Posted on

Herding Code 122: Bert Belder on porting Node.js to Windows

This episode of Herding Code the guys talk to Bert Belder, a Node.js developer who’s working on the native Windows port.

  • Kevin asks how Bert got started with Node.js. Bert explains that he was working on a PHP based system which had a good amount of logic in Javascript, and he started looking to node as a way to consolidate that logic.
  • K. Scott ask Bert about how you’d go about sharing Javascript between client and server.
  • Kevin asks how the Windows port of Node.js got started, and whether there was any resistance to it.
  • Jon asks if the eventing model in Windows was one of the more difficult things Bert had to work on. Bert explains that getting REPL (read evaluate print loop) to work on Windows as it did on Unix was actually one of the more difficult challenges at the beginning.
  • K. Scott asks about experience of running Node.js. Bert talks about how Node.js runs as an executable, and it’s up to you to set up an HTTP server.
  • Kevin asks about the team that’s working on the Windows port.
  • Kevin asks Bert for more info on technical challenges on getting Node.js not only working on Windows, but really performing. Bert describes some challenges in implementing I/O Completion Ports, spawning child processes, etc.
  • Jon says that some of the initial negative feedback he’d seen on the announcement of the Windows port was concern over negative performance implications for the Unix version, and Bert says that one of the criteria they’re working under is that Unix performance not be degraded.
  • Kevin asks Bert if he had a background in high performance networking, of if he’d been  figuring things out as he went along. Bert explains that he and the team have had a good amount of time to work on this, so they had a pretty good idea of how to solve this.
  • Jon asks about performance testing, and Bert describes some of the load tests that they use.
  • K. Scott asks about if most Node.js modules will work on Windows. Bert says he guesses 90% will work, and the ones that won’t are making operating-system specific assumptions.
  • Jon asks Bert about his work on libuv, the abstracted platform layer for Node.js, and Kevin asks about the process of designing this abstraction layer.
  • Kevin asks Bert if there are places where Windows is more flexible or powerful than Linux. Bert explains that since node was written for Unix first, it wasn’t built to exploit Windows advantages, but with Windows kernel mode HTTP stack might be useful in the future.
  • K. Scott asks if the port could have been aided by open source projects like Cygwin. Bert explains how Cygwin isn’t really helpful in making node.js work well on Windows.
  • Jon asks if dropping support for Cygwin in Node.js will affect users, and Bert says the only effect may be for addons which are written so as to be Unix-only.
  • Kevin asks about Windows support for NPM.
  • There’s a discussion about the different hosting options on Windows including Azure and iisnode. Kevin asks about service hosting to allow socket level access (below IIS).
  • Kevin asks if there’s a plan for cross-platform system support for modules that need native access. Bert talks about changes to npm to support binary hosting so operating system specific binaries can be automatically downloaded.
  • Kevin asks if there will be a focused effort to get node module authors to support Windows.
  • Jon asks about multi-core scenarios. Bert talks about different options, explaining that iisnode can help with this, but he hasn’t seen options for interprocess communication in iisnode or in in either multi-core scenarios.
  • Kevin asks about the use of gyp for the node.js build process and about Bert’s development environment on Windows.
  • Kevin wraps up by asking Bert how to pronounce his twitter handle.

Show Links:

Download / Listen:

Herding Code 122: Bert Belder on porting node.js to Windows

[audio://herdingcode.com/wp-content/uploads/HerdingCode-0122-Bert-Belder-on-porting-nodejs-to-Windows.mp3]
Posted on

Herding Code 121: Sara Chipps updates us on Girl Develop It at one year

This episode of Herding Code Kevin and Jon catch up with Sara Chipps to find out how Girl Develop It is going.

  • Kevin jumps right into it by asking Sara about what’s been going on over the past year. Sara goes back to what was on their minds as they were first getting started with their first class, and how that’s grown to 6 cities worldwide, and their original New York chapter offers 20 classes a month with 5 teachers.
  • Kevin asks where the other teachers came from, and Sara spins a tall tale about flamethrower classes. Well, maybe it’s true, but I find it a bit suspect.
  • Kevin asks about where the classes are held.
  • Jon asks about how the money part works out. Sara explains how the class fees, donations, and teacher payments all work out.
  • Kevin asks Sara about about what tools they teach, and Sara mentions Aptana.
  • Jon asks for some success stories and Sara tells a few.
  • Kevin asks how many students go through several classes; Sara says they see about 25% frequent fliers.
  • Jon asks if there’s some measurement of how much the students actually learn. Sara explains that the classes include a good amount of hands-on work and homework, and that she and the other teachers continue to learn how to gage when students are getting lost.
  • Kevin asks if there are some students that just don’t get it, and Sara says that some students have a tough time understanding that a single missing character can break a whole program. Everyone commiserates about this fun part of software development. Jon speculates that young women may be missing out on some of the split between cold logic and reason because they don’t play enough video games.
  • Jon notes that a lot of real world computer programming involves problem solving and support network and asks if students are equipped with those things. Sara talks about how anyone watching her code will see a good amount of debugging; additionally she teaches students about how to use StackOverflow so they can get their questions solved.
  • Kevin asks how the curriculum and courses have evolved over the past year. Sara and Jon talk about the amount of time and effort involved in preparing decent training materials.
  • Kevin asks about changes in teaching approach over the course of the past year. Sara said she’s moved from code-only to using some slides, and that when writing code it’s important to walk through it in pretty good detail.
  • Kevin asks what’s been different from expectations, and Sara talks about both the amount of interest and community goodwill.
  • Kevin asks about how Girl Develop It has spread to other cities, and asks about how much of the content is shared between cities.
  • Jon asks about how the branding and design is handled.
  • Jon asks if there are advanced classes or seminars.
  • Jon asks Sara what sort of projects she’s been working on lately. Sara says it’s pretty much all Javascript lately.
  • Kevin asks if there will be a node.js class (drink!)
  • Kevin asks what’s next for Girl Develop It, which prompts Sara to talk about the first Girl Develop It hack-a-thon. Apparently these are like a guy hack-a-thon except with less pizza and body odor and more resort and catering.
  • Kevin asks if Girl Develop It could develop into a full time gig. Sara says that all the leaders love developing and don’t want to give that up, so they’re still trying to figure that out.
  • Jon asks how listeners can support Girl Develop It. Sara lists a range of options, including book, laptop donations, and meeting space. Kevin asks about cash contributions.
  • Jon and Kevin note that there are no West Coast US branches. Sara mentions that a bay area location may start soon.
  • Kevin asks about the 15% male attendance in Girl Develop It, and Sara explains how that works.
  • Jon asks if there’s potential for virtual events and video recordings. Sara explains that, while it sounds great logistically, it misses out on a lot of the most important aspects of the Girl Develop It classroom experience. They’ve got trial running in the Columbus branch, though, so they’ll see how it goes.
  • Sara teases about an interesting hack-a-thon project they did recently using the Aviary API’s called Stash Your Stash, which removes moustaches from photos because "they’re super creepy!"

Show Links:

Download / Listen:

Herding Code 121: Sara Chipps updates us on Girl Develop It at one year

[audio://herdingcode.com/wp-content/uploads/HerdingCode-0121-Sara-Chipps-updates-us-on-Girl-Develop-It.mp3]
Posted on

Herding Code 120: Ryan Stewart on RIAs and All Things Adobe

This episode of Herding Code the guys talk to Ryan Stewart, a developer evangelist at Adobe.

  • Scott K asks about the pricing of Adobe products. Ryan explains why things are priced as they are and talks about the subscription model alternative.
  • Jon talks about the open other free or inexpensive alternatives for beginning Adobe development since the formats are generally open sourced. Ryan agrees and also explains that Adobe’s entire pricing model is built around tools, whereas Microsoft’s includes both tools and servers.
  • Jon asks about the licensing around Flash Media Server. Ryan explains that it’s not something general developers will need to deal with.
  • Scott K speculates around the idea of appliances from Adobe which would be complete video / media processing systems. Ryan says he thinks that’s interesting but he doesn’t expect anything like that to happen.
  • Jon asks about the general trends away from some rich internet abuses in the past. Ryan and Jon talk about the polyfill approach for using Flash and RIA technologies to augment browsers when the features aren’t supported. Ryan and Scott K talk about how developers and the tech press have quickly forgotten that many of the new emerging browser capabilities (typography, media, animation) are modeled after capabilities that RIA technologies initially pioneered.
  • Scott K points out that this is a good segue to the new Edge tool. Ryan talks about how Edge is a designer tool that creates CSS3 and JavaScript animation using the timeline that Flash designers are used to. Jon mentions that this makes more sense to him when thinking about Adobe primarily as a developer tools company rather than a platform oriented company.
  • Jon asks about Adobe support for HTML5 / CSS3 development in Edge with Flash fallback. Ryan talks about how they generally keep them separate, and if browser-based animation isn’t supported it just won’t play.
  • Kevin asks about how Edge affects the accessibility of the underlying content.
  • Jon asks about how Flash Builder 4.5 allows for developing native iOS and Android applications. Ryan explains how it works and clarifies how it complies with Apple developer guidelines. There’s a mention of the popular Mono apps which run on iOS.
  • Twitter question from Chris Edwards: "What are the are the best tools for automated testing of Flash UI’s" – Ryan recommends HP Quick Test Pro
  • Scott K asks about Adobe Air – it seemed great, but seems to have kind of fallen out of favor. What’s the deal there? Ryan talks about how Air was both a great, bold idea, but also a new challenge for Adobe, in that Air applications are much longer running than most Flash apps.
  • Jon asks about some annoyances in installing updates for Air, Flash, etc. Ryan explains some of the reasons for the updates. Jon asks about the possibility to add in more of an auto-update experience.
  • Jon asks Ryan about some of the new features in Flash. Ryan talks about a lot of features, including Stage Video and 3D GPU support and graphics features. Scott K. asks if there are opportunities for leveraging WebGL, and Ryan says that there have been discussions about that but nothing’s in progress yet.
  • Ryan asks the guys what they’re expecting at at BUILD, and they all clam up. Scott K. ask about Flash on Windows Phone. Ryan says it’d be great, but he’d be surprised given the Silverlight support on Windows Phone. Kevin’s happy that speculation will finally stop. Scott K. talks it’s good that developers are having to care about memory and CPU usage again.
  • Ryan talks about the difficulty of bringing richness and creativity to the client without adversely impacting performance.
  • Scott K asks about the current state of Adobe Labs. Ryan points out that the Adobe MAX conference is coming up in October, so Labs will probably be pretty quiet until then.
  • Jon asks about Muse. Ryan explains that really targeted at print designers who want to create web content, so it’s not really a tool for web designers or developers.
  • Jon asks about some Adobe client products which are developed in Air.
  • Scott K. asks about what big surprises may be happening at MAX this year. Ryan and and Scott K. speculate a bit more about BUILD, and Ryan tells listeners who recognize him at BUILD to please say hi. Jon says he might sneak in if he can locate a catering costume.

Show Links:

Download / Listen:

Herding Code 120: Ryan Stewart on RIAs and all things Adobe

[audio://herdingcode.com/wp-content/uploads/HerdingCode-0120-Ryan-Stewart-on-RIAs-and-all-things-Adobe.mp3]
Posted on

Herding Code 119: On The Writing Technical Books (with Jesse Liberty, Phil Haack, and Brad Wilson)

This episode of Herding Code the guys talk to Jesse Liberty, Phil Haack, and Brad Wilson about writing technical books. Jesse has written dozens of technical books, and both Brad and Phil worked with Jon and K. Scott on the recently released ASP.NET Professional MVC 3 book. What’s it like to write a book? Why do it at all? How does the process work? How is it changing? Is Angry Birds your favorite story? Join us for a very literary version of Herding Code to find out.

  • We start with a listener question from @tpdorsey (Terrence Dorsey): “*Printed* books? If so, why bother? I ask this as a print writer and editor for 17 of last 20 years.”
  • There’s a discussion of the value that the editorial process adds to books as compared to blog posts.
  • Twitter question from @schwarty: “Best way to work a first time deal? Submit queries? Try to team up with co-writers?” Phil, Brad, Jon, K. Scott, and Jesse tell stories about how they got their first book deals.
  • Phil notes that most authors get a start in other mediums – blogs, magazines, possibly StackOverflow in the future.
  • Jesse talks about his investigation of self-publishing. In the end, he decided that the editorial process tipped the scales towards working with a publishing company.
  • Jon references Twitter questions from @devhammer and @jglozano on dealing with procrastination, making time to write, etc. He points to positive pressure of working for a publisher under a schedule and a contract to get the book out the door. Brad talks about how peer pressure from other authors to keep up is also helpful.
  • There’s discussion about the challenge writing up an outline before starting the book. Phil talks about how the publishing industry in general is pretty stuck on older technologies like FTP, and in general the process feels like waterfall software development.
  • Jesse says that if you’re reasonably on schedule, publishers are pretty flexible about changes to the outline.
  • Brad answers listener questions from @stevenproctor “Have you found e-readers coming into their own to influence how you think about book?” and “has it changed how you think about layout/presentation for cross format reading” saying that it was difficult to write without seeing what the end result would look like, but he was very happy when he saw the end result, both in print and the e-book format.
  • Jon talks about how working with book templates from various publishers has taught him to appreciate the use of styles in Word, explaining a case where he was able to search for code snippets based on the styles that were used.
  • Kevin asks some interesting questions about whether working on the book puts pressures on when to ship the product, and whether writing a book about a product points out features that should be changed. Brad explains that the king of product changes due to writing (blogging in this case) is Scott Guthrie.
  • Phil talks about how Eilon (the technical reviewer) was good at keeping him from digging too far into minute details that nobody would care about. Jon explains that Eilon pointed out that the Controllers chapter was going way into the weeds before actually showing the most common use case, and Brad says that he’s in good company with an example from Charles Petzold’s book on Windows programming.
  • Jon talks about the tricky chicken-and-egg situation with trying to explain the MVC pattern in depth, since an in-depth explanation of the Model, View, and Controller requires an in-depth understanding of the other components. Jon asks Jesse about how he handles that, and Jesse talks about the importance of having a model user and getting volunteer readers as you’re writing.
  • Jon talks about the benefit of working with other authors. Brad talks about the this is more important with products with quick release cycles, and Phil compares book writing with software development techniques.
  • There’s a discussion on ensuring a consistent voice in a book with more than one author.
  • Phil, Jon, and Brad discuss the conflict between beginner and advanced content. Are the experts who are asking for advanced content representative of most readers? Phil points out that writing advanced content is a lot more fun, but limits the audience. Can a book please everyone? Is there a way to include beginner and advanced content? Jesse describes some ways he handles this, and says that it’s important to set expectations.
  • Jon talks about the decision to remove NerdDinner, referring instead to the MVC Music Store tutorial.
  • Jesse says that it’s remarkable that books are still selling well, since they’re competing with blogs, tutorials, videos, etc. He speculates that the main selling point for books now is in context: telling a story. Jon asks how he manages to do that.
  • Phil talks about the decision to put all the book’s code samples in NuGet (triggering the Haacked NuGet Drinking Game clause).
  • Jesse talks about the mismatch between the publisher’s requirements for a flow of completed chapters and the software developer’s desire to refactor.
  • Jon asks K. Scott how writing magazine articles compares to writing for a book.
  • That’s pretty much it.

Show Links:

Download / Listen:

Herding Code 118: On The Writing Technical Books (with Jesse Liberty, Phil Haack, and Brad Wilson)

[audio://herdingcode.com/wp-content/uploads/HerdingCode-0119-On-The-Writing-Of-Technical-Books.mp3]
Posted on

Herding Code 118: Paul Betts on SassAndCoffee

This episode of Herding Code the guys talk to Paul Betts about SassAndCoffee, a NuGet package that adds runtime Sass and CoffeeScript compilation to ASP.NET.

  • Jon asks Paul about his role on the Office Labs team [Spoiler alert! Since this podcast, Paul has started a new job at GitHub!]
  • Jon asks Paul about why he got interested in Sass and Coffee for web development.
  • Jon asks Paul how an Office dev gets away with liking Ruby and Python.
  • Paul describes Sass (Syntactically Awesome Stylesheets), a meta-language for CSS that adds in variables, nested rules, inheritance, mixins, etc.
  • Paul explains how SassAndCoffee is designed to eliminate fiddling, so he’s gone to great lengths to make the NuGet package just work without any setup.
  • Kevin points out that this is at run-time rather than at build. Paul talks about the advantages of run-time compilation (especially interactive CSS edit / refresh) as well as potential downsides (performance, potential for compilation errors). Scott K also mentions that it might be useful for CDN deployment and continuous integration.
  • Jon asks about the new package definition file the 1.0 release, and Paul explains how it tells Visual Studio that the .coffee files are to be included with the project build.
  • Paul explains how the CoffeeScript compiler works using an HttpHandler, Jurrassic, and V8. Jon asks if he’d looked at IronJS, and Paul describes why that didn’t work for him.
  • Paul explains the hurdles he went through to get V8 running under an ASP.NET HttpHandler, since V8 assumes that it will always be accessed from a single threaded process.
  • Jon asks Paul about his use of uglify.js for Javascript optimization and compression.
  • Scott K asks about the ability to swap out other compilers, e.g. the Google Closure Compiler.
  • Paul talks about some of the commits he’s had recently, including support for Nancy, better cache configuration and some useful refactoring. Jon and Paul discuss how some refactoring patches – especially blind Resharpering – are less than helpful.
  • The discussion shifts to how Paul got Sass working without requiring the user to have a local Ruby installation, including some crazy tricks with the DLR’s platform resource library to embed a portion of the Ruby standard library as an embedded resource in the NuGet package via a virtual R: drive.
  • Jon asks if it’s possible for others to reuse Paul’s Ruby embedding technique in other applications.
  • Jon, K. Scott, and Paul discuss commenting policy, and Paul explains why he liberally commented certain sections of the code.
  • Paul mentions how the V8 integration falls back to Jurassic in case it can’t run for some reason.
  • Kevin asks if Paul looked into creating native ports of Sass or CoffeeScript, and Paul explains why he decided to set up compilers for the original versions rather than port them.
  • Jon talks about how he’d looked at Zen Coding and mused about how to implement it in Visual Studio, deciding as Paul had that it’s better
  • Twitter question from @elijahmanor about IDE support for Sass and CoffeeScript. Note: since the podcast, the Mindscape Web Workbench has made this available.
  • Scott K points out that the MVC 4 roadmap includes support for recipes, which should help with extending the IDE via NuGet.
  • Paul wraps up the nerdy internals of SassAndCoffee by explaining how he’s handling loading the 32 and 64 bit versions of the V8 C++ CLI DLL depending on the user’s CPU architecture.
  • Jon asks about what server-side hosting requirements are needed to run SassAndCoffee.
  • Kevin asks whether the V8 API was easy to work with, and Paul says no.
  • Jon mentions that he appreciated all the thanks that Paul had in the readme, and Paul talks about how great it feels to thank people.
  • Jon asks what’s next for SassAndCoffee, and Paul says he wants to add Compass.
  • Jon (jokingly) mentions that he’s a fan of BrainScript and asks for support.
  • Paul and Jon discuss some of the nuttier esoteric languages they’ve seen.
  • Jon asks what else Paul’s up to; Paul talks about the Reactive Extensions book he’s working on with Jesse Liberty. Jon asks Paul a bit about what’s new and interesting with Reactive Extensions.
  • Jon asks Paul about the interaction between Reactive Extensions and Interactive Extensions, and Paul explains by way of a Reactive Extensions history lesson.
  • There’s a discussion on the (intentional) lack of a ForEach operator in LINQ
  • K Scott tells his terrifying story from NDC in which Eric Lippert was sitting in his talk.
  • Scott K asks about Reactive UI, another of Paul’s projects, an MVVM framework which leverages Reactive Extensions.

Show Links:

Download / Listen:

Herding Code 118: Paul Betts on SassAndCoffee

[audio://herdingcode.com/wp-content/uploads/HerdingCode-0118-Paul-Betts-on-SassAndCoffee.mp3]