had a year of high school english that taught this. Most of us kids did fine.
The assignments worked as follows:
- Write an multipage essay on this topic; you have a week. (first quarter, repeat 2-3 times)
- Here is your exam question. You can bring one 3x5" notecard with you tomorrow, but you'll need to write the actual multipage essay in 55 minutes. (second quarter; repeat 2-3 times)
- You will be given a question about the book we read, and will have 55 minutes to write a multipage essay answering the question. No supplementary materials allowed (except the book). (last quarter; repeat 2-3 times)
Grading for things like the overall structure of the essay and sentence flow got stricter as the course progressed. So did grading on reading comprehension.
It's physically more straining than to type on any computer keyboard, even with very stiff keys. And I didn't know how to touch-type at the time. So, it was pecking with two fingers. With all that said... three words per minute isn't a lot. At all. By the time I was finished typing, I could probably do better than that.
To me, this quote describes someone who rather hates / struggles with technology. I had a bunch of math professors like this: they couldn't draw diagrams to save their lives, they couldn't write formulas in any editor, and so often we'd have very poorly hand-drawn formulas photographed and inserted as images into MS Word documents or similar incomprehensible junk.
This doesn't necessarily mean they are bad at the very narrow field they specialized in, but it usually means they are very bad at virtually everything else adjacent to that field. Also, this often means that they've made their career in academia by being the first person in that narrow field they specialize in.
This is also my impression of Dijkstra: he had some novel, but also kind of low-hanging fruit ideas... at the time when he was almost alone in that field. Had he started today, he'd be very lucky to get a tenured position, but most likely would end up being a TA / dropped out and just do something else. Even though there's still a lot of politics and all sorts of un-meritocratic ways of getting into prestigious positions in academia, you need to be very good. A lot of the times it's being very good at cheating, but there isn't really any more room for greenfield academics who instantaneously propel themselves to the top of the academic hierarchy.
I'm not really sure how to describe it for me, as I have partial aphantasia, and a kind of weird shape-synesthesisa which also blends with my sensibilities around physical taste. It can really only be described as "feeling" a shape, where a shape represents some quantity of information for a topic, if it were the case that I'm not using receptors in my hands to touch something, but instead, it's the shape feeling its own shape relative to other shapes, and it's not as much a touch-like sensation as a visceral, semi-intuitive experience of the information of whatever that shape-object-whatever represents.
I do not think that thinking in this way was fully formed from a young age (i.e. it used to be an extremely active story-based imagination), it seems to have developed in some manner as an (informationally) dissociative coping mechanism because my autism is strong enough that the world does not stop overstimulating me. When I am in this headspace, I am generally unaware of sensory input, there is only me, and the (generally mathematical) "problem" at hand. Sometimes I can go for hours, working on a problem, moving shapes and concepts around, and fitting them together in a way that explains more while taking less information to do so. This can be both a gift and a hindrance.
I generally have to limit my budget for sensory stimulation and for 'new' things because my brain is so sensitive to information. This generally is more of a disability than a help, on the whole, though it is nice to be able to solve problems in the way that I do, that is one comfort to me.
I do have a tendency to want to fit things in my brain before thinking about them, but the 'onboarding' process can take a while. I used to get in trouble all the time as a kid before this sense by trying to do mental math all of the time, I would sit, think for a long time about a problem, write an answer down, and move on. Unfortunately, instead of being encouraged, it was something I got punished for (I was homeschooled), and there wasn't much in the way of what I would consider appropriate adaptivity in that regard.
Problems are fun, and the hardest part about them I think is that we are generally crap at phrasing them and looking for solutions, most of my work seems to just be cutting through the cruft of whatever trends-du-jour or symbolic obfuscation shenanigans are going on for a given topic. Once you sort out the structure of a given mathematical problem, the answer seems to fall into place well, and one does not always need compute for that.
So, all of that to say, each person thinks and processes things in a different way, and if you're into information theory, then one might say the kl divergence (in a manner of speaking) between how you encode and solve problems might be very high with djikstra's methods of thinking! It depends upon what unique strengths and weaknesses you have, and that can take years to find whichever special problem solving mode fits best to your sensibilities and gifts. For example, I have significant trouble around symbols that have potentially multiple meanings (like, as is often the case with many formal mathematical definitions), and since those take a disproportionate amount of "processing time" for me, my problem solving methods have evolved to replace and/or avoid them unless necessary. That is likely a silly solution for many people!
So, while you may not entirely fit Djikstra's natural proclivities, there might be elements that fit you. But whatever fits you best as a problem-solving technique is likely, frustratingly, and beautifully, entirely unique and specific to you. <3 :'))))
I also try to (and mostly succeed at) hold a whole problem in my head at once. This made me fairly successful at architecting software systems with many moving parts, but it's an incredible drain on my energy, and I would sometimes find myself so exhausted I would go to sleep for hours (i.e. not a simple nap) in the middle of the day. I credit this ability for my current burnout, honestly! It's deeply satisfying to do, but maintaining it week after week amidst all the other responsibilities of my previous job (especially the interpersonal ones) was crippling.
Yes
No one knows the answer, it's an incredibly over discussed topic, and we won't know for sure for many years.
I think those points still apply to AI intelligence today. However, the power of today's AI greatly outstrips anything Djikstra would have seen in his day.
Here's the context since it's been quoted without providing it.
But I think perhaps you've missed his subtle use of language.
Fish swim, obviously. What submarines do in water isn't usually described a swimming.
That's a bit odd in when you think about it as in both cases the objective is to get from A to B, while in the water. In fact if you look at only the outcomes, what fish and submarines do when swimming looks almost identical. They move (or perhaps be stationary in a current), they tend to be efficient about it, they try to not make a lot of noise, they even use very similar mechanisms to move vertically.
Despite that it almost seems that to swim you have to be fish, or a sea snake, or a blue bottle or water bug - but not a submarine. And going by the discussions here to think you have to be a human, or a dog, or just about anything but a computer. And that's true no matter how closely a computer can emulate tasks we say require thinking in a human.
A conclusion you might draw then is whether a submarine swims deciding belongs in the domain of linguistics, not computer science. And he's saying that's true for whether computers "think" too.
So there's a theatrical game being played when interacting with these devices that makes them valuable to the people playing that game.
More generally, when the adtech companies selling the current round of AI do the same, without any irony, it's usually a mixture of charlatanism, selling and legal avoidance.
(Eg., that "ChatGPT wrote X" is a kind of theatrical game wherein OpenAI are the material beneficiaries, and most others, are the loosers).
And it’s perfectly reasonable to say a machine “thinks.” It’s just good to understand that it’s a metaphor and not a literal description of what the machine is doing. I avoid saying machines think because it’s confusing, but in principle it’s fine.
> There is a great deal of often heated debate about these matters in the literature of the cognitive sciences, artificial intelligence, and philosophy of mind, but it is hard to see that any serious question has been posed. The question of whether a computer is playing chess, or doing long division, or translating Chinese, is like the question of whether robots can murder or airplanes can fly — or people; after all, the “flight” of the Olympic long jump champion is only an order of magnitude short of that of the chicken champion (so I’m told). These are questions of decision, not fact; decision as to whether to adopt a certain metaphoric extension of common usage.
2. Asking a networking guy about the use of the term "stack" in his talk. A little unfair, given that the two uses (push/pop stack vs. network protocol stack) are unrelated, but still a bit funny.
The immediate second thought that came was the famous quote “I don't know how many of you have ever met Dijkstra, but you probably know that arrogance in computer science is measured in nano-Dijkstras”, from Alan Kay should I believe https://www.goodreads.com/quotes/69528-i-don-t-know-how-many...
There's the old adage that you "can lead a horse to water but you can't make him drink." I wonder what those who give up on humility think they're actually gaining?
BASICs are perhaps the language HN readers are most likely to have seen which (in some cases) have the GOTO feature Dijkstra wrote the letter about, if you've seen goto in C for example, or C++, that resembles the go-to statement superficially, but in practice it's not very dangerous because it's constrained. Suppose my C++ f() recursive function adds a "goto" - it's not allowed to jump into a different function g(), it's not even allowed to jump into a different iteration of the same function f(), it's only allowed to change which part of the current iteration of the current function happens next, and it can't skip the destruction or initialization of variables, or other book keeping.
The same applies to Dijkstra.
You could start with this account of the Dijkstra-Zonneveld Algol 60 compiler written by Kruseman Aretz: https://ir.cwi.nl/pub/4155.
We are talking about the person that institutionalized the "we people aren't smart enough to do X" line of thinking from software engineering.
Of course, he was also quick to call a spade a spade, but I haven seeing any case of him denouncing something that wasn't obviously bad.
But Dijkstra is the poster child for that sort of thing.
> together with his wife, he purchased a Volkswagen bus, dubbed the Touring Machine, which they used to explore national parks
Maybe not so surprising for a guy who would not indulge in a vcr, cinema, smartphone or even a computer :) more info would be neat if anyone knows from a longer biographic. Photographer? Bird watcher? Hiker? Camper?
It's part of a longer interview with him. He also does crossword puzzles.
A sampling:
> “The competent programmer is fully aware of the limited size of his own skull. He therefore approaches his task with full humility, and avoids clever tricks like the plague.” (Dijkstra, 1972)
> “The use of COBOL cripples the mind; its teaching should, therefore, be regarded as a criminal offense.” (Dijkstra, 1982)
> “Object-oriented programming is an exceptionally bad idea which could only have originated in California.” (Dijkstra, 1989)
Unfortunately, when OOP has become fashionable, its proponents have tried to convince everybody that OOP is the best paradigm for absolutely all programming problems, not only for those where OOP is indeed the best, and they have been rather successful for some time, which was bad for the software industry in general, resulting in many sub-optimal programs.
You end up being forced to co-mingle your data structure design, with your data processing design. Which tends to be rather unhelpful.
Keeping your data structures fairly separate from the processes/functions that operate on them makes it much easier to build composable software. Data structures tend to be very hard to change over time, because the migration process is always tricky. Any structure that’s exposed outside your programs address space, whether that be via an API, storage on disk or in a DB, needs a migration path so new code can deal with old structures correctly. On the other hand, algorithms operating on that data are trivial to change, there’s no migration risk, and indeed we should expect those algorithms to change often, as the requirements of the software change.
In an OOP world, because you’re strongly encouraged to tightly bind your data structures to the algorithms that operate on them. You quickly end up in a horrible situation where changing your algorithms is very difficult without also being forced to change your data structures. Suddenly what should be a simple and easy change (introducing a new way of processing data), becomes difficult because coupling created by objects makes it hard to change the processing without also changing the data structures.
In a simple “write-once” world, OOP is fine. But once you want to write software that’s expected to evolve and adapt over years, as business requirements change, OOP quickly becomes more a hinderance than help.
The fact that a language is strong typed and you get compilation errors if you forgot something is a big plus.
OOP is fundamentally about no static variables.
strong typing has nothing to do with OOP. static variable are replaced by singletons and other similar in spirit objects.
OOP today is a 'No true Scotsman' concept much like communism, agile, and other vague by design terms. You can't argue against it because every individual has at least one definition in his head and it shifts through the conversation toward the one that's not refuted by the claims.
the intent of such concepts is to hit you in the feels and trigger some ideals inside you, so that you associate the positive feelings you have toward those ideals to the concept.
if you argue about it as it is used in practice you will inevitably have someone bring up that java style OOP is not true OOP and that you should look intro Smalltalk or some other language that implement "Real OOP".
since arguing about specifics is a losing battle, lets argue about the bigger picture, if we take the goals of alain kay like he talked about in many of his talks his goals was to make systems more like biogology, but as a system designer the last thing you want to do is that. we don't know much about biology, people in that field a still reverse engineering pre-existing designs to this day and not designing much of their own from scratch. If you design a system you want to have the most control and foresight in the dynamics of your systems, uncontrolled & unintended emergent effects are your source of problems.
entangling data and behavior make your conceptual design state-space size explode, when you go full OOP you become an ontologist philosophizing about what is an X and what is an X-manager, X-Provider, ... and less of a system designer trying to make sure your system does not land in the wrong states.
One of the most influential books about OOP - "Design Patterns: Elements of Reusable Object-Oriented Software" - talks about composition, separating interface from implementation etc - not about hierarchies ( other than to prefer composition ).
Composition has been used since LISP I and ALGOL 60 in almost all programming languages.
"Separating interface from implementation" is also the main feature of the programming languages based on abstract types, starting with Alphard and CLU, which are not OOP languages and which predate the time when Smalltalk has become known to the public, launching the OOP fashion.
An abstract type is defined by an interface, i.e. by a set of functions that have arguments of that type and this is a programming language feature that is completely independent of the OOP features like member functions, virtual functions and inheritance.
All OOP languages have some kind of abstract types, though with a different point of view on the relationships between individual objects and types a.k.a. classes and the functions that operate on them, but there have been many languages with abstract data types without the OOP features.
Moreover "separating interface from implementation" has also been the main feature of all programming languages based on modules, starting with Mesa, Modula and Ada.
The features that identify an OOP language are the idea that the functions belong to individual objects, not to types, hence the member functions, the substitution of the union types (of the right kind, like in Algol 68, not of the pathetic kinds encountered in Pascal and C and derived languages) with virtual functions (to provide an alternative form of dynamic polymorphism, which is preferable for closed-source software, by allowing changes without recompilation, unlike with tagged unions where there are select/case/switch statements that must be recompiled when the union is extended), and the inheritance in a class hierarchy.
There are many languages that are multi-paradigm, like C++ or D, where you can choose whether to write a program in an OOP style or in a totally different style, but there are also languages where it is difficult to avoid using the OOP features, or even impossible, because all the available data types may be derived from some base "object" type, inheriting its properties.
That is not at all unique to OOP, and in fact OOP makes the problem undecidable due to ad-hoc subtyping. More structured languages like ML, Haskell, and Rust are much easier to reason about and have much stronger type systems.
Most people using OOP languages prefer composition over hierarchies and I almost never see those complex class structures modelled on data ( for some of the reasons you give ).
In terms of evolvability - encapsulation - a key feature of OOP - is a key tool in enabling that.
And also, can you give me a huge non-OOP codebase that shows in practice how it is better than the potential OOP implementation?
As for huge codebases, everyone knows that line of codes does not equal quality.
Nothing. It's just one of many useful tools we can use to develop software. It's like with food. Sometimes some people loudly claim that a certain food is unhealthy, and then a few months or years later you read the exact opposite. Fortunately, there are low-pass filters.
And it is the wrong choice, if you need a script and can make that function in python.
OOP, particularly Java, is good when you have a lot of functionality, especially if you can use already existing massive ecosystem.
If I would have to write something from scratch, I would not use OOP, but if you need something fairly complicated, chance is there is already a generic solution available, and you need to costomize it, where Java is pretty good.
You can find some criticism points in there
When you're learning OOP, you deal with questions like "is a Circle an Ellipse or is an Ellipse a Circle?" F'in' neither, actually. You end up "modelling" "business objects" in "code", leading to monstrosities like Hibernate.
In reality, you have the real-world domain and you have the in-the-computer domain, and trying to make one look like the other is a mistake. OOP is a dandy way of managing some forms of complexity, but it's not often an especially good way of looking at most problems, much less the only way.
The phrase "considered harmful" has become a meme and is the go-to phrase (pun intended) for essayists looking to criticise some aspect of computing
https://www.cs.utexas.edu/users/EWD/transcriptions/EWD12xx/E... "I must confess that I was very slow on appreciating LISP’s merits. My first introduction was via a paper that defined the semantics of LISP in terms of LISP, I did not see how that could make sense, I rejected the paper and LISP with it."
For example in a language without GOTO inside an if statement the guarding condition is known to be a predicate (at least until another operation changes it) whereas in a language with unrestricted GOTO there is no guarantee whatsoever that the guard holds since execution could have jumped past it.
"At the I F I P Congress in 1971 I had the pleasure of meeting Dr. Eiichi Goto of Japan, who cheerfully complained that he was always being eliminated."
"For many years, the go to statement has been troublesome in the definition of correctness proofs and language semantics....Just recently, however, Hoare has shown that there is, in fact, a rather simple way to give an axiomatic definition of go to statements; indeed, he wishes quite frankly that it hadn't been quite so simple."
I cannot find anything written by Hoare about it, but Knuth goes on to describe it: the idea is that labels have associated preconditions and GOTOs have the semantics { precondition(L) } goto L { false }.
"Informally, a(L) [my "precondition(L)"] represents the desired state of affairs at label L; this definition says essentially that a program is correct if a(L) holds at L and before all "go to L" statements, and that control never "falls through" a go to statement to the following text. Stating the assertions a(L) is analogous to formulating loop invariants. Thus, it is not difficult to deal formally with tortuous program structure if it turns out to be necessary; all we need to know is the "meaning" of each label."
It is a very nice paper. Some sources:
http://www.kohala.com/start/papers.others/knuth.dec74.html
Thanks for sharing that paper! As an aside, it's really remarkable just how readable the state of the art papers such as the one you shared there from that era are. Either computing science has greatly advanced to the point where clarity is no longer achievable without years of specialized preparatory study or the quality of writing has regressed.
"I don't know how many of you have ever met Dijkstra, but you probably know that arrogance in computer science is measured in nano-Dijkstras"
Which, I dunno, feels kind of funny coming from him of all people.
Not sure what he meant by "object-oriented", but he - as the other members of the IFIP - was well aware of Simula 67, and he belonged to the faction that rejected van Wijngaarden's proposal and regarded Simula 67 as "a more fruitful development" [1]. So - if he said or wrote that at all - it's likely related to what Kay understood by the term, or the implementation as a dynamically typed, originally interpreted language done at Xerox PARC, not to the kind of "object-orientation" for which Simula 67 is recognized today (remember that the term "object-oriented" was originally not applied to Simula 67 by the public, but to Smalltalk).
[1] https://www.researchgate.net/publication/2948437_Edsger_Dijk...
See “TUG LINES”, Issue 32 [1]
> it's likely related to what Kay understood by the term, or the implementation as a dynamically typed, originally interpreted language done at Xerox PARC
I think the comment was referring to what they were doing at Xerox PARC - yes.
There are more recent writings of him taking jabs at Californian universities for teaching Java to freshmen instead of what he was doing in Austin - teaching functional programming via Haskell. [2]
[1] Dijkstra, E.W. Quoted by Bob Crawford. TUG lines, Journal of the Turbo User Group, 32, Aug.–Sep. (1989)
[2] https://www.cs.utexas.edu/users/EWD/OtherDocs/To%20the%20Bud... (2001)
Unfortunately I wasn't able to find [1] anywhere on the web; the CHM only seems to have issues up to 22. Do you have a link where I can access it?
The other one [2] is also interesting, but the claim is not against OO, but "that functional programs are much more readily appreciated as mathematical objects than imperative ones, so that you can teach what rigorous reasoning about programs amounts to", and that "Java [..] is a mess", which again applies to the language quality (see also Brinch Hansen's paper https://dl.acm.org/doi/10.1145/312009.312034), not the paradigm.
My point today is that, if we wish to count lines of code, we should not regard them as "lines produced" but as "lines spent": the current conventional wisdom is so foolish as to book that count on the wrong side of the ledger.
https://www.cs.utexas.edu/users/EWD/transcriptions/EWD10xx/E...
Except I think people have often overly focused on "lines of code" as the unit for this metric, leading many to overrate terse code, even when it is very dense.
But I'm not sure what wording would capture this idea succinctly enough to include it in a pithy quote.
APL (and it's many cousins) is a bridge too far for me, but I can see the appeal to having the entire program on a single screen.
In one short letter he makes an insightful critique yet misses the the point so hard I can practically hear the woosh 42 years later.
It's funny that you used Go as an example. Like 60% of its raison d'être is noticing that more lines of code is often better if each line is simple. And I think it's mostly right about that. Its inscrutable naming conventions are, on the other hand, a poor choice. (Notwithstanding that java's are a poor choice in the opposite direction.)
" a) 2 ≤ i < 13
b) 1 < i ≤ 12
c) 2 ≤ i ≤ 12
d) 1 < i < 13
There is a smallest natural number. Exclusion of the lower bound —as in b) and d)— forces for a subsequence starting at the smallest natural number the lower bound as mentioned into the realm of the unnatural numbers. That is ugly, so for the lower bound we prefer the ≤ as in a) and c).
Consider now the subsequences starting at the smallest natural number: inclusion of the upper bound would then force the latter to be unnatural by the time the sequence has shrunk to the empty one. That is ugly, so for the upper bound we prefer < as in a) and d). We conclude that convention a) is to be preferred."
I despise this guy
You either have an empty range be -1 .. 0, which is ugly, or 0 .. -1, which is also ugly. Thus, start-inclusive, end-exclusive.
This and column-major order for matrices (a vector is a column, not a row)
We can be thankful that this applied physicist in particular decided to dedicate time to think carefully about the needs of who would be his colleagues and successors, when he was pretty much inventing programming as a profession; he told of the anecdote of when he got married, and he wasn't able to write in "programmer" as his occupation in the Dutch paperwork, because it didn't exist as a job category yet.
And then, once you remove empty ranges, there's no reason to to choose a range style based on which you prefer for an empty range.
The problem is not that he self-aggrandizes or is boastful; it's that he keeps tearing other people down. And that he was very willing to offer sweeping pronouncements on matters that he had no practical experience in. He basically stopped touching computers in the early 1970s, and certainly stopped writing practical programs. What basis then, did he have for dismissing object oriented programming, which for all its flaws has done decidedly more for human progress than program verification has?
And as the tense exchange with Backus shows, he was quick to dismiss the relevance of functional programming, which certainly has done a lot for the kind of mathematical reasoning he advocates for programming (Apparently he taught his students Haskell later. Did he ever apologize to Backus for his wrong headed initial assessment?).
Would computer science miss ANYTHING relevant if he had stopped working in the field in 1975? His output basically consisted of proving propositions over cherry picked toy problems. I don't think anybody doubts that this is possible, but is this a viable approach to build real systems (not only large ones, but also ones whose requirements evolve over time)? I consider this, at best, highly unproven, and yet he advocates this as the sole approach to be taught in CS education. Knuth noted, correctly, that neither Dijkstra's education nor any of the practical systems he built (Algol compiler, etc) were based on this approach.
After reading the article it seems that he mostly produced hot takes, while others did the actual heavy lifing. Goto considered harmful? And some algorithm that would probably be found by someone else?
Will Linus get a monument made of pure gold when he dies?
He won a turing award for advocating for structured control flow. That is the modern standard style for programming. He contributed a number of algorithms including the shortest path algorithm.
"You can hardly blame M.I.T. for not taking notice of an obscure computer scientist in a small town in the the Netherlands."
I suspect geeks like Dijkstra because he is "edgy" more than because he is correct.
Also, Object-Oriented programming was actually invented in Norway, even though Alan Kay of Smalltalk fame tried to take credit.
He was right about GOTO though, but many developers did not even understand his argument but just read the headline and concluded "GOTO bad".
https://www.quora.com/What-did-Alan-Kay-mean-by-I-made-up-th...
I don’t know how amenable he was to his arguments possibly being incorrect, but I love working with/knowing people who are that combination of outspoken and not-overly-stubborn. People like that are a firehose of ideas and knowledge, even if not everything they say is correct. They also usually are “passionate” about their work and at least competent enough to have unorthodox opinions that don’t just sound blatantly stupid.
Most people are too timid or low-ability to be outspoken at all.
Not enough of us, that's how many.
Also, if he said this back when the majority of programs were written in assembly I think it makes a lot more sense.
That was never the claim. The claim was humans using "Structured programming" produced code with less bugs than humans using other methods. "Humans" being the key word here, a compiler will happily produce perfect programs every time just using goto's (because there is no choice in assembler). But when humans use goto's it's usually a disaster.
History has vindicated him big time. Many modern languages don't have goto's, precisely for this reason.
While perhaps less directly impactful on the industry, in computer science teaching and academia, there are few names who loom as large as his.
He has touched so many aspects of computer science that we now consider fundamental.
Dijkstra advocated "single entry, single exit" for each control block. Programs should be composed of such blocks. Makes for very neat flowcharts. Good for entry and exit conditions.
Single entry wasn't that controversial. Single exit, though, means no "break", or "continue" for loops, and no early returns from functions. This forces a rather convoluted style. Try writing a loop of the form "get thing, if done quit, process thing, put thing" in that style. You have to have two calls to "get thing", or extra flags.
Turns out you don't need that for program proofs, once you have machine assistance to make sure all the cases were covered. It makes for cleaner hand proofs, though.
(I never met Dijkstra. Knew people who worked with him.)
Sounds like pretty standard functional programming a la Standard ML or Haskell, or am I missing something?
It may be standard for them, but it's why functional languages are niche while C, C++, C#, Java, JavaScript and Python dominate.
To be clear, I agree that it could be a reason, along with a multitude of others. I think that discerning which is the most substantial reason (if any such exist) is hard if not impossible.
I think the pattern that emerges is that the popular languages I listed are multi-discipline, they can be adapted to whichever paradigm you prefer, even if it's a little cumbersome, and over time they adopt the key features of other languages, while retaining their existing benefits. In other words... you can get the best features of Haskell in Python, but you can't get the best features of Python in Haskell?
Haskell's best features relative to more mainstream languages are HM-style type inference, higher-kinded types, and typeclasses - none of which are possible in a language without real static types.
I believe all kinds of Python frameworks too like to incorporate FP into their APIs, and of course, there are list comprehensions.
The fact that Eric Meijer was very active in the Haskell community, I think, pretty much confirms that.
I'm not saying either approach is generally the right one, but I think it's interesting that best practice can be the polar opposite of his "single exit" recommendation
> The guard is a proposition, which must be true before the statement is executed. At the start of that statement's execution, one may assume the guard to be true. Also, if the guard is false, the statement will not be executed. The use of guarded commands makes it easier to prove the program meets the specification. The statement is often another guarded command.
e.g.
function Foo (Value : integer) : boolean;
label return;
begin
if Value < 0 then
begin
Foo := false;
goto return;
end;
Foo := Bar(Value);
return:
end;It is a different way of thinking though and it takes a while before your mind stops reaching for while loops.
There's Scheme. There may be constructs like DO and WHILE (though I don't remember if these are standard or just common extensions of Scheme), but they're often just macros implemented using inner functions and tail calls.
Beautiful place btw, of course there are many buildings that didn't exist in his time, I remember the huge and old Atlas building, construction started in 1958.
You can go anywhere by bicycle from residential areas of Eindhoven to the TU/e, the city centre, supermarkets, train station, Philips Stadium, even neighbooring towns like Nuenen or Oirschot can be reached on roads dedicated to bikes.
About a decade ago, over the course of a few months I skimmed through the archive of his notes[1] and at least perused the majority and dug deeper into the more tantalizing ones. I dare say that it was one of the most educational periods of my professional life.
Also, I find his more acerbic observations amusing, but they’re mostly confined to non-technical notes and easily skipped. And who among us hasn’t shared frustration against this or that unfortunate fad in our field?
[1] https://www.cs.utexas.edu/users/EWD/transcriptions/transcrip...
> The purpose of abstraction is not to be vague, but to create a new semantic level in which one can be absolutely precise.
> The question of whether a machine can think is the same as the question of whether a submarine can swim.
> In this respect the programmer does not differ from any other craftsman: unless he loves his tools it is highly improbable that he will ever create something of superior quality.
The article is very well written and is full of insights into Dijkstra's thought process.
Looking forward to reading more EWDs.
Dijkstra had some real achievements. He was one of the giants, no doubt. But he wasn't the giant. He was highly opinionated. He had his way of working, and anybody who did it any other way was wrong (and probably either stupid or lazy). That's not the optimal route to improving a field, especially when you aren't 100% right.
These are some of my favorites:
From Tony Hoare:
They removed recursion from Algol 60, on the grounds of its alleged inefficiency. Edsger's answer was that recursion was a useful programming tool, and every workman should be allowed to fall in love with their tools.
He would repeatedly stress the importance of choosing words carefully: "If you have to use your hands, then there is something wrong with your words", he would say, adding "imagine there is a blind man in your audience, how would you speak?"
Over time, I came to learn and appreciate that Prof. Dijkstra's class was operating at two different planes. There was the lesson and then, the lesson within the lesson. At one level, the goal was to solve the presented problem. At the second and more richer meta-level was the approach for arriving at the solution.
Prof. Dijkstra taught us that in computing science, complexity comes for free; one has to work hard for simplicity.
In his technical writing, he used language like a precision tool. For him, precision did not necessarily imply ease of reading, and he stated that the reader also had to make an effort to understand.
Professionally, Edsger's impact on me is best summarized by his ATAC Rule 0: "Don't make a mess of it." It made me strive for simplicity in notation and modelling throughout my working life and take unpleasant complexity as an indication of a possible lack of comprehension.
I decided against writing with a fountain pen. He never took issue with this, except in a letter that he posted seven weeks before his death: "I hope you will overcome your resistance and learn how to fill a pen without soiling your fingers, or otherwise you are denying yourself one of the joys of life." These were his last written words to me.
A famous quote of EWD is the following: 'I mean, if 10 years from now, when you are doing something quick and dirty, you suddenly visualize that I am looking over your shoulders and say to yourself "Dijkstra would not have liked this", well, that would be enough immortality for me' (EWD1213). I can testify that this actually works.
When I write papers, even now, I still see Edsger over my shoulder going "Tsk! Tsk!" [...] I found that being less sloppy not only helped my readers understand what I had written, but most importantly, helped me from confusing myself.
I have an invisible Edsger inside my head which looks over my shoulder when I am writing and quietly goes "Tut tut" if I write something that is muddled or not accurate. I don't always listen to that voice but know I should.
At faculty meetings when Edsger was not present it was common for someone to say "If Edsger were here he'd say such-and-such."
He came to work in the summer wearing a big Texas cowboy hat - they are made to keep you cool in the Texas sun. Often he would have on a cowboy's string tie. So from the waist up, he looked more Texan than I did. But he almost always wore shorts and sandals, which ruined the cowboy image completely.
Edsger critiqued not the person but only what they said, and later one could drink a beer and laugh as if nothing happened. Technical differences and shortcomings should be treated this way.
Douglas Engelbart aroused Dijkstra's ire so much that he needed a whole page of vituperative prose to offload his emotions (EWD387). What has Engelbart done to provoke this outburst? One only has to refer to "Engelbart's Law", see Wikipedia. Its reasoning seems to run as follows: look at what mere printing has done as a tool for thought; the system demonstrated is so much more powerful than printing that it must quickly lead to Intellect Augmentation. This way Engelbart showed no appreciation for the rich culture developed over centuries. What makes printing a powerful tool for thought is mostly due to other things than technology. Much of the power of this culture comes from publishers and editors, who sniff out what is worth printing and hold back what is not. Another important component of this culture is provided by libraries and librarians. Much is due to scholarly societies, which started printing their proceedings and to commercial publishers, which created journals, each with their editorial board and unseen bevy of reviewers. Most of all it is due to the idea of a university.> The other two speakers that gave three one-hour lectures, Dr. McKay from IBM, Yorktown Heights and Professor Engelbart, SRI, Menlo Park, were both terrible. McKay spoke undiluted IBMerese for three full hours and I am not going to give any further comments; I only heard the first hour —like many participants— and that was enough (too much). Because I had an urgent letter to write I missed Engelbart's first lecture —it was not really a lecture, he showed a movie— but I attended his next two performances. He was not only terribly bad, he was dangerous as well, not so much on account of the product he was selling —a sophisticated on-line text-editor that could be quite useful— as on account of the way in which he appealed to mankind's lower instincts while selling it. The undisguised appeal to anti-intellectualism and anti-individualism was frightening. He was talking about his "augmented knowledge workshop" and I was constantly reminded of Manny Lehman's vigorous complaint about the American educational system that is extremely "knowledge oriented", failing to do justice to the fact that one of the main objects of education is the insight that makes quite a lot of knowledge superfluous. (Sentences like "the half-life of a fresh university graduate is five years" are only correct if you have crammed the curriculum with volatile knowledge, erroneously presented as stuff worth knowing.) His anti-individualism surfaced when he recommended his gadget as a tool for easing the cooperation between persons and groups possibly miles apart, more or less suggesting that only then you are really "participating": no place for the solitary thinker. (Vide the sound track of the Monsanto movie showing some employees: "No geniuses here: just a bunch of average Americans, working together."!) The two talks I heard were absolutely insipid, he had handed out a paper "An augmented knowledge workshop.": the syntactical ambiguity in the title is characteristic for the level of the rest of the article. As a result of his presentations I have told a few of the participants that I had found, thanks to this seminar, a new software project. "Because in the years to come there will be a crippling shortage of competent programmers, I shall develop a software package, called "The Instruction Interpreter". From the moment of its completion, users do no longer need to program, they just give their instructions to the system." (This is only an edited version of one of the paragraphs of the Engelbart article!) I would have liked to start a discussion with him but I knew that my lack of mastery of the understatement would have made me too rude for English ears if I had spoken. Finally —after a more than two-hour effort in the middle of the night in sorting out his muddle— I decided that he was not worth the trouble. (One of the most offending conclusions I ever came to!)
It may be big because of industrial applications, but as a discipline does it deserve to be bigger than a section in an applied math dept or elec engg dept in universities?
https://news.ycombinator.com/item?id=24942671 - Oct 30, 2020 with 196 comments
A lot of good discussion in there.