Read on Twitter
Heck with it, feel like killing an evening. Give me any tech topic and I'll give you a connected essay pitch that 1) has a solid shot of going viral and 2) people will actually find insightful and worth reading
"This code broke a compiler" Find a case where the code segfaults *during compilation*, talk about why and what the fix was, and then how compilers are tested to avoid these weird things in the future https://twitter.com/tekknolagi/status/1316908606516846593
"Write code to write your code": using template strings and programs to fill out a program that'd otherwise be super tedious. Maybe config files?
(I do this a lot with python. Esp helps for DSLs and some spec languages. https://docs.python.org/3/library/string.html#template-strings) https://twitter.com/palvaro/status/1316908965138206720
(I do this a lot with python. Esp helps for DSLs and some spec languages. https://docs.python.org/3/library/string.html#template-strings) https://twitter.com/palvaro/status/1316908965138206720
"21st century logic programming": how a lot of the early problems with prolog come from it being developed 50 years ago, and how with much faster computers we can do more with backtracking to make better logic programming affordances https://twitter.com/enkiv2/status/1316909957502164992
"Why do Makefiles require tabs?" A historical perspective on the development of make and makefiles https://twitter.com/ihnorton/status/1316911401613615105
"Software teaching wetware": use applying cogsci to PL design to give techies a brief intro to cogsci as a field, so they can go "ah, I'm annoyed by feature X because of principle XYZ!" https://twitter.com/ijks_w/status/1316911780627681281
I've been informed that Automatic Programming is actually something different. New topic: we all know it's doomed to fail, but how is it doomed to fail? Let's compare AP coming from academia to AP coming from snake-oil sellers and see how they differ https://twitter.com/palvaro/status/1316912872065302531
"Ruby vs Python": talk about how you'll compare the two, write sample programs in each, then veer hard left into what what CPython and YARV are doing under the hood and how the different things they do reflect different interests in the communities https://twitter.com/HeyChelseaTroy/status/1316908439835185160
"When NoSql makes sense", but as all of the kinds of data which are still relational but more specialized. So specialized graph, object, hierarchical, taxonomic, document (as in XML documents) schema'd databases. https://twitter.com/spion/status/1316913751115894789
"Everyone does UI development wrong", I dunno, probably something about UX labs and one-way mirrors, I don't think anybody in small-medium tech companies actually does cold runs where they watch an actual user use the UI https://twitter.com/asolove/status/1316916956189044737
I mean isn't this actually an important thing? A lot of legacy scientific computing code is TMI in Fortran, so if you want to integrate ML with any of that you'll have to have 2-way communication https://twitter.com/DJGagneDos/status/1316918017893257218
"Elliptic Curves are Hard": make a small change to the safe set of parameters and show it opens up some kind of attack. Bonus: explain what the heck a test vector is https://twitter.com/bmastenbrook/status/1316922324969455617
Seems like a good essay topic already, tbh https://twitter.com/arachnocapital2/status/1316918828538974208
"Interesting things people have done in the Fediverse that are not Mastadon"
bonus: "Interesting things people have done in the Fediverse that are not just alternatives of centralized services" https://twitter.com/jgoldschrafe/status/1316919634399035398
bonus: "Interesting things people have done in the Fediverse that are not just alternatives of centralized services" https://twitter.com/jgoldschrafe/status/1316919634399035398
"How much data would your roll of scotch tape store, if it was magnetic tape instead". Bonus: how you actually read, store, and write data on magnetic tape https://twitter.com/garrett_wollman/status/1316920657918263296
"The Rise and Fall of UML": a sprawling epic that covers the method wars, CASE tooling, Rational's ascendance and the Three Amigos, the first UML conferences, the IBM acquisition, and UML's gradual decline into obscurity https://twitter.com/kylemisc/status/1316924918932656128
Why ORMs are hard in theory, but going the other way, showing why it's so hard to represent object data in a relational schema. What do object database O"R"Ms look like? https://twitter.com/sklrmths/status/1316920545775132674
You don't know how tempting it is to just phone these all in with "the history of $thingymajig"
Just implement a prototype and the clickbait writes itself https://twitter.com/rob_rix/status/1316921884949217283
"The Hardest Problem Naming Things", but not from the done-to-death perspective of finding the right name for something among many options, but the deeper challenge of giving something a name that we don't as a culture have a name for yet https://twitter.com/blagh/status/1316921885901324290
"Tech Credit": if tech debt exists then so must its opposite, code that continually pays your project dividends. Or maybe debt that's owed to you, somehow. Hm. https://twitter.com/fabioperalves/status/1316922419081302018
"Why jargon is good, actually." Jargon as a form of roadbuilding vs a form of gatekeeping https://twitter.com/hyperpape/status/1316922592956194817
"Lesser-known optimization tradeoffs", things like power consumption, heat output, stability, anything that's not memory or simplicity. Find a case where performance somehow comes at the cost of worse user interfaces https://twitter.com/shelbyspees/status/1316916486003449856
Discuss a historical case where one became the other, and how it happened https://twitter.com/vkolgi/status/1316929825853067265
Play this straight, and then share how their foreparents talked about *them* https://twitter.com/mengwong/status/1316928470899970049
Use this as a motivating example to introduce constraint satisfaction problems https://twitter.com/bmastenbrook/status/1316923199951654916
"How HIPAA affects your codebase" https://twitter.com/LewisEmilyR/status/1316934651420811264
Oh geez, I know nothing about Kubernetes. But it's a distributed systems thing so a formal specification of part of it would go over well with the HN crowd https://twitter.com/rickasaurus/status/1316938133624283137
"Beyond text": A theoretical explanation of why canonical text is problematic, and then a *persuasive* demo of a *productive* technique opened by an alternative. Frame as speculative theorycrafting to avoid "but legacy!!!" arguments https://twitter.com/lhochstein/status/1316938891971203073
How holograms work and the technical challenges of consumerizing them. Easy. https://twitter.com/flowinthezone/status/1316935758524092416
Ugh don't know anything at all here. The *physical* challenges of making the chips, maybe? Instruction sets and quantum tunneling? Hypothetical chipsets for photonic computing? Drawing a blank here https://twitter.com/Paul_Bone/status/1316933315350376448
No interesting twist on this needed, just "how to keep up with a subfield" https://twitter.com/nuttycom/status/1316936614241837057
"PHP: A Fractal of Bad Design" https://twitter.com/Alaskan_Emily/status/1316935971485700098
"Bullshitting bullshit interview questions": taking common bullshit interview questions and breaking them over your knee with fringetech: APLs, TLA+, constraint solvers, prolog. Anything that turns a hard problem into a one-liner https://twitter.com/wilton_quinn/status/1316964711179685894
"The science case for code review": Crunch all the research, see what you get https://twitter.com/dotnwat/status/1316966200631611393
ok going to bed, thanks for giving me an excuse to not do actual things, night all
