Sign in with Twitter


James Iry @jamesiry San Francisco, CA

Some people when faced with a problem think, I know, I'll use distributed computing. Now they have n^2 problems. I corral Cascading cats at Concurrent.

214 Following   3,375 Followers   7,473 Tweets

Join Twitter 1/15/09

Reverse Tweets
Looks like Horton and Twitter are teaming up to give a webinar on @Scalding and @Cascading http://t.co/gyn9wDXHYc
Retweeted by James Iry@cowtowncoder @capotej @bodil Okay, maybe I misunderstood what was meant by "sharing code client/server"@cowtowncoder @capotej @bodil Example. Game computes collisions on clients for speed. Re-computes on server for validation and consistency.@capotej @bodil Which means a certain amount of code sharing (or at least strong similarity) is inevitable.@capotej @bodil In many client/server situations you can't trust the client for security reason but you want computation there for speed…Trial slogan: "we pass the buck to you so you don't have to."
I'd like to raise a toast to the French http://t.co/u4cWTXzsii
@deech Run away. It's a trap!A game that claims to be beyond boxes, yet comes in a box. Not only post-box but post-structuralist. http://t.co/Abml3v1YJJ
Retweeted by James Iry
@bodil Needs sauce.@bodil Oh, it gets pretty hat there sometimes. http://t.co/vz2epaVO6b@bodil That's because it is in fact "hat sauce," a sauce designed to make eating your hat palatable by making it spicy.@jamesiry not sure about the capitalization though, what about "Bike shed."?
Retweeted by James Iry@cwensel Yo, hook a brother up.From now on every new project will be code-named "bike shed."@extempore2 @dbgrandi @aloiscochard No, no, paradigms are organized religions (with churches, martyrs, and bishops) not mere superstitions.Gnu calls weak vtables and some similar things "vague linkage." You couldn't invent this much accidental complexity if you tried.On my BART commute this morning I learned about C++ "weak vtables." They're the perfect obscure collision between several C++ peculiarities@shebang @deech I do want to say "not these types" when dealing with effects system.@cwensel Did you try port forwarding? I mean on the vacuum cleaner.could not even sort out how to work our office vacuum cleaner, getting this tez cluster running is probably a lost cause
Retweeted by James IryWent to hear Bach last night. Very disappointed. Was just some cover band.
@djspiewak Even then, a if you hack a PR based git bisect I'm still going to be annoyed if I can't bisect into a PR with lots of commits.@djspiewak If git bisect worked at the level of PRs I might accept that. But since it doesn't I have minimum standards on commits.@djspiewak I should say it used to have an anything goes policy. That changed and bisecting through history has a clear delineation.@djspiewak to long term health of the project.@djspiewak And I've lived with a large code base that had anything goes commits. Sometimes re-doing a PR before committing is essential…@djspiewak Which is my original point. Gitbhub doesn't maintain enough history on the PR that gets a push -f.@djspiewak people who based work off the WIP are screwed.@djspiewak If your commits don't meet our standards we have two choices. Close PR and ask for reissue. Or ask for push-f. Either way…@djspiewak So don't push -f to a collaborative branch.@djspiewak Until my PR is merged, my branch is WIP.Most computer science papers' evaluation sections look like this: http://t.co/el40woPlrq
Retweeted by James Iry@davetron5000 @github But i don't necessarily want to squash into one commit, just into a better organized series of clean commits.@davetron5000 @github …because I need to make sure the squishing process didn't introduce new, unreviewed changes.@davetron5000 As for one final squish, I'd be happy with that if @github let me compare the result of the squished PR with the previous PRs@davetron5000 @github Not "true" history since arbitrary rebasing happened before the PR was issued.@davetron5000 @github Don't rebase a PR? But then you've got millions of janky little commits to fix typos the reviewers found.Any suggestions on how to use @github to review pull requests that get rebased during the review process? The lack of history is frustratingOMG, homotopy type theory is so HoTT.FAQ: Yes, that means I'm not doing any Clojure. Lisps are too mainstream. Homotopy type theory is cool now.
Retweeted by James Iry
@rit @runarorama Yes, the version composed by the skalds of old! Or was that scalds?@bodil @rit @milessabin I thought that was Palpatine Reforms.@veronica @brendaneich Given their relative recent success, I would expect it to be the other way around.@runarorama Pfft. Flip phones are so last century.Say "cloud" again! Say! "cloud" again! I dare you! I double-dare you, motherfucker! Say "cloud" one more goddamn time! — Me to @cwensel@rit @milessabin But as heavy cavalry rose the pike formation came back into vogue and was a major part of the field until guns took over.@rit @milessabin Swords get all the glory, but pikes got the work done.If I understand my twitter feed correctly, Scotland will soon vote on independence from Apple.A visual proof that neural nets can compute any function http://t.co/QpBI8yhJay
Retweeted by James IryI was thinking about the phases of the moon and then I realized…I'm only happy when it wanes.If anyone on this airplane tries to judge me for watching vapid TV shows about rich teenagers, I'll just be like, "I got a Ph.D. yesterday."
Retweeted by James IryDon't kid yourself: "deprecated" doesn't mean you'll ever remove it. It just means, "I has a sad about this code!"
Retweeted by James Iryapple event starting, hoping they announce shifting from swift to idris
Retweeted by James Iry
Near the New England Conservatory, student approaches me "I like your shirt, F# is an under appreciated note"
Retweeted by James Iry@rickasaurus That a nicer motto than "your current self is the the result of dozens of botched merges."In which @agl__'s definition of "shallow" differs from most everyone's. https://t.co/0NyZl3JC6k On proving things about C.
Retweeted by James IryThat google brings this up when you image search "Best Gaming Mouse" is one of life's simple joys http://t.co/n7s9GSkb5v
Retweeted by James IryLegit talk outline... #lucacardellifest http://t.co/yMgAzCtOvL
Retweeted by James Iry
Forks everywhere! #scala #whatwg #w3c #markdown http://t.co/GilffgIiGV
Retweeted by James IryI'm in interstellar space & you make me feel like a star. Thanks for the #launchiversary wishes! I left Earth 37 years ago today.
Retweeted by James Iry
@runarorama Also composition. Asteroid = basically rock and/or metal. Comet = significant amount of ice (H20, C02, CO, CH4, etc)@campster it looks like @AuerbachKeller took "gamer" to mean anybody who plays games rather than a segment that sees gaming as an identity... turns out Idris doesn't even support proof by contradiction. Future work, maybe?
Retweeted by James Iry@shinypb @peterseibel They seem very perky.
Retweeted by James IryOkay, Canada, it’s time for the hard truth about Tim Hortons http://t.co/2lPfi72pPW
Retweeted by James Iry@chrisamaphone That or a pain in the ass.
@curious_reader So far everybody's talking nicely about merge and binary compatibility. We'll see.@deech "Rain: it never leaks but it pours."Ugh. The .doc format resume can die in the hellfire from which it spawned.
@deech @raganwald Isn't the common wisdom "don't knock it till you've tried it"
"in the end, there are only two kinds of development - incremental, and excremental." - @Gilad_Bracha@myriachan @johnregehr What are you fighting about? Have a link?@MurderMondays @BadAstronomer That is NOT support.@MurderMondays @BadAstronomer 'If [you think games] are “under attack” from [...] “social justice warriors”: please leave this website.'@djspiewak Fundamentally, if you are going to have both an effects system and reusability then you'll need effect polymorphism at least.@djspiewak Type safe is fine, but Java doesn't give you the power to use that type safety well. http://t.co/uGTcDUv3bt@djspiewak Except that whole checked exception thing. That was a bit of a clunker.Scala, java, and core algorithmic stuff? Come work with me. http://t.co/DjGKzmNzxTDoes StackOverflow count as "publish" in "publish or perish?" If so, @pigworker is nigh immortal at this point.I wish that I had been taught in university that the proper way to design an algorithm is to first understand the properties of the problem
Retweeted by James Iry@jeremyhorwitz You aren't going to move to a remote cabin, write manifestos, and mail bombs to people are you? @dsquirrely @hetrestaurant@raganwald Your boolean is my function that takes two continuations. - Alonzo Church@laurencetratt @samth @johnregehr Yes. Implementation defined means it has to tell you what it's gonna do. UB means chaos.
wondering if its reasonable to glue tacks to our book shelves to discourage baby from scaling them
Retweeted by James Iry"Experiments in Code Typography" http://t.co/zHhfVzPhZW
Retweeted by James Iry@pchiusano Now let me tell you about all the times I've seen InterruptedException mishandled so that interrupts don't work.I'd be a lot happier person if OpenJDK moved to git.
Retweeted by James Iry
@johnregehr Also I half seriously believe that any C program that doesn't deliberately have some UB is probably in the wrong language.@johnregehr I'm curious how much exploiting UB helps real programs vs how much it helps in benchmarks.proposal for a friendly dialect of C: http://t.co/WKRF8WZbiA
Retweeted by James IryThe latest Programming Scala, 2nd Ed., Early Release contains the drafts for all the chapters. http://t.co/elT1h5mqeX
Retweeted by James Iry@mubay Avery Ting Iry@mubay Avery T'ing Gwan-Bi Iry.Hey twitter verse, any feedback on the quality of http://t.co/grImw0T06U ?@marius @jorgeo @jakedonham S&W is fine as long as you ignore everything it says about grammar http://t.co/y0P0HH8TTW@peterseibel @marius @jakedonham "Strunk & White is largely BS anyway" is something only a true pedant can say or appreciate. Thumbs up.Classic First Lines of Novels in Emojis http://t.co/7QDc8OfRHG
Retweeted by James IryIs Coffeescript Inaction qualitatively different from Java Inaction?
« Prev1234567