POPULARITY
Bringing a little tidy magic to creating flowcharts in R, how data.table is addressing recent shifts in R's C APIs, and another showcase of R's visualization prowess in the realm of brain imaging. Episode LinksThis week's curator: Ryo Nakagawara - @R_by_Ryo@mstdn.social (Mastodon) & @rbyryo.bsky.social (Bluesky) & @R_by_Ryo) (X/Twitter)Flowcharts made easy with the package {flowchart}Use of non-API entry points in data.tableIntro to working with volume and surface brain data Entire issue available at rweekly.org/2025-W04Supplement Resourcesflowchart - R package for drawing participant flow diagrams directly from a dataframe using tidyverse https://bruigtp.github.io/flowchartMermaid - Generation of diagrams like flowcharts or sequence diagrams from text in a similar manner as markdown https://github.com/mermaid-js/mermaidDiagrammeR - Graph and network visualization using tabular data in R https://rich-iannone.github.io/DiagrammeR/RNifti https://github.com/jonclayden/RNiftigifti https://github.com/muschellij2/giftiCRAN Cookbook https://r-consortium.org/posts/user-friendly-technical-cookbook-style-cran-guide-for-new-r-programmers-ready/2024 Posit Year in Review https://posit.co/blog/2024-posit-year-in-review/Supporting the showUse the contact page at https://serve.podhome.fm/custompage/r-weekly-highlights/contact to send us your feedbackR-Weekly Highlights on the Podcastindex.org - You can send a boost into the show directly in the Podcast Index. First, top-up with Alby, and then head over to the R-Weekly Highlights podcast entry on the index.A new way to think about value: https://value4value.infoGet in touch with us on social mediaEric Nantz: @rpodcast@podcastindex.social (Mastodon), @rpodcast.bsky.social (BlueSky) and @theRcast (X/Twitter)Mike Thomas: @mike_thomas@fosstodon.org (Mastodon), @mike-thomas.bsky.social (BlueSky), and @mike_ketchbrook (X/Twitter) Music powered by OCRemixSecrets Abound (Matoya's Cave) - Final Fantasy: Random Encounter - Midgarian Sky - https://encounter.ocremix.org/Succumb to the Wilderness - Wild Aarms: ARMed and DANGerous - Level 99 - https://armed.ocremix.org/
Topics covered in this episode: NumPy 2.0 release date is June 16 Uvicorn adds multiprocess workers pixi JupyterLab 4.2 and Notebook 7.2 are available Extras Joke Watch on YouTube About the show Sponsored by Mailtrap: pythonbytes.fm/mailtrap Connect with the hosts Michael: @mkennedy@fosstodon.org Brian: @brianokken@fosstodon.org Show: @pythonbytes@fosstodon.org Join us on YouTube at pythonbytes.fm/live to be part of the audience. Usually Tuesdays at 10am PT. Older video versions available there too. Finally, if you want an artisanal, hand-crafted digest of every week of the show notes in email form? Add your name and email to our friends of the show list, we'll never share it. Brian #1: NumPy 2.0 release date is June 16 “This release has been over a year in the making, and is the first major release since 2006. Importantly, in addition to many new features and performance improvement, it contains breaking changes to the ABI as well as the Python and C APIs. It is likely that downstream packages and end user code needs to be adapted - if you can, please verify whether your code works with NumPy 2.0.0rc2.” NumPy 2.0.0 Release Notes NumPy 2.0 migration guide including “try just running ruff check path/to/code/ --select NPY201” “Many of the changes covered in the 2.0 release notes and in this migration guide can be automatically adapted in downstream code with a dedicated Ruff rule, namely rule NPY201.” Michael #2: Uvicorn adds multiprocess workers via John Hagen The goal was to no longer need to suggest that people use Gunicorn on top of uvicorn. Uvicorn can now in a sense "do it all” Steps to use it and background on how it works. Brian #3: pixi Suggested by Vic Kelson “pixi is a cross-platform, multi-language package manager and workflow tool built on the foundation of the conda ecosystem.” Tutorial: Doing Python development with Pixi Some quotes from Vic: “Pixi is a project manager, written in Rust, that allows you to build Python projects without having Python previously installed. It's installable with Homebrew (brew install pixi on Linux and MacOS). There's support in VSCode and PyCharm via plugins. By default, pixi fetches packages from conda-forge, so you get the scientific stack in a pretty reliable and performant build. If a package isn't on conda-forge, it'll look on PyPI, or I believe you can force it to look on PyPI if you like.” “So far, it works GREAT for me. What really impressed me is that I got a Jupyter environment with CuPy utilizing my aging Nvidia GPU on the FIRST TRY.” Michael #4: JupyterLab 4.2 and Notebook 7.2 are available JupyterLab 4.2.0 has been released! This new minor release of JupyterLab includes 3 new features, 20 enhancements, 33 bug fixes and 29 maintenance tasks. Jupyter Notebook 7.2.0 has also been released Highlights include Easier Workspaces Management with GUI Recently opened/closed files Full notebook windowing mode by default (renders only the cells visible in the window, leading to improved performance) Improved Shortcuts Editor Dark High Contrast Theme Extras Brian: Help test Python 3.13! Help us test free-threaded Python without the GIL both from Hugo van Kemenade Python Test 221: How to get pytest to import your code under test is out Michael: Bend follow up from Bernát Gábor “Bend looks roughly like Python but is nowhere there actually. For example it has no for loops, instead you're meant to use bend keyword (hence the language name) to expand calculations and another keyword to join branches. So basically think of something that resembles Python at high level, but without being compatible with that and without any of the standard library or packages the Python language provides. That being said does an impressive job at parallelization, but essentially it's a brand new language with new syntax and paradigms that you will have to learn, it just shares at first look similarities with Python the most.” Joke: Do-while
Topics covered in this episode: PyCon 2024 is up? Ruff formatter is production ready gil--; Why is the Django Admin “Ugly”? Extras Joke Watch on YouTube About the show Sponsored by Scout APM Connect with the hosts Michael: @mkennedy@fosstodon.org Brian: @brianokken@fosstodon.org Show: @pythonbytes@fosstodon.org Join us on YouTube at pythonbytes.fm/live to be part of the audience. Usually Tuesdays at 11am PT. Older video versions available there too. Michael #1: PyCon 2024 is up? May 15 - May 23, 2024 - Pittsburgh, Pennsylvania Conference breakdown: Tutorials: May 15 - 16, 2024 Main Conference and Online: May 17 - 19, 2024 Job Fair: May 19, 2024 Sprints: May 20 - May 23, 2024 Tickets aren't on sale yet Unfortunately, I'm not going (see health and safety guidelines) Attendance numbers over time on Wikipedia Brian #2: Ruff formatter is production ready We reported the alpha release in September It's fast, 30x faster than Black Provides >99.9% compatibility with Black, with a list of known deviations More configurable Bundled with ruff, ruff format Still in Beta, but considered production-ready Integration extensions for VSCode and PyCharm Michael #3: gil--; The Python Steering Council has now formally accepted PEP 703 ("Making the Global Interpreter Lock Optional in CPython") The global interpreter lock will remain the default for CPython builds and python.org downloads. A new build configuration flag, --disable-gil will be added to the configure script that will build CPython with support for running without the global interpreter lock. "In short, the SC accepts PEP 703, but with clear provisio: that the rollout be gradual and break as little as possible, that we can roll back any changes that turn out to be too disruptive – which includes potentially rolling back all of PEP 703 entirely if necessary (however unlikely or undesirable we expect that to be)." Removing the global interpreter lock requires substantial changes to CPython internals, but relatively few changes to the public Python and C APIs. The implementation changes can be grouped into the following four categories: Reference counting Memory management Container thread-safety Locking and atomic APIs Brian #4: Why is the Django Admin “Ugly”? Vince Salvino Some great quotes from the article: "The Django admin is not ugly, rather, no effort was made to make it a beautiful end-user tool.” - Ken Whitesell “The admin's recommended use is limited to an organization's internal management tool. It's not intended for building your entire front end around.” - Django docs “The Django admin was built for Phil.” - Jacob Kaplan-Moss “Even in the 0.9x days we used to have a image that said “Admin: it's not your app”.” - Curtis Maloney As Curtis put it, “encouraging people to build their own management interface, and treat admin as a DB admin tool, has saved a lot of people pain... the effort to customise it grows far faster than the payoffs.” Extras Brian: Local Conferences: Big Potential Michael: Data Science Jumpstart with 10 Projects course is out! PSF is X-ed out (or are they?) GPT4All is pretty excellent Fosstodon invites from us (expires Nov 7 2023) Joke: Searching YouTube for bug fixes
Watch on YouTube About the show Sponsored by Microsoft for Startups Founders Hub. Connect with us Michael: @mkennedy@fosstodon.org Brian: @brianokken@fosstodon.org Show: @pythonbytes@fosstodon.org Join us on YouTube at pythonbytes.fm/stream/live to be part of the audience. Usually Tuesdays at 11am PT. Older video versions available there too. Brian #1: PEP 703 - Making the GIL Optional in CPython Author: Sam Gross Sponsor: Łukasz Langa Draft status, but on Standards Track, targeting Python 3.12 Suggested by: Will Shanks “The GIL is a major obstacle to concurrency.” Especially for scientific computing. PEP 703 proposes adding a --without-gil build configuration to CPython to let it run code without the global interpreter lock and with the necessary changes needed to make the interpreter thread-safe. PEP includes several issues with GIL and sckikit-learn, PyTorch, Numpy, Pillow, and other numerically intensive libraries. Python's GIL makes it difficult to use modern multi-core CPUs efficiently for many scientific and numeric computing applications. There's also a section on how the GIL makes many types of parallelism difficult to express. Changes primarily in internals, and not much exposed to public Python and C APIs: Reference counting Memory management Container thread-safety Locking and atomic APIs Includes information on all of these challenges. Distribution C-API extension authors will need access to a --without-gil Python to modify their projects and supply --without-gil versions. Sam is proposing “To mitigate this, the author will work with Anaconda to distribute a --without-gil version of Python together with compatible packages from conda channels. This centralizes the challenges of building extensions, and the author believes this will enable more people to use Python without the GIL sooner than they would otherwise be able to.” Michael #2: FerretDB Via Jon Bultmeyer A truly Open Source MongoDB alternative MongoDB abandoned its Open-Source roots, changing the license to SSPL making it unusable for many Open Source and Commercial Projects. The core of our solution is a stateless proxy, which converts MongoDB protocol queries to SQL, and uses PostgreSQL as a database engine. FerretDB will be compatible with MongoDB drivers and will strive to serve as a drop-in replacement for MongoDB 6.0+. First release back in Nov 2022 I still love you MongoDB ;) Brian #3: Four tips for structuring your research group's Python packages David Aaron Nicholson Not PyPI packages, but, you know, directories with __init__.py in them. Corrections for mistakes I see frequently Give your packages and modules terse, single-word names whenever possible. Import modules internally, instead of importing everything from modules. Make use of sub-packages. Prefer modules with very specific names containing single functions over modules with very general names like utils, helpers, or support that contain many functions. Michael #4: Quibbler Quibbler is a toolset for building highly interactive, yet reproducible, transparent and efficient data analysis pipelines. One import statement and matplotlib becomes interactive. Check out the video on the repo page. Extras Brian: And now for something completely different: turtles talk Michael: More RSS recommendations FreshRSS a self-hosted RSS and Atom feed aggregator. Feedly (for AI) Flym for Android Readwise is very interesting RSS for courses at Talk Python New article: Dev on the Road Joke: Testing the program Joke: Every Cloud Architecture
Welcome to Code Completion! We are a group of iOS developers and educators hoping to share what we love most about development, Apple technology, and completing your code! This time, we are joined by special guest, James McDougall! Be sure to follow them at https://twitter.com/JamesWMcDougall. Follow us @CodeCompletion (https://twitter.com/CodeCompletion) on Twitter to hear about our upcoming livestreams, videos, and other content. Today, we discuss: * James's path from audio engineer to app developer * How to get an interview when you don't have experience * Fudging your resumes, but not lying to the hiring manager Also, join us for #CompleteTheCode and Compiler Error, two segments that test both your knowledge and our knowledge on Swift, Apple, and all things development! Your hosts for this week: * Dimitri Bouniol (https://twitter.com/dimitribouniol) * Fernando Olivares (https://twitter.com/fromjrtosr) Be sure to also sign up to our monthly newsletter (https://codecompletion.io/), where we will recap the topics we discussed, reveal the answers to #CompleteTheCode, and share even more things we learned in between episodes. You are what makes this show possible, so please be sure to share this with your friends and family who are also interested in any part of the app development process. Sponsor This week's episode of Code Completion is brought to you by Not Phở. Search for Not Phở on the iOS and macOS App Stores today to give it a try. https://apps.apple.com/app/apple-store/id1525104124?pt=14724&ct=CodeCompletion1&mt=8 Complete the Code Be sure to tweet us (https://twitter.com/intent/tweet?text=%23CompleteTheCode%20cc%2F%20%40CodeCompletion&original_referer=https%3A%2F%2Fcodecompletion.io) with hashtag #CompleteTheCode (https://twitter.com/hashtag/CompleteTheCode) if you know the answer! Compiler Error The UTF-32 view provides access to each character as a 4-byte UInt32, representing the complete range of all 4 billion unicode code points, but in their original, possibly decomposed, forms. The UTF-8 view optimizes for memory at a disadvantage to manipulation speed, since each unicode character may be represented by one or more bytes, making length calculations costly, but necessary for many C APIs. Easily accessibly by treating a string like a collection, the default view represents normalized characters as they would be displayed on screen, even if they were originally decomposed. The UTF-16 view is optimal when interacting with unicode libraries such as Objective-C’s NSString since it vends UInt16 code-points, as is standard in many programing models.
In this episode, we cover 3 major announcements in the world of open-source:1. Github makes it all free and lives to its mission - "Every developer on Earth should have access to Github. Price shouldn't be a barrier". 2. Exciting PyTorch 1.5 release including PyTorch Serve, PyTorch Elastic, Full parity C++ APIs and more. A treat for all deep learning users.3. Confluent, a company founded by Apache Kafka developers, raises 250$ MM funding and kicks of project Metamorphosis to take the streaming platform to the next level.References:https://github.blog/2020-04-14-github-is-now-free-for-teams/https://www.confluent.io/press-release/confluent-series-e-round-project-metamorphosis/https://pytorch.org/blog/pytorch-1-dot-5-released-with-new-and-updated-apis/
Welcome to the History of Computing Podcast, where we explore the history of information technology. Because understanding the past prepares us for the innovations of the future! Today we're going to look at Java. Java is an Indonesian island with over 141 million people. Java man lived there 1.7 million years ago. Wait, wrong java. The infiltration of coffee into the modern world can really trace its roots to ancient coffee forests on the Ethiopian plateau. Sufis in Yemen began importing coffee in the 1400s to make a beverage that would aid in concentration and as a kind of spiritual intoxication. Um, still the wrong java… Although caffeine certainly has a link somewhere, somehow. The history of the Java programming language dates back to early 1991. It all started at Sun Microsystems with the Stealth Project. Patrick Naughton had considered going to NeXT due to limitations in C++ and the C APIs. But he stayed to join Stealth, a secret team of engineers led by a developer Sun picked up from Carnegie Mellon named James Gosling . Stealth was formed to explore new opportunities in the consumer electronics market. This came up when Gosling was writing a program to port software from perf to vax and emulating hardware as many, many, many programers had done before him. I wonder if he realized when he went to build the first Java compiler and the original virtual machine code that would go on to write a dozen books about Java and it would consume most of his professional life. I wonder how much coffee he would have consumed if he had. They soon added Patrick Sheridan to the team. The project was later known as the “Green” project and with the advent of the web, somewhat pivoted into more of a web project. You see, Microsoft and the clones had some runaway success but Apple and other vendors were a factor in the home market. But Sun saw going down market as the future of the company. They added a few more people and rented separate offices in Menlo Park. Lisa Friendly was the first employee in the Java Products Group. Gosling would be lead engineer. John Gage would direct the project. Jonni Kanerva would write Java FAQ1. The team started to build C++ ++ —. Sun founder Bill Joy wanted a language that combined the the best parts of Mesa and C. In 1993, NCSA gave us Mozilla. That Andreessen guy was on the news saying the era of the desktop was over. These brilliant designers knew they needed an embedded application, one that could even be used in a web browser, or an applet. The language was initially called “Oak,” but was later renamed “Java” in 1995, supposedly from a list of random words but really due to massive consumption of coffee imported from the island of Java. By the way, it only aids in concentration up to a point. Then you get jumpy. Like a Halfling. It took the Java team 18 months to develop the first working version. It is unknown how much Java they drank in this time. Between the initial implementation of Oak in the fall of 1992 and the public announcement of Java in the spring of 1995, around 13 people ended up contributing to the design and evolution of the language. They were going to build a language that could sit on top of the operating systems on the market. This would allow them to be platform agnostic. In 1995, the team announced that the evolution of Mosaic, Netscape Navigator, would provide support for Java. Java gave us Write Once, Run Anywhere platform independence. You could run the code on a Mac, on Solaris, or on Windows. Java derives its syntax from C and many of the object oriented features were influenced by C++. Several of Java's defining characteristics come from—or are responses to—its predecessors. Therefore, Java was meant to build on these and become a simple, object-oriented, distributed, interpreted, robust, secure, architectural neutral, portable, high performance, multithreaded, and dynamic language. Before I forget. The "Mocha Java" blend pairs coffee from Yemen and Java to get a thick, syrupy, and highly caffeinated blend that is often found with a hint of cinnamon or clove. Similar to all other computer language, all innovation in the design of the language was driven by the need to solve a fundamental problem that the preceding languages could not solve. To start, the creation of C is considered by many to have marked the beginning of the modern age of computer languages. It successfully synthesized the conflicting attributes that had so troubled earlier languages. The result was a powerful, efficient, structured language that was relatively easy to learn. It also included one other, nearly intangible aspect: it was a programmer's language. Prior to the invention of C, computer languages were generally designed either as academic exercises or by bureaucratic committees. C was designed, implemented, and developed by real, working programmers, reflecting how they wanted to write code. Its features were honed, tested, thought about, and rethought by the people who actually used the language. C quickly attracted many followers who had a near-religious zeal for it. As such, it found wide and rapid acceptance in the programmer community. In short, C is a language designed by and for programmers, as is Java. Throughout the history of programming, the increasing complexity of programs has driven the need for better ways to manage that complexity. C++ is a response to that need in C. To better understand why managing program complexity is fundamental to the creation of C++, consider that in the early days of programming, computer programing was done by manually toggling in the binary machine instructions by use of the front panel or punching cards. As long as programs were just a few hundred instructions long, this worked. Then came Assembly and Fortran and then But as programs grew, assembly language was invented so that a programmer could deal with larger, increasingly complex programs by using symbolic representations of the machine instructions. As programs continued to grow, high-level languages were introduced that gave the programmer more tools with which to handle complexity. This gave birth to the first popular programing language; FORTRAN. Though impressive it had its shortcomings as it didn't encourage clear and easy-to-understand programs. In the 1960s structured programming was born. This is the method of programming championed by languages such as C. The use of structured languages enabled programmers to write, for the first time, moderately complex programs fairly easily. However, even with structured programming methods, once a project reaches a certain size, its complexity exceeds what a programmer can manage. Due to continued growth, projects were exceeding the limits of the structured approach. To overcome this problem, a new way to program had to be invented; it is called object-oriented programming (OOP). Object-oriented programming (OOP) is a programming methodology that helps organize complex programs through the use of inheritance, encapsulation, and polymorphism. In spite of the fact that C is one of the world's great programming languages, there is still a limit to its ability to handle complexity. Once the size of a program exceeds a certain point, it becomes so complex that it is difficult to grasp as a totality. While the precise size at which this occurs differs, depending upon both the nature of the program and the programmer, there is always a threshold at which a program becomes unmanageable. C++ added features that enabled this threshold to be broken, allowing programmers to comprehend and manage larger programs. So if the primary motivation for creating Java was the need for a platform-independent, architecture-neutral language, it was to create software to be embedded in various consumer electronic devices, such as microwave ovens and remote controls. The developers sought to use a different system to develop the language one which did not require a compiler as C and C++ did. A solution which was easier and more cost efficient. But embedded systems took a backseat when the Web took shape at about the same time that Java was being designed. Java was suddenly propelled to the forefront of computer language design. This could be in the form of applets for the web or runtime-only packages known as Java Runtime Environments, or JREs. At the time, developers had fractured into the three competing camps: Intel, Macintosh, and UNIX. Most software engineers stayed in their fortified boundary. But with the advent of the Internet and the Web, the problem that the portability of software between platforms suddenly got important in ways it hadn't been since the forming of ARPANET. Even though many platforms are attached to the Internet, users would like them all to be able to run the same program. What was once an irritating but low-priority problem had become a high-profile necessity. The team realized this pressing need and later made the switch to refocus Java from embedded, consumer electronics to Internet programming. So while the desire for an architecture-neutral programming language provided the initial spark, the Internet ultimately led to Java's large-scale success. So if Java derives much of its character from C and C++, this is by intent. The original designers knew that using familiar syntax would make their new language appealing to legions of experienced C/C++ programmers. Java also shares some of the other attributes that helped make C and C++ successful. Java was designed, tested, and refined by real, working programmers. Not scientists. Java is a programmer's language. Java is also cohesive and logically consistent. If you program well, your programs reflect it. If you program poorly, your programs reflect that, too. Put differently, Java is not a language with training wheels. It is a language for professional programmers. Java 1 would be released in 1996 for Solaris, Windows, Mac, and Linux. It was released as the Java Development Kit, or JDK, and to this day we still refer to the version we're using as JDK 11. Version 2, or 1.2 came in 1998 and with the rising popularity we had a few things that the burgeoning community needed. These included event listeners, Just In Time compilers, and change thread synchronizations. 1.3, code named Kestrel came in 2000, bringing RMI for CORBA compatibility, synthetic proxy classes, the Java Platform Debugger Architecture, Java Naming and Directory Interface in core libraries, the HostSpot JVM, and Java Sound. Merlin, or 1.4 came in 2002 bringing the frustrating regular expressions, native XML processing, logging, Non-Blocking I/O, and SSL. Tiger, or 1.5 came in 2004. This was important. We could autobox, get compile time type safety in generics, static import the static part of a class, annotations for declarative programming, and run time libraries were mapped into memory - a huge improvements to how JVMs work. Java 5 also gave us the version number change. So JDK 1.5 was officially recognized as Java 5. JDK 1.6, or Mustang, came in 2006. This was a big update, bringing monitoring and management tools, compiler access gave us programmatic access to javac and pluggable annotations allowed us to analyze code semantically as a step before javac compiles the code. WebStart got a makeover and SE 6 unified plugins with webstart. Enhanced XML services would be important (at least until he advent of son) and you could mix javascript up with Java. We also got JDBC 4, Character Large Objects, SwingWorker, JTable, better SQL datatypes, native PKI, Kerberos, LDAP, and honestly the most important thing was that it was stable. Although I've never written code stable enough to encounter their stability issues… Not enough coffee I suppose. Sun purchased Oracle in 2009. Wait, no, that's one of my Marvel What If comic book fantasies where the world was a better place. Oracle bought Sun in 2009. After ponying up $5.6 billion dollars, Oracle had a lot of tech based on Sun products and seeing Sun as an increasingly attractive acquisition target by other companies, Oracle couldn't risk someone else swooping in and buying Sun. With all the turmoil created, it took 5 years during a pretty formative time on the web, but we finally got Dolphin, or 1.7, which came in 2011 and gave us compressed, 64-bit pointers, strings in switch statements, the ability to make a binary integer and use underscores in literals, better graphics APIs, more cryptography algorithms, and a new I/O library that gave even better platform compatibilities. Spider, or 1.8, came along in 2014. We got the ability to Launch JavaFX application Jars, statically-linked JNI libraries, a new date an time API, annotation for java types, unsigned integer arithmetic, a JavaScript runtime that allowed us to embed Javascript code in apps - whether this is a good idea or not is still tbd. Lambda functions had been dropped in Java 7 so here we also got lambda expressions. And this kickstarted a pretty interesting time in the development of Java. We got 9 in 2017, 10 and 11 in 2018, 12, 13, and 14 in 2019. Of these, only 8 and 11 are LTS, or commercial Long Term Support releases, basically meaning we got the next major release after 8 in 2018 and according to my trend line should expect the next LTS in 2021 or 2022. JDK 13, when released later in 2019, will give us text blocks, Switch Expressions, improved memory management by returning unused heap memory to the OS, improves application class and data sharing, and brings back the legacy socket API. But it won't likely be an LTS release. Today there are over 45 billion active Java Virtual Machines and java remains arguably the top language for micro service, ci/cd environments, and a number of other use cases. Other languages have come. Other languages have gone. Many are better in their own right. Some are not. Java is not perfect. It was meant to reduce complexity. But as languages evolve they become more complex. A project with a million lines of code is monolithic and probably incorporates plugins or frameworks like spring security as an example, that make code even more complex. But Java is meant to reduce cyclomatic complexity, to allow for a language that is simple enough for a professional to pick up quickly and only be as complex as the quality of the code being compiled. I don't personally love Java. I respect it. And I adore high-quality programmers and their code in any language. But I've had to redo so much work because other languages have come and gone over the years that if I were to be starting a new big monolithic web-app today, I'd probably use Java every time. Which isn't to say that Java isn't useful in micro-service architectures. According to what's required from the contract testing on a service, I might use Java, Go, node, python or even the formerly hipster Ruby. Although I don't love drinking PBR… If I'm writing an Android app, I need to know Java. No matter what the lawyers say. If I'm planning on an enterprise webapp, Java needs to be in the conversation. But usually, I can do the work in a fraction of the time using something like python. But most big companies speak Java. And for good reason. Because of the write once run anywhere approach and the level of permissions a JRE needs, there have been security challenges with running Java on desktop computers. Apple deprecated Java on the Mac in 2010. Users could still instal lications and is the gold standard for those. I'm certainly not advocating going back to the 90s and running Java apps on our desktops any more. No matter what you think of Java, one thing you have to admit, the introduction of the language and the evolution have had a substantial impact on the IT industry and it will continue to do so. A great takeaway here might be that there's always a potential alternative that might be better suited for a given task. But when it comes to choosing a platform that will be there in a decade or 3, getting support, getting a team that can scale, sometimes you might end up using a solution that doesn't immediately seem as well suited to a need. But it can get the job done. As it's been doing since James Gosling and the rest of the team started the project back in the early 90s. So thank you listeners, for sticking with us through this episode of the History of Computing Podcast. We're lucky to have you.
Automatic generation of FFI types between Rust and C APIs. Show Notes bindgen the official guide API docs cbindgen a blog post by one of the main authors of the crate rust-ffi Binding libraries for scripting languages PyO3 Neon Rustler a great write-up from the folks at Discord Helix Sponsors Thanks to Parity for sponsoring the show and hiring Rust developers! Patreon Sponsors Adam Green Aleksey Pirogov Alexander Kryvomaz Alexander Lozada Alexander Payne Andrew Dirksen Andrew Thompson Anthony Deschamps Anthony Scotti Arlen Haftevani Arlo (Hyena) Arun Kulshreshtha Behnam Esfahbod Benjamin Manns Benjamin Wasty Brandon ‘Spanky’ Mills Brian Casiello Brian Manning Brian McCallister Bryan Stitt Caryn Finkelman Cass Costello Cat Dad Chap Lovejoy Charlie Egan Chip Chris Palmer Christoffer Ceutz Cristian Paul Dan Abrams Daniel Daniel Bross Daniel Collin Daniel Mason David Carroll David Hewson Derek Morr Dominic Cooney Doug Reeves Douglas Correa Edmund Kump Eduard Knyshov Embark Studios Eugene Bulkin Evan Stoll Fabio (decathorpe) Fabio Correa Freeman P. Pascal Gaveen Prabhasara Graham Wihlidal Henri Sivonen Ian Jones Hoàng Đức Hiếu Hugo Josefson “Jake”“ferris”" Taylor" Jako Danar James Cooper James Hagans II Jason Bowen Jeff May Jendrik Illner Jerome Froelich JockeTF Joar Wandborg Johan Andersson John Rudnick Jon Jonah Jonathan Knapp Jonathan Turner Joseph Hain Joseph Mou Joseph Schrag Joe Percy Justin Ossevoort Kai Yao Kazutaka Mise Keith Gray Kilian Rault Kyle Lee Jenkins Luca Schmid Luiz Irber Lukas Eller Malnormalulo Martin Heuschober Masashi Fujita Matt Rudder Matthew Brenner Matthias Ruszala Max Jacobson Max R.R. Collada Messense Lv Micael Bergeron Michael Mc Donnell Michael Melanson Michael Sanders Nathan Sculli Nick Coish Nick Gideo Nick Stevens Nicolas Pochet Oladapo Fadeyi Olaf Leidinger Oliver Uvman Oluseyi Sonaiya Ovidiu Curcan Pascal Patrick O’Doherty Paul Naranja Paul Osborne Peter Scholtens Peter Tillemans Pierre-Antoine Champin Ralph Giles Ramon Buckland Randy MacLeod Raph Levien Richard Dallaway Rob Tsuk Robbie Clarken Robert Chrzanowski Romain Chossart Ryan Blecher Ryan Osial Satoshi Yoshikawa Scott Moeller Sebastián Ramírez Magrí Simon Dickson Simon G Soren Bramer Schmidt Steve Jenson Steven Knight Steven Murawski Stuart Hinson Sven Huster Tim Brooks Tim Small Tim Süberkrüb Tom Prince Toolmaker’s Guild Ty Overby Tyler Harper Victor Kruger Will Greenberg Zak van der Merwe Zachary Snyder Zach Peters Zaki (Thanks to the couple people donating who opted out of the reward tier, as well. You know who you are!) Become a sponsor Patreon Venmo Dwolla Cash.me Flattr PayPal.me Contact New Rustacean: Twitter: @newrustacean Email: hello@newrustacean.com Chris Krycho GitHub: chriskrycho Twitter: @chriskrycho
In the Application Security News, Breach at IT outsourcer Wipro, SCP serves the file it wants, Confluence Path traverses to RCE, another Local PrivEsc on Windows, easier sandboxing for C and C++ APIs, and Computer Science plus Ethics! Full Show Notes: https://wiki.securityweekly.com/ASW_Episode58 Follow us on Twitter: https://www.twitter.com/securityweekly
This week, we welcome Thomas Hatch, the creator of the Salt open source software project, and is the CTO of SaltStack, the company behind Salt! In the Application Security News, Breach at IT outsourcer Wipro, SCP serves the file it wants, Confluence Path traverses to RCE, another Local PrivEsc on Windows, easier sandboxing for C and C++ APIs, and Computer Science plus Ethics! To learn more about SaltStack, visit: https://securityweekly.com/saltstack Full Show Notes: https://wiki.securityweekly.com/ASW_Episode58 Visit our website: https://www.securityweekly.com Follow us on Twitter: https://www.twitter.com/securityweekly Like us on Facebook: https://www.facebook.com/secweekly
In the Application Security News, Breach at IT outsourcer Wipro, SCP serves the file it wants, Confluence Path traverses to RCE, another Local PrivEsc on Windows, easier sandboxing for C and C++ APIs, and Computer Science plus Ethics! Full Show Notes: https://wiki.securityweekly.com/ASW_Episode58 Follow us on Twitter: https://www.twitter.com/securityweekly
This week, we welcome Thomas Hatch, the creator of the Salt open source software project, and is the CTO of SaltStack, the company behind Salt! In the Application Security News, Breach at IT outsourcer Wipro, SCP serves the file it wants, Confluence Path traverses to RCE, another Local PrivEsc on Windows, easier sandboxing for C and C++ APIs, and Computer Science plus Ethics! To learn more about SaltStack, visit: https://securityweekly.com/saltstack Full Show Notes: https://wiki.securityweekly.com/ASW_Episode58 Visit our website: https://www.securityweekly.com Follow us on Twitter: https://www.twitter.com/securityweekly
FreeBSD 11.2 has been released, setting up an MTA behind Tor, running pfsense on DigitalOcean, one year of C, using OpenBGPD to announce VM networks, the power to serve, and a BSDCan trip report. ##Headlines FreeBSD 11.2-RELEASE Available FreeBSD 11.2 was released today (June 27th) and is ready for download Highlights: OpenSSH has been updated to version 7.5p1. OpenSSL has been updated to version 1.0.2o. The clang, llvm, lldb and compiler-rt utilities have been updated to version 6.0.0. The libarchive(3) library has been updated to version 3.3.2. The libxo(3) library has been updated to version 0.9.0. Major Device driver updates to: cxgbe(4) – Chelsio 10/25/40/50/100 gigabit NICs – version 1.16.63.0 supports T4, T5 and T6 ixl(4) – Intel 10 and 40 gigabit NICs, updated to version 1.9.9-k ng_pppoe(4) – driver has been updated to add support for user-supplied Host-Uniq tags New drivers: + drm-next-kmod driver supporting integrated Intel graphics with the i915 driver. mlx5io(4) – a new IOCTL interface for Mellanox ConnectX-4 and ConnectX-5 10/20/25/40/50/56/100 gigabit NICs ocs_fc(4) – Emulex Fibre Channel 8/16/32 gigabit Host Adapters smartpqi(4) – HP Gen10 Smart Array Controller Family The newsyslog(8) utility has been updated to support RFC5424-compliant messages when rotating system logs The diskinfo(8) utility has been updated to include two new flags, -s which displays the disk identity (usually the serial number), and -p which displays the physical path to the disk in a storage controller. The top(1) utility has been updated to allow filtering on multiple user names when the -U flag is used The umount(8) utility has been updated to include a new flag, -N, which is used to forcefully unmount an NFS mounted filesystem. The ps(1) utility has been updated to display if a process is running with capsicum(4) capability mode, indicated by the flag ‘C’ The service(8) utility has been updated to include a new flag, -j, which is used to interact with services running within a jail(8). The argument to -j can be either the name or numeric jail ID The mlx5tool(8) utility has been added, which is used to manage Connect-X 4 and Connect-X 5 devices supported by mlx5io(4). The ifconfig(8) utility has been updated to include a random option, which when used with the ether option, generates a random MAC address for an interface. The dwatch(1) utility has been introduced The efibootmgr(8) utility has been added, which is used to manipulate the EFI boot manager. The etdump(1) utility has been added, which is used to view El Torito boot catalog information. The linux(4) ABI compatibility layer has been updated to include support for musl consumers. The fdescfs(5) filesystem has been updated to support Linux®-specific fd(4) /dev/fd and /proc/self/fd behavior Support for virtio_console(4) has been added to bhyve(4). The length of GELI passphrases entered when booting a system with encrypted disks is now hidden by default. See the configuration options in geli(8) to restore the previous behavior. In addition to the usual CD/DVD ISO, Memstick, and prebuilt VM images (raw, qcow2, vhd, and vmdk), FreeBSD 11.2 is also available on: Amazon EC2 Google Compute Engine Hashicorp/Atlas Vagrant Microsoft Azure In addition to a generic ARM64 image for devices like the Pine64 and Raspberry Pi 3, specific images are provided for: GUMSTIX BANANAPI BEAGLEBONE CUBIEBOARD CUBIEBOARD2 CUBOX-HUMMINGBOARD RASPBERRY PI 2 PANDABOARD WANDBOARD Full Release Notes ###Setting up an MTA Behind Tor This article will document how to set up OpenSMTPD behind a fully Tor-ified network. Given that Tor’s DNS resolver code does not support MX record lookups, care must be taken for setting up an MTA behind a fully Tor-ified network. OpenSMTPD was chosen because it was easy to modify to force it to fall back to A/AAAA lookups when MX lookups failed with a DNS result code of NOTIMP (4). Note that as of 08 May 2018, the OpenSMTPD project is planning a configuration file language change. The proposed change has not landed. Once it does, this article will be updated to reflect both the old language and new. The reason to use an MTA behing a fully Tor-ified network is to be able to support email behind the .onion TLD. This setup will only allow us to send and receive email to and from the .onion TLD. Requirements: A fully Tor-ified network HardenedBSD as the operating system A server (or VM) running HardenedBSD behind the fully Tor-ified network. /usr/ports is empty Or is already pre-populated with the HardenedBSD Ports tree Why use HardenedBSD? We get all the features of FreeBSD (ZFS, DTrace, bhyve, and jails) with enhanced security through exploit mitigations and system hardening. Tor has a very unique threat landscape and using a hardened ecosystem is crucial to mitigating risks and threats. Also note that this article reflects how I’ve set up my MTA. I’ve included configuration files verbatim. You will need to replace the text that refers to my .onion domain with yours. On 08 May 2018, HardenedBSD’s version of OpenSMTPD just gained support for running an MTA behind Tor. The package repositories do not yet contain the patch, so we will compile OpenSMTPD from ports. Steps Installation Generating Cryptographic Key Material Tor Configuration OpenSMTPD Configuration Dovecot Configuration Testing your configuration Optional: Webmail Access iXsystems https://www.forbes.com/sites/forbestechcouncil/2018/06/21/strings-attached-knowing-when-and-when-not-to-accept-vc-funding/#30f9f18f46ec https://www.ixsystems.com/blog/self-2018-recap/ ###Running pfSense on a Digital Ocean Droplet I love pfSense (and opnSense, no discrimination here). I use it for just about anything, from homelab to large scale deployments and I’ll give out on any fancy for a pfSense setup on a decent hardware. I also love DigitalOcean, if you ever used them, you know why, if you never did, head over and try, you’ll understand why. . Unfortunately, while DO offers tremendous amount of useful distros and applications, pfSense isn’t one of them. But, where there’s a will, there’s a way, and here’s how to get pfSense up and running on DO so you can have it as the gatekeeper to your kingdom. Start by creating a FreeBSD droplet, choose your droplet size (for modest setups, I find the 5$ to be quite awesome): There are many useful things you can do with pfSense on your droplet, from OpenVPN, squid, firewalling, fancy routing, url filtering, dns black listing and much much more. One note though, before we wrap up: You have two ways to initiate the initial setup wizard of the web-configurator: Spin up another droplet, log into it and browse your way to the INTERNAL ip address of the internal NIC you’ve set up. This is the long and tedious way, but it’s also somewhat safer as it eliminates the small window of risk the second method poses. or Once your WAN address is all setup, your pfSense is ready to accept https connection to start the initial web-configurator setup. Thing is, there’s a default, well known set of credential to this initial wizard (admin:pfsense), so, there is a slight window of opportunity that someone can swoop in (assuming they know you’ve installed pfsense + your wan IP address + the exact time window between setting up the WAN interface and completing the wizard) and do . I leave it up to you which of the path you’d like to go, either way, once you’re done with the web-configurator wizard, you’ll have a shiny new pfSense installation at your disposal running on your favorite VPS. Hopefully this was helpful for someone, I hope to get a similar post soon detailing how to get FreeNAS up and running on DO. Many thanks to Tubsta and his blogpost as well as to Allan Jude, Kris Moore and Benedict Reuschling for their AWESOME and inspiring podcast, BSD Now. ##News Roundup One year of C It’s now nearly a year that I started writing non-trivial amounts of C code again (the first sokol_gfx.h commit was on the 14-Jul-2017), so I guess it’s time for a little retrospective. In the beginning it was more of an experiment: I wanted to see how much I would miss some of the more useful C++ features (for instance namespaces, function overloading, ‘simple’ template code for containers, …), and whether it is possible to write non-trivial codebases in C without going mad. Here are all the github projects I wrote in C: sokol: a slowly growing set of platform-abstraction headers sokol-samples - examples for Sokol chips - 8-bit chip emulators chips-test - tests and examples for the chip- emulators, including some complete home computer emulators (minus sound) All in all these are around 32k lines of code (not including 3rd party code like flextGL and HandmadeMath). I think I wrote more C code in the recent 10 months than any other language. So one thing seems to be clear: yes, it’s possible to write a non-trivial amount of C code that does something useful without going mad (and it’s even quite enjoyable I might add). Here’s a few things I learned: Pick the right language for a problem C is a perfect match for WebAssembly C99 is a huge improvement over C89 The dangers of pointers and explicit memory management are overrated Less Boilerplate Code Less Language Feature ‘Anxiety’ Conclusion All in all my “C experiment” is a success. For a lot of problems, picking C over C++ may be the better choice since C is a much simpler language (btw, did you notice how there are hardly any books, conferences or discussions about C despite being a fairly popular language? Apart from the neverending bickering about undefined behaviour from the compiler people of course ;) There simply isn’t much to discuss about a language that can be learned in an afternoon. I don’t like some of the old POSIX or Linux APIs as much as the next guy (e.g. ioctl(), the socket API or some of the CRT library functions), but that’s an API design problem, not a language problem. It’s possible to build friendly C APIs with a bit of care and thinking, especially when C99’s designated initialization can be used (C++ should really make sure that the full C99 language can be used from inside C++ instead of continuing to wander off into an entirely different direction). ###Configuring OpenBGPD to announce VM’s virtual networks We use BGP quite heavily at work, and even though I’m not interacting with that directly, it feels like it’s something very useful to learn at least on some basic level. The most effective and fun way of learning technology is finding some practical application, so I decided to see if it could help to improve networking management for my Virtual Machines. My setup is fairly simple: I have a host that runs bhyve VMs and I have a desktop system from where I ssh to VMs, both hosts run FreeBSD. All VMs are connected to each other through a bridge and have a common network 10.0.1/24. The point of this exercise is to be able to ssh to these VMs from desktop without adding static routes and without adding vmhost’s external interfaces to the VMs bridge. I’ve installed openbgpd on both hosts and configured it like this: vmhost: /usr/local/etc/bgpd.conf AS 65002 router-id 192.168.87.48 fib-update no network 10.0.1.1/24 neighbor 192.168.87.41 { descr "desktop" remote-as 65001 } Here, router-id is set vmhost’s IP address in my home network (192.168.87/24), fib-update no is set to forbid routing table update, which I initially set for testing, but keeping it as vmhost is not supposed to learn new routes from desktop anyway. network announces my VMs network and neighbor describes my desktop box. Now the desktop box: desktop: /usr/local/etc/bgpd.conf AS 65001 router-id 192.168.87.41 fib-update yes neighbor 192.168.87.48 { descr "vmhost" remote-as 65002 } It’s pretty similar to vmhost’s bgpd.conf, but no networks are announced here, and fib-update is set to yes because the whole point is to get VM routes added. Both hosts have to have the openbgpd service enabled: /etc/rc.conf.local openbgpdenable="YES" Conclusion As mentioned already, similar result could be achieved without using BGP by using either static routes or bridging interfaces differently, but the purpose of this exercise is to get some basic hands-on experience with BGP. Right now I’m looking into extending my setup in order to try more complex BGP schema. I’m thinking about adding some software switches in front of my VMs or maybe adding a second VM host (if budget allows). You’re welcome to comment if you have some ideas how to extend this setup for educational purposes in the context of BGP and networking. As a side note, I really like openbgpd so far. Its configuration file format is clean and simple, documentation is good, error and information messages are clear, and CLI has intuitive syntax. Digital Ocean ###The Power to Serve All people within the IT Industry should known where the slogan “The Power To Serve” is exposed every day to millions of people. But maybe too much wishful thinking from me. But without “The Power To Serve” the IT industry today will look totally different. Companies like Apple, Juniper, Cisco and even WatsApp would not exist in their current form. I provide IT architecture services to make your complex IT landscape manageable and I love to solve complex security and privacy challenges. Complex challenges where people, processes and systems are heavily interrelated. For this knowledge intensive work I often run some IT experiments. When you run experiments nowadays you have a choice: Rent some cloud based services or DIY (Do IT Yourself) on premise Running your own developments experiments on your own infrastructure can be time consuming. However smart automation saves time and money. And by creating your own CICD pipeline (Continuous Integration, Continuous Deployment) you stay on top of core infrastructure developments. Even hands-on. Knowing how things work from a technical ‘hands-on’ perspective gives great advantages when it comes to solving complex business IT problems. Making a clear distinguish between a business problem or IT problem is useless. Business and IT problems are related. Sometimes causal related, but more often indirect by one or more non linear feedback loops. Almost every business depends of IT systems. Bad IT means often that your customers will leave your business. One of the things of FeeBSD for me is still FreeBSD Jails. In 2015 I had luck to attend to a presentation of the legendary hacker Poul-Henning Kamp . Check his BSD bio to see what he has done for the FreeBSD community! FreeBSD jails are a light way to visualize your system without enormous overhead. Now that the development on Linux for LXD/LXD is more mature (lxd is the next generation system container manager on linux) there is finally again an alternative for a nice chroot Linux based system again. At least when you do not need the overhead and management complexity that comes with Kubernetes or Docker. FreeBSD means control and quality for me. When there is an open source package I need, I want to install it from source. It gives me more control and always some extra knowledge on how things work. So no precompiled binaries for me on my BSD systems! If a build on FreeBSD fails most of the time this is an alert regarding the quality for me. If a complex OSS package is not available at all in the FreeBSD ports collection there should be a reason for it. Is it really that nobody on the world wants to do this dirty maintenance work? Or is there another cause that running this software on FreeBSD is not possible…There are currently 32644 ports available on FreeBSD. So all the major programming language, databases and middleware libraries are present. The FreeBSD organization is a mature organization and since this is one of the largest OSS projects worldwide learning how this community manages to keep innovation and creates and maintains software is a good entrance for learning how complex IT systems function. FreeBSD is of course BSD licensed. It worked well! There is still a strong community with lots of strong commercial sponsors around the community. Of course: sometimes a GPL license makes more sense. So beside FreeBSD I also love GPL software and the rationale and principles behind it. So my hope is that maybe within the next 25 years the hard battle between BSD vs GPL churches will be more rationalized and normalized. Principles are good, but as all good IT architects know: With good principles alone you never make a good system. So use requirements and not only principles to figure out what OSS license fits your project. There is never one size fits all. June 19, 1993 was the day the official name for FreeBSD was agreed upon. So this blog is written to celebrate 25th anniversary of FreeBSD. ###Dave’s BSDCan trip report So far, only one person has bothered to send in a BSDCan trip report. Our warmest thanks to Dave for doing his part. Hello guys! During the last show, you asked for a trip report regarding BSDCan 2018. This was my first time attending BSDCan. However, BSDCan was my second BSD conference overall, my first being vBSDCon 2017 in Reston, VA. Arriving early Thursday evening and after checking into the hotel, I headed straight to the Red Lion for the registration, picked up my badge and swag and then headed towards the ‘DMS’ building for the newbies talk. The only thing is, I couldn’t find the DMS building! Fortunately I found a BSDCan veteran who was heading there themselves. My only suggestion is to include the full building name and address on the BSDCan web site, or even a link to Google maps to help out with the navigation. The on-campus street maps didn’t have ‘DMS’ written on them anywhere. But I digress. Once I made it to the newbies talk hosted by Dan Langille and Michael W Lucas, it highlighted places to meet, an overview of what is happening, details about the ‘BSDCan widow/widower tours’ and most importantly, the 6-2-1 rule! The following morning, we were present with tea/coffee, muffins and other goodies to help prepare us for the day ahead. The first talk, “The Tragedy of systemd” covered what systemd did wrong and how the BSD community could improve on the ideas behind it. With the exception of Michael W Lucas, SSH Key Management and Kirk McKusick, The Evolution of FreeBSD Governance talk, I pretty much attended all of the ZFS talks including the lunchtime BoF session, hosted by Allan Jude. Coming from FreeNAS and being involved in the community, this is where my main interest and motivation lies. Since then I have been able to share some of that information with the FreeNAS community forums and chatroom. I also attended the “Speculating about Intel” lunchtime BoF session hosted by Theo de Raddt, which proved to be “interesting”. The talks ended with the wrap up session with a few words from Dan, covering the record attendance and made very clear there “was no cabal”. Followed by the the handing over of Groff the BSD goat to a new owner, thank you’s from the FreeBSD Foundation to various community committers and maintainers, finally ending with the charity auction, where a things like a Canadian $20 bill sold for $40, a signed FreeBSD Foundation shirt originally worn by George Neville-Neil, a lost laptop charger, Michael’s used gelato spoon, various books, the last cookie and more importantly, the second to last cookie! After the auction, we all headed to the Red Lion for food and drinks, sponsored by iXsystems. I would like to thank the BSDCan organizers, speakers and sponsors for a great conference. I will certainly hope to attend next year! Regards, Dave (aka m0nkey) Thanks to Dave for sharing his experiences with us and our viewers ##Beastie Bits Robert Watson (from 2008) on how much FreeBSD is in Mac OS X Why Intel Skylake CPUs are sometimes 50% slower than older CPUs Kristaps Dzonsons is looking for somebody to maintain this as mentioned at this link camcontrol(8) saves the day again! Formatting floppy disks in a USB floppy disk drive 32+ great indie games now playable on OpenBSD -current; 7 currently on sale! Warsaw BSD User Group. June 27 2018 18:30-21:00, Wheel Systems Office, Aleje Jerozolimskie 178, Warsaw Tarsnap ##Feedback/Questions Ron - Adding a disk to ZFS Marshall - zfs question Thomas - Allan, the myth perpetuator Ross - ZFS IO stats per dataset Send questions, comments, show ideas/topics, or stories you want mentioned on the show to feedback@bsdnow.tv
In the lead up to WWDC, Ben and Jelly take an early look at some of the changes coming to the Swift language this year with version 3, which like one of the biggest releases so far. With widespread changes in the wording of method names, improvements to how C APIs are handled, and a Swift version of Foundation, it’s going to be a big change to both your code, and how you read it.
Opening theme by Larry King (@laking). The case for a modernized AppKit in OS X 10.9 a la Chameleon/TwUI. The lack of official Apple Objective-C wrappers around old C APIs such as Keychain and Address Book. Springboard #10 SMOP Underscore Should Apple add type inference to Objective-C? Haswell Retina MacBook Pro predictions and the possibility of having only the integrated GPU in the 15". gfxCardStatus The only 15" with integrated-only GPU The first unibody 15" with high-res, matte LCD Haswell integrated GPU explained Mac Pro speculation. John on Ad Hoc Larrabee Intel's custom CPU for the first MacBook Air Thunderbolt problems on Xeon boards Marco's Mac Pro post this week Scaling down the Mac Pro/"xMac" challenges Thunderbolt 2 Asus' 4K monitor The first 27" iMac's panel economics Major iOS 7 API wishes: better inter-app communication mechanisms, "default app" associations, and periodic background updates. Quick WWDC tips. Casey's WWDC predictions Casey's WWDC tips from last year Sponsored by: Tonx: Freshly roasted coffee delivered straight to your door. New customers can get a free AeroPress by signing up for a Standard subscription by June 17. Squarespace: The all-in-one platform that makes it easy to create your own website. Use coupon code ATP6 for 10% off.