Quantcast
Sign in with Twitter

Username:

Occasional OCaml programmer

78 Following   1,697 Followers   902 Tweets

Join Twitter 6/21/09

Reverse Tweets
@ndm_haskell @TacticalGrace To program well, you also need human feedback. To be good, code must be understandable by others.
7/27
2014
Can't believe @avsm tricked me into *wanting* to learn OCaml! My spare time was *already* negative, but I'm staring at "Real World OCaml".
Retweeted by Yaron Minsky
7/24
2014
Very cool work on unikernels and cloud security with MirageOS and @XenProject - Go see @avsm talk tomorrow at 11 a.m. F150 #OSCON2014
Retweeted by Yaron Minsky@diptanu @avsm Hmm. I'm not entirely sure that's a compliment. Does that mean OCaml is the new Verilog?
7/23
2014
There's a real OCaml book out, finally. (And one of the authors is a Jane Streeter, natch.) http://t.co/CKU8u3z6Nk
Retweeted by Yaron MinskyLook who is here signing Real World OCaml... #oscon http://t.co/BqfK34BtjQ
Retweeted by Yaron Minsky
7/22
2014
Income Inequality Is Not Rising Globally. It's Falling (as w/ violence, dataphobic journalism has misled people) http://t.co/9Ny5B9Oi0W
Retweeted by Yaron MinskyForget resources- the real causes of war are ideas, particularly irredentist nationalism. http://t.co/4k7kCS5qHg
Retweeted by Yaron Minsky
7/20
2014
Got my copy of Real World OCaml and a pound of caffeinated mints in the mail see you next week
Retweeted by Yaron MinskyContract law should be turned into a formal language, with static analyzers and solvers. More Haskell, less Perl.
Retweeted by Yaron Minsky
7/16
2014
@marcaruel Sure, but do you re-review post-rebase? Read the interdiff? Or just hope that the resolution was done correctly?
7/15
2014
@marcaruel How do you handle merge conflicts that happen between initial review and final linearization?
7/14
2014
"Read Real World OCaml... It might be the best computer language book I’ve read in my life." Quite a compliment! http://t.co/VhP0hEa7NcI tried to learn OCaml like 9 times before it stuck. So I wrote a Beginner's guide to OCaml Beginner's Guides. http://t.co/v5Bgp5GB7I
Retweeted by Yaron MinskyA sprawling, funny, erudite post from @OlinShivers on the nature of names: https://t.co/Eara9aQfotHunting a packet of death on @openmirage unikernels: fuzzing, confusion, data leakage, and triumph at http://t.co/igiGernsnE
Retweeted by Yaron Minsky
7/10
2014
the only "static vs. dynamic" argument i'm willing to take seriously (between matthias and bob) apparently happened: http://t.co/8onz114SoY
Retweeted by Yaron Minsky
7/8
2014
@yminsky 700 packages in OPAM now! Just need 300 more by September for the feel-good factor...
Retweeted by Yaron MinskyThoughts on automating your release prices to ensure that you never break the build, and the ensuing complications. https://t.co/JY0lUfvWEs
7/7
2014
All told, I think @OneAcreFund is an excellent organization to support if you want to buy the maximum good for your charity dollar.From their research, @OneAcreFund doubles the income of the farmers they work with. Having seen their operation, I believe the numbers.A great NGO: @oneacrefund helps poor farmers by selling training and materials on credit. Clients will only pay if they see the value.
7/4
2014
Don’t put so much faith in the ability of any programming language to solve all your software engineering problems.
Retweeted by Yaron Minsky
7/2
2014
Most large code cleanups involve losing some small bits of functionality, because those are the tendrils keeping systems intertwined.
Retweeted by Yaron Minsky
6/28
2014
Some thoughts on going back and rereading Lamport's classic clocks paper. https://t.co/aI49p7Z30J
6/26
2014
Just had a lovely dinner with Neal Stephenson after his talk at Jane Street. He is just as charming and interesting as you might imagine.My Iron post was clearly judged dull by reddit. Maybe most folk are (sanely) less obsessed than I about code review. http://t.co/Nvh3fEG6gG@eikke We actually embed review comments directly in the code. First class out of band support for such comments would be good.@LukeHoersten Short version: review the diff-of-diffs. Our patdiff4 (to be released soon) is a tool to help. https://t.co/EKG1h9m4aUI wonder if code review should be a first-class part of version control. Seems like part of merging should be understanding the result.@eriangazag It's tricky. It's useful that the underlying hg graph is a dag. But the feature structure is just a tree.
6/25
2014
A post on hierarchical features in Iron, our new review tool. This is my favorite part of the design of the system. https://t.co/HCx3Xk8WtJ@yminsky in particular, he motivates causal clocks with a casually fair mutual exclusion algorithm, whose value is unclear.Just did Lamport's clocks paper in our systems seminar. Surprised to find that the motivation given for causal clocks is subtly circular.
6/24
2014
One week before the CFP deadline! http://t.co/HGRqnvcyUC
Retweeted by Yaron MinskyThe "Call Me Maybe" series by @aphyr should be required reading for distributed systems folks. Great stuff!
Retweeted by Yaron Minsky
6/20
2014
Working hard on an Iron post covering hierarchical features and the release workflows they enable. It's subtle stuff...
6/19
2014
@sperbsen @avsm very close. We built this bit before subrepos existed.
6/17
2014
@marius Sigh. I did say "mostly accurate"...@cicatriz Our new code review tool: https://t.co/QIdQSB356tInteresting to see a (mostly correct) transcript of my qcon talk on PCWorld, of all places. Oddly clickbaitish title. http://t.co/YP99xMDjsa@avsm we'll probably add support for connecting features when a feature in one repo requires a feature in another.@avsm I don't know that this would scale to a large number of dependencies. And iron is mostly unaware of this structure for now.@avsm We deal with our multi repo projects by having a file in a repo that directories the revision of each dependency.@yminsky It's a surprisingly general purpose mechanism, enabling both dependent features and independent management of sub-projects.My favorite idea in Iron is hierarchical features: this supports a PR against a PR in a way that preserves review.
6/16
2014
According to InfoWorld, Swift got generics from Java. Lovely how the technology world eats its ancestors. http://t.co/3nprii0YU2
6/15
2014
@a_cowley @craigstuntz That's a really interesting point, and not one I'd thought of.@chrisamaphone Agreed.@yminsky @chrisamaphone If you want to ensure that your hearers are properly put in their place, say “it’s *just* a monoid” …
Retweeted by Yaron Minsky@cobbpg Right. If you understand that this is PL jargon (not even math jargon really), and use it accordingly, then it's fine.@chrisamaphone or say, "you need an associative operator". The identity isn't the interesting bit.@chrisamaphone in other words, act as if almost no one knows the word monoid. Because that's the truth.@chrisamaphone say: "you need an associative operator and an identity element" rather than "you need a monoid".@cobbpg I'm not opposed to using the term, but when explaining, you should appeal to well-worn terminology, which monoid is not.@missingfaktor you need a full sentence that includes the word associative. Takes a bit longer to say, but at least you'll be understood...I also tend to think the term monoid is more trouble than it's worth. "associative" is better known by children and mathematicians alike.
6/14
2014
More on Iron, our new code review tool. Scrutinize your code in style: https://t.co/6LBLPFMn4C@jacobmcarthur Hah! I was not referring to you nor anyone at Jane Street.Nothing more painful than watching an FP hipster explain monoids as of they're deep and mysterious. Simple ideas should be explained simply!@mattmight I think that counts as a bang...@LukeHoersten In OCaml, and hg only for now, but an interested contributor should be able to add git support without too much trouble.
6/13
2014
@khrabrov Right now it's Mercurial only. But porting to git should not be hard, and the OCaml Labs folks are potentially interested.A quick post about our new code review tool. https://t.co/EKG1h9m4aU
6/12
2014
Is there Algebraic Data Type Notation? If so, can we all start using it instead of JSON post haste?
Retweeted by Yaron Minsky
6/10
2014
This Friday, I'll be talking about the past, present and future of FP and OCaml at Jane Street. https://t.co/6S4FCE1CoL
6/9
2014
@hmansell @TacticalGrace @copumpkin Ugh.@mwotton @TacticalGrace @copumpkin sorry, I meant / not +.@mwotton @TacticalGrace @copumpkin Yes, if your language supports non total functions, like hd or +.@TacticalGrace @copumpkin Anyone know what recovery mechanism Swift offers, if any?@TacticalGrace @copumpkin Sure, though monitors bear some resemblance to exception handlers. Anyway, some recovery mechanism is needed.@copumpkin @TacticalGrace unless your language guarantees totality, of course...@copumpkin @TacticalGrace Aren't exception handlers at some level a necessity if you're going to recover from some subcomponent failing?
6/7
2014
Facebook: PHP + OCaml = Hack Apple: Objective C + OCaml = Swift Microsoft: C# + OCaml = F#
Retweeted by Yaron Minsky
6/4
2014
@dom_dere I very much doubt I'll like Swift as much as OCaml, but that doesn't prevent me from being happy about the progress it represents.@dom_dere Don't get me wrong, recursive ADTs are valuable. Don't know why they omitted them, but there may be good reasons.@dom_dere The ability to exhaustively enumerate finite cases is a very common case. Recursive structures are a rarer form of case analysis.@dom_dere It's not great; but in my experience most of the value is in the non-recursive case anyway.I worry though, that Apple's instincts will lead it to keep Swift closed-source. That will surely limit the language's success.@lojikil Do you mean Algol?Swift's design is so much less depressing than Go's. Confirmation that algebraic data types are not too hard for real world programmers.It's interesting to compare Swift and Go. One has caught up to the 1990s in language design, with the other firmly in the 60s.
Retweeted by Yaron MinskyAh, turns out I'm wrong: the HTML is here. That said, Apple wasn't going out of its way to tell folk about it. https://t.co/RYli5Q505bJoy: Apple's Swift has variants and pattern matching. Sadness: their documentation is only available in the proprietary ibook format.“4000 new developer APIs” is exactly the opposite of what you want.
Retweeted by Yaron Minsky
6/2
2014
“Buy an index fund. This is the most actionable, most mathematically supported, short-form investment advice ever.” http://t.co/E8DULOSy3p
Retweeted by Yaron Minsky
5/30
2014
@ezyang you might find this poster helpful: https://t.co/hxVcXnYSPq
5/29
2014
here I am begging the CPU people to support integer overflow traps: http://t.co/UDJw6amQRa
Retweeted by Yaron Minsky
5/28
2014
@jeanqasaur Traders usually have 6 screens, believe it or not. Thanks for coming to visit! It was a really fun talk.@sperbsen @conal Though I do think "why scheme matters" should discuss macros prominently.@sperbsen @conal I know, but I still disagree. I think we should discuss over beers at ICFP! Too hard to resolve in 140 char chunks.@sperbsen @conal Ah, and there we part company in a way that mere tweets will likely not be able to resolve...@a_cowley @jonsterling Proofs do often have bugs. But important theorems have too many interesting implications...@sperbsen @conal But ML and Haskell are more similar than different when it comes to productivity.@sperbsen @conal it's not just the title. He asks, why are Miranda like languages so effective, and then spends a lot of time on laziness.@a_cowley @jonsterling still: in 20th century, very rare for important theorem to be widely thought true and then found false.
5/23
2014
@conal @EyalL Sounds perfect! I'd like that title better, although it would sell less well on the corner newsstand.@conal @headinthebox Got it! I guess that was Erik. Erik, can you explain what you mean?@bitemyapp @conal Also Lennart Augustsson. And lots of other Haskell hackers, current and former.@bitemyapp @conal Be serious. I've heard it from SPJ directly.
5/22
2014
« Prev1234567

0