Yupp, it’s already here. If you have no clue what TW Radar is about, feel free to check my post for the last year edition (http://gebski.tumblr.com/post/51321405824/bandit-on-your-6) or go straight to the Radar’s website: a picture’s worth the thousand words: http://www.thoughtworks.com/radar/#/
Short map:
- Adopt - good stuff, mature stuff
- Trial - promising, do some dogfeeding or try in non-critical scenario
- Assess - not for everyone, not every time; PoC it first
- Hold - no go
To the point then: what’s new, what’s cool, what’s interesting and are there any surprises - let’s go through the categories:
Techniques:
http://www.thoughtworks.com/radar/#/techniques
- DOM abstraction + JS testing on Node.js gets straight to Adopt - can’t agree more.
- Windows infrastructure automation finally gets some love (appears in Adopt)! It’s awesome and I’m all for, but IMHO there’s still very little contribution in the community (when compared to *nixes).
- Server-less provisioning with Chef/Puppet appears in Trial (both with provisioning testing!) - personally, I’m much more for the distributed version (client-server).
Tools:
http://www.thoughtworks.com/radar/#/tools
- Plenty, plenty of new entries - that’s both good and bad. Good, because things are going on and bad, because the potential maturity may not be sufficient for enterprise.
- Personally I’m happy with the following entries: D3 in Adopt (it requires a lot of effort, but the results may be ground-braking), grunt.js, Logstash & Phantomjs in Trial; Docker, Octopus & Xamarin in Assess. Where’s Icenium though?
- TFS remains in Hold. Die in fire.
Platforms:
http://www.thoughtworks.com/radar/#/platforms
- Now it should get really sexy, and it does, but not exactly in the way you expect: Adopt category didn’t change since last year, so there’s no new platform that got that proven within last 12 months. Not many changes in Trial as well - Hadoop 2.0 didn’t make a breakthrough either.
- Plenty of new stuff in Assess category: TypeSafe’s Akka (my favourite!), Google’s SPDY, Twitter Storm (as an adopter, can’t agree more) and at last but not least - Web Components as a standard.
- There’s also a lot of interesting stuff going on in Hold: bane of CMS as a platform for complex IT solutions (hail to that! do you hear that, Sharepoint?), death wish to traditional, rigid-structured enterprise DWHs.
Languages & frameworks:http://www.thoughtworks.com/radar/#/languages-and-frameworks
- No surprises - Adopt stays the same as it was a year ago (but it was some kind of surprise then). I really wonder how many people really use Clojure in production code and how hard is it to hire a Clojure programmer.
- Some new interesting contestants in Trial category - Go (ain’t it too early yet?), Hive (yepp, it’s so much more convenient that writing the Map/Reduce jobs on your own) and … Reactive Extensions. About the last one - personally, I am a big fan and I find it a true masterpiece created by ultrasmart minds - but in my opinions people still are learning how (or rather “where”) to use it. It’s that different to stuff that was available before.
- Assess is a big bag of new stuff. Most noticeable: Elixir (Erlang ain’t so hipsta anymore? :D), TypeScript (my new web dev love) and Yeoman (Yo + Grunt + Bower, can’t work w/o them).
To summarize:
There are no big surprises, in general I totally agree with the TW report (well, that’s why I’ve investigated many of these technologies before), but I’d still add few points that are missing:
- web app frameworks are completely missing and IMHO it’s a mistake - they matured & proven themselves a lot within last year
- if MongoDB is in Adopt for platforms, where should Apache Cassandra be? ;P
- Personally, I give all my votes to bring Apache Cordova to Adopt - I believe that it’s a matter of (short) time until native dev dies on mobile - unification is the key, so everyone can make money of that.
- I am very happy to see that TW keeps “big enterprise solutions” in Hold category for the next year - it should be emphasized as strong as possible - XXI century says ‘no’ to humongous, untestable clusterf*cks of 20yo code, that give job to armies of drone consultants who don’t know any honest work.