Discover and read the best of Twitter Threads about #SwiftLang
Most recents (9)
New update to Ferrite has been released — mostly minor fixes, performance improvements, that sort of thing. It does also include a big update to the User Guide, now featuring a full search facility: 
You can just type some text and it'll search for it throughout the guide — but it'll also suggest chapters or keywords you might be interested in:
The User Guide used to be written in raw HTML, but that got increasingly difficult to manage due to all the variations (iPhone/iPad, iOS 12/13, print version etc). Now, it's written in an unusual way — in #SwiftLang of all things, using some similar techniques to SwiftUI:
Hey, internet: just in case you're not one of those engineers who hears "Hey, compilers!" and comes running delightedly, here's a thread on why I think #MLIR is going to transform the machine learning industry:
(1/n)
for more technical details, ref: ai.google/research/pubs/…
(1/n)
for more technical details, ref: ai.google/research/pubs/…
We're seeing an explosion of domain-specific accelerated hardware (yay!).
@GoogleAI's TPUs are an example of this, sure; but you also have @Apple's specialized chips for iPhones, @BaiduResearch's Kunlun, and literally *thousands* of others.
The way that you program for
(2/n)
@GoogleAI's TPUs are an example of this, sure; but you also have @Apple's specialized chips for iPhones, @BaiduResearch's Kunlun, and literally *thousands* of others.
The way that you program for
(2/n)
@GoogleAI @Apple @BaiduResearch these chips right now isn't easy.
Developer-facing tooling is essentially non-existent (or quite bad, even for TPUs); many models are built, and then found to be impossible to deploy to edge hardware; etc.
You also have very rad, but specialized types (e.g., bfloat16).
(3/n)
Developer-facing tooling is essentially non-existent (or quite bad, even for TPUs); many models are built, and then found to be impossible to deploy to edge hardware; etc.
You also have very rad, but specialized types (e.g., bfloat16).
(3/n)
If you can't make it to the TensorFlow Dev Summit this week, you can still follow along on the live stream. The agenda for the day is at, tensorflow.org/dev-summit/age…. #MachineLearning #TFDevSummit #TensorFlow
…the official #GoogleCoral announcement is now up on the @GoogleDevs blog, developers.googleblog.com/2019/03/introd…. Full details of @Google's new #EdgeTPU hardware in my piece for @HacksterIO. #MachineLearning #Edge #IoT #TFDevSummit blog.hackster.io/say-hello-to-g…
I've been sitting on this for months. Say "Hello" to the @SparkFun Edge, sparkfun.com/news/2886. More soon. #TensorFlow #IoT #Edge #TFDevSummit
Since there seems to be another squall of #swiftlang hate passing overhead (e.g. ) I just want to take a moment to thank the Swift devs & say how much I like the language. Contrary to the linked article, it’s made my business _more_ tenable, not less.
Are there changes I’d like to see? Of course! No language is ever perfect — but that’s why there’s a community process for enhancements.
Plus, outside of that process, I’ve found it much easier than in Obj-C to roll my own enhancements via library extensions, yet still migrate to official ones if/when they become available (e.g. Codable, CaseIterable, DispatchQueue API improvements just to pick 3).
Playing today with msgpack objc impl at #swiftlang: optionals = you're fucked, every .min/.max = you're fucked
Tomorrow: Test a message with some null on it, 😰
Found wtf is happening to optionals here:
When proposing changes to #SwiftLang every time you think “It’s okay, we’ll write a migrator to handle the changes automatically”, remember
it’s not actually true for production code, because the underlying compiler isn't reliable enough in the face of type errors. SEG11 all over
The migrator relies on the compiler, the compiler relies on valid types, migration causes invalid types, migration is broken.
iOS 8 went to beta 7.
iOS 9 went to b6.
iOS 10b6 came out, pretty late in the beta cycle to be landing huge semantic changes to #SwiftLang 3
iOS 9 went to b6.
iOS 10b6 came out, pretty late in the beta cycle to be landing huge semantic changes to #SwiftLang 3
+ the migrator warns you off re-running it to migrate from previous Swift 3 betas, so you’re pretty much on your own. 🎉
Also, some breaking bugs in the clang Obj-C → Swift translation layer (have fun implementing a UIActivity) aren’t helping.
Wondering if #SwiftLang 3 migration was really handled the right way. I get there are advantages to "ripping off band-aid" in one go, but…
I suspect combining certain changes increases migration difficulty exponentially, not linearly. Multi small migrations might've been better?
feel that reliably machine-migratable 1:1 changes like renaming should be separate from complex semantic shifts like Seq. index refactoring
I must admit, I find the #SwiftLang dynamism "debate" perplexing. Did I miss a deprecation notice? Are ObjC RT, IB, KVC/KVO not in Xcode 8?
I'd understand if this was people complaining about Swift-on-Linux, but it's not, it's about iOS and Mac app development.
Apple has built an empire on these technologies, yet people seem concerned they'll just throw them under the bus? I don’t get it.
