4 ms·
"Pen Science" is to Writing as CS is to Web 2.0
Do you agree with this analogy?
- cperciva 19y agoNo. A better analogy would be that English literature is to blog posts as Computer Science is to Web 2.0: Blog posts and Web 2.0 are very marginally influenced by -- and are studied on the margins of -- English literature and Computer Science respectively, but they're both primarily sociological phenomena. EDIT: Looking back at the topic, I'm not sure if the submitter is aware of Dijkstra's comment about the nature of Computer Science: "Computer Science is no more about computers than astronomy is about telescopes". Comparing Computer Science to "Pen Science" (which is presumably the study of pens) simply demonstrates a lack of understanding of what Computer Science is.
- amichail 19y agoI'm no fan of Dijkstra, at least when it comes to his views of what CS should be all about. He wanted to turn a rather boring field into something that is much worse. The point of the analogy is that implementation -- essentially what CS is all about -- is not interesting just as the way you construct a pen or use it to write grammatical sentences is not interesting.
- cperciva 19y ago"The point of the analogy is that the implementation -- essentially what CS is all about..." Computer Science is not about how to build a computer. Computer Science is about what you can do with a computer. Google's pagerank is computer science. Recommendation systems, like Amazon and Netflix use, are computer science. Computer science is about discovering new ways to design bridges -- not about building them.
- amichail 19y agoComputer Science is not about how to build a computer. Implementation here refers to software as well. Computer Science is about what you can do with a computer. Not really. The vast majority of CS research papers are not about new uses of a computer.
- extantproject 19y agoComputer Science is about understanding what computing is; it's about what it means to compute. Let me guess: you're not a big fan of Knuth either?
- amichail 19y agoKnuth has done lots of useful work. He represents some of the best that CS has to offer in terms of addressing implementation issues. But clearly, we need to go beyond this and think about apps too.
- ochiba 19y agoWhereas the original "pen science is to writing as CS is to web 2.0" analogy might be awfully skewed, you do have to ask yourself how frequently it is necessary to apply the kind of algorithms and techniques [that CS is in part comprised of] in a typical Web 2.0 application.
- amichail 19y agoWell maybe some advanced algorithms are important (e.g., encryption), but they are part of libraries. You do not need to worry about how they work just as you do not need to worry about the physics behind pens or the way a word processor is constructed when writing a novel.
- extantproject 19y agoKnowing how algorithms work can easily be the difference between a program finishing in ten minutes as opposed to never.
- cperciva 19y agoI also ask myself how frequently it is necessary to apply the plot and characterization -- or for that matter, grammar and spelling, apparently -- to blog posts. Gulliver's Travels is better writing than most blog posts, just like the FreeBSD kernel is better code than Reddit. If most Web 2.0 applications don't apply many concepts from Computer Science, it says more to me about the quality of Web 2.0 applications than it does about the importance of Computer Science.
- amichail 19y agoWeb 2.0 apps should of course make use of many concepts from Computer Science -- just as professional writers should make use of the latest word processing software. But you can still have the case that the web 2.0 apps are much more interesting than the CS concepts on which they are based.
- cperciva 19y ago
- pius 19y agoIt's a pretty terrible analogy, to be honest with you.
- Zak 19y agoLinguistics is to writing as CS is to computing. Web 2.0 is a subset of computing.
- henning 19y agoNo. The basic work on relational databases was done by computer science researchers. And actually people in labs built the Internet, come to think of it. And they pioneered the ideas behind dynamic languages like Ruby. Before there was Matz or Larry there was Alan Kay and Guy Steele. What would we do without computer science researchers?
- neilk 19y agoResponse #1: By which you mean, one makes the other possible? In the middle ages, can you imagine the trouble it was to make your own ink, collect and maintain quills of the appropriate type, train oneself to wield the quill skillfully and not cause inkblots, and finally deal with the fatigue of manipulating such a large object? Nowadays pens are so cheap and easy to use you never even have to think about it. So you move onto other questions, like what to write. Response #2: This statement is idiotic. Computer science is essential to the collection of technologies people call "Web 2.0". The most advanced Web 2.0 sites allow for quick search across enormous databases, and efficient, many-to-many messaging. These problems are hard. Using the appropriate algorithm is essential to scalability. Look at Facebook; they deploy Stanford whiz kids to invent their own infrastructure (hmmm, that sounds familiar) and they're growing faster than anybody else in this space.
- amichail 19y agoIt's true that scalability is hard once you get users. But by then you will have VC funding to hire experts who could do that stuff for you. So it's not essential knowledge for entrepreneurs.
- neilk 19y agoI checked out your blog. Someone who blogs about Scala doesn't need schooling from me on CS. ;) I think I see what you're doing now. Ok, to tell you the truth -- this is something I've wondered about recently myself. I happen to know about the codebases of several successful Web 2.0 sites post-acquisition. And in many cases, they used the naive solution and got away with it until acquisition. It's difficult to blog about these things in public, but I've circulated similar speculations privately, that maybe all CS knowledge is a huge waste of time for entrepreneurs. Using naive solutions has definitely worked for some startups. But it comes at a fairly heavy price. You are essentially betting everything on being acquired before you get so slow that your own users start hating you. Once acquired, you then get a reprieve from death, and then you spend about one year building your 2.0 version that actually scales decently (all the while fighting fires with the clunky 1.0 version). Meanwhile your users scratch their heads wondering why, once you got all that money, you stopped building any new features. This pattern has been repeated many times for successful web 2.0 startups. Contrast that situation with a company like Google or Facebook. One of Google's main advantages was scaling, from their algorithms to their ability to create vast computing infrastructures. Facebook acquired this sort of talent a little bit later in their history, but they are definitely there today. So when they got offers to buy their company for millions of dollars -- even a billion dollars -- they were able to say no. Because they knew they could grow faster by themselves. Last year Zuckerberg was considered an idiot for turning down a billion dollars. But he knew that F8 was in the pipeline, and he was very smart to say no. Being able to scale means freedom. Ignoring scalability is a high-stakes bet. But either strategy might work, depending on what the startup's business model is, and the talents and inclinations of the founders.