Dhananjay Nene

A passionate programmer, software architect, and a enthusiastic observer of internet technologies and usage. Having worked in a variety of roles, the essential passion that drives me is that of a Programmer. Architecture and Design issues excite me, C++, Java and Python I am quite competent at, and Scala is one where I would like to get there.

Tweets for 2010-02-08

Couldn't find a good Object-XML mapper for #python .. now painstakingly creating one #
@psnively I was actually searching 4 mappers, but thanks for the tip .. will checkout generateDS to see if it saves me effort n time in reply to psnively #
@psnively Was unclear in earlier tweet. I meant a runtime mapper which took a object – xml bindings as a map parameter. in reply to psnively #
RT: @debasishg #Ruby Blocks .. a good one .. http://yehudakatz.com/2010/02/07/the-building-blocks-of-ruby/ #
I get amused when Ruby / Python proponents get overly passionate about differences. eg. Ruby in syntax sugar & Python in functions. (1/2) #
But there are very few things that either language is overly superior to the other. Choose the one with a better feel and go with it (2/2) #
Rails 3 Beta is Out — A Retrospective http://ff.im/-fB2vn #
Scala Self-Type Annotations for Constrained Orthogonality http://ff.im/-fB2vo #
Django Advent – About Django 1.2 http://ff.im/-fB3UR #

February 8th, 2010 in Tweets | No comments

Tweets for 2010-02-07

@IndianGuru isn't that a bit of a risky thought ? I thought metaprogramming is best avoided till there is no effective alternative in reply to IndianGuru #
@IndianGuru What I meant was it shouldn't take understanding metaprogramming (which should be optional) to grok a language in reply to IndianGuru #
@IndianGuru Its a truism to say you understand X better if you understand MetaX. (1/2) /cc @arnab_deka in reply to IndianGuru #
@IndianGuru What I'm not so sure of is whether it requires one to understand MetaRuby to understand (start thinking in) Ruby /cc @arnab_deka in reply to IndianGuru #
@IndianGuru I think I grokked it eventually.. the distinction was between understanding / being competent in vs. thinking in /cc @arnab_deka in reply to IndianGuru #

February 7th, 2010 in Tweets | No comments

Tweets for 2010-02-05

Martin Fowler: ConversationalStories http://ff.im/-fsW4G #
NoSQL Means Never Having to Store Blobs Again http://ff.im/-fsWZd #
High Availability Principle : Concurrency Control http://ff.im/-fsXds #

February 5th, 2010 in Tweets | No comments

Tweets for 2010-02-03

No I will not blog about hiphop Bracing for a day full of hiphop related posts after the last week was completely spoilt by ipad posts #
What are the properties of "Functional Programming Languages"? (from LTU) http://ff.im/-fiFjM #
@psnively actually the point I was making was that Jinja2, Twig and Liquid are virtually identical templating langs on different host langs in reply to psnively #
RT: @alex_gaynor Blogged: my thoughts on hip hop php: http://alexgaynor.net/2010/feb/02/thoughts-hiphop-php/ #
He He Reddit Clone in 10 mins & 91 lines Clojure http://is.gd/7CLT5 97 lines PHP http://is.gd/7CLY9 & Bash/Git 14 secs http://is.gd/7CM8j #
Android and the Linux kernel community http://ff.im/-fkHnV #
What Pythonistas Think of Ruby http://ff.im/-fkLnt #

February 3rd, 2010 in Tweets | No comments

Tweets for 2010-02-02

Continuous Delivery: Continuous Delivery http://ff.im/-fgkow #
Software on the Cheap http://ff.im/-fgkKX #
Jonathan Andrew Wolter: Simplicity is Better for Deploying in Production Web Architectures http://ff.im/-fglIZ #
Improving clarity by removing temporal coupling http://ff.im/-fgmZW #
Pattern matching in Scala for expressions http://ff.im/-fgnhX #
RT: @_pythonic An interview with Michael Foord (@voidspace) about python http://ff.im/-fgrpH #
Will LLVM become the JVM equivalent for dynamic languages ? Now powers Python, Ruby, and PHP implementations #
Nice comment by @rickardoberg wrt EventSourcing http://is.gd/7yrr8 on @KentBeck's post on NoSQL http://is.gd/7yrAQ for PolySQL persistence #
Hmm.. PolySQL Persistence: n The strategy of using different database types (relational, key value, graph) for different subsets of data. #
@codemonkeyism Oh he absolutely rocks .. I still remember the JMX based microkernel for JBoss in reply to codemonkeyism #
@sbtourist I could be wrong. I thought @rickardoberg's use of EventSourcing was a more concrete manifestation of @debasishg's messaging in reply to sbtourist #
Its not a promise, its a guess http://is.gd/7yGE7 by @dhh Meant for a parallel universe ? I can't imagine this world buying it #
@psnively while EventSourcing is based on EventCalculus I think the level of complexity seems way way way too distant. in reply to psnively #
@psnively though I shall follow up on Event Calculus to see if it could have applications in simpler domains e.g. business txn processing in reply to psnively #
@codemonkeyism apparently seems from 50%scala/50%php earlier http://is.gd/7yXGs /cc @debasishg in reply to codemonkeyism #
IMO Scala is a near logical choice for PHP sites to migrate to when facing scaling challenges while lacking facebook budgets #
re: last tweet .. PHP/Scala .. I meant near most likely (not near logical) #
@davetron5000 despite attempts to railsify PHP, any migration from PHP would require a substantial change anyways & for a site facing (1/2) in reply to davetron5000 #
@davetron5000 extreme scaling challenges such as four square, the decision is likely to lean more towards Scala/Java than Python/Ruby (2/2) in reply to davetron5000 #
@psnively Hmm.. I could easier explain Event sourcing to a biz guy who superficially understands software than Event calculus to a developer in reply to psnively #
@davetron5000 agreed. Sometimes I am able to make my changes using python faster than it takes the java compiler to bootstrap itself in reply to davetron5000 #
@davetron5000 I now believe OS, devices & IO should be written in C, plumbing (eg messaging) on JRE/C, & business logic/rules in Python/Ruby in reply to davetron5000 #
@davetron5000 I considered Scala/Clojure in JRE.. imo reqd 4 the highest speed / concurrency. FWIW I use lot of FP constructs in python now. in reply to davetron5000 #
@indrajitr wrt JS .. I sooo totally agree. Groovy perhaps due to personal biases I could never quite get used to so wouldn't know. in reply to indrajitr #
@psnively I come from the (biased?) point of view that a large proportion of apps today written on JRE never needed that complexity (1/2) in reply to psnively #
@psnively Wrt economics & business agility, most apps are better off with quickly modifiable biz tier (2/2), otoh plumbing shud b robust in reply to psnively #
@psnively I generally use economics followed by (business) agility as more important criteria than technology per se with my tech choices in reply to psnively #
@michaelg is it C++ .. I thought they just compiled PHP to LLVM bytecode in reply to michaelg #
@kaleidic Yes .. but especially so when they need to continuously react to changing business requirements.. thats the business agility angle in reply to kaleidic #
@kaleidic Having said that .. I am a Scala novice in reply to kaleidic #
@debasishg nice post .. especially liked the replay / hotswap "fun extrema" in reply to debasishg #
@psnively I think most biz apps need to deal with MQ, DB, XML, UI (JSP/Templating) .. and the JEE learning curve still high for Scala in reply to psnively #
@psnively I have this feeling for pure algorithms, Scala is just as productive as python / ruby and in more complex cases more productive in reply to psnively #
@psnively ha! though one would've imagined that the mainstream choices would naturally reflect good economics in reply to psnively #
@psnively sometimes yes and sometimes no VB was one of the best economics friendly mainstream product even though it was #sillytech for many in reply to psnively #
@psnively OTOH, Java (1.3+) is an extraordinarily superb runtime which unfortunately became much more mainstream than necessary in reply to psnively #
@psnively Having said that (Java 1.3 superb runtime) my big issue with Java is productivity and JEE is the sheer learning curve for avg apps in reply to psnively #
@psnively other biz factors- qual/qty of dev, dev migration across projects, ability 2 make tech choices.. different across geographies in reply to psnively #
@davetron5000 You nailed the part on tweak / reload cycle in reply to davetron5000 #
@davetron5000 I changed my choice of templating language to Jinja2 due to the tweak-reload cycle factor in reply to davetron5000 #
@wmartinez Third which works best especially with skilled teams – design as an outcome of coding (or design implicit in coding) in reply to wmartinez #
From the Facebook blog : HipHop for PHP: Move Fast : http://developers.facebook.com/news.php?blog=1&story=358 #
So autoconverting PHP to C++ "only halves" CPU utilisation ? I'm not questioning their ability .. but doesn't quite seem right. #
@psnively You're right. But makes me wonder .. was CPU the constraining bottleneck & if not, what was the cost/benefit of hiphop (1/2) in reply to psnively #
@psnively the exercise would be substantially useful only if CPU utilisation came down from 100% to 50% (not from 60% to 30%) (2/2) in reply to psnively #
@psnively this makes for a good contextual reference too http://shootout.alioth.debian.org/u32q/benchmark.php?test=all&lang=php&lang2=gpp in reply to psnively #
@davetron5000 JSTL (and TagLibs) are a result of a meme which believed HTML & Java don't go together and should be sandboxed from each other in reply to davetron5000 #
@michaelg you're right .. earlier reports now obviously incorrect stated they were compiling to LLVM in reply to michaelg #
@psnively Ahh the pleasure of mashing turing complete & incomplete languages w/ sandboxing & w/o restarts http://is.gd/7zKtb @davetron5000 in reply to psnively #
@psnively FWIW check out the similarity with http://is.gd/5O4aQ and http://is.gd/5O5h3 #near_universal_templating in reply to psnively #

February 2nd, 2010 in Tweets | No comments

Tweets for 2010-02-01

A Guide to Branching in Mercurial http://ff.im/-fbnnR #
Google summer of code 2010 is under way http://is.gd/7sHeP http://is.gd/7sH1S 2009 site http://socghop.appspot.com/ #

February 1st, 2010 in Tweets | No comments

Tweets for 2010-01-29

@jneira neither do you groan at its limitations .. .. .. until you have it. The grass is always greener on the other side. in reply to jneira #

January 29th, 2010 in Tweets | No comments

Tweets for 2010-01-28

@codemonkeyism hmm .. I think the word "dead" probably has a far higher correlation than votes in reply to codemonkeyism #
Answering Baranovskiy’s JavaScript quiz http://ff.im/-f11BB #

January 28th, 2010 in Tweets | No comments

Tweets for 2010-01-26

@surya_s Both offer similar capabilities. Git is more fine grained and flexible, mercurial simpler. Both good choices for a DVCS in reply to surya_s #
@surya_s check out http://importantshock.wordpress.com/2008/08/07/git-vs-mercurial/ re: git/mercurial in reply to surya_s #

January 26th, 2010 in Tweets | No comments

Tweets for 2010-01-22

RT: @dstanek I believe that REST is as much an architectural style as SOA. I don't care what these guys say: http://bit.ly/6IIgXH #
Firefox 3.6 brings joy to Web devs, not just users. A review by @arstechnica http://ff.im/-eDOQl #
Locating and managing the IS security function http://ff.im/-eDOSg #

January 22nd, 2010 in Tweets | Comments Off

Tweets for 2010-01-21

Constructor over-injection anti-pattern http://ff.im/-eCjc5 #
The CPU Clock principle of software releases (Oct 2009) by Bertrand Meyer http://ff.im/-eCksK #
@codemonkeyism Yes, tho am not sure if always avoidable. In the xample better inject the factory than the object. eg ConnPoll not Connection in reply to codemonkeyism #
@codemonkeyism Rather than focusing on constructor weight, I would focus on pooling or lazy construction (through injected factories) in reply to codemonkeyism #

January 21st, 2010 in Tweets | Comments Off

Tweets for 2010-01-20

From @dailywtf : An acquired realtime forex trading system http://ff.im/-ez5AL #

January 20th, 2010 in Tweets | Comments Off

Tweets for 2010-01-19

Programming Styles http://ff.im/-euXpN #
Code-used Time-used Shapes – A lang shootout modeling of time n brevity http://ff.im/-euY87 #
The Maybe Type and Its Consequences http://ff.im/-euYe0 #
Owners vs Caretakers http://ff.im/-evD1C #
The big bang was a result of dereferencing the null pointer #
@codemonkeyism is Option similar to the Maybe monad or is it something different (sorry for my ignorance) ? in reply to codemonkeyism #

January 19th, 2010 in Tweets | Comments Off

links for 2010-01-18

January 19th, 2010 in Bookmarks | Comments Off

Tweets for 2010-01-18

Hard core multicore with TestNG « Otaku, Cedric's weblog http://ff.im/-erKIu #
@DocOnDev Nice one .. really enjoyed it. in reply to DocOnDev #
More on multithreaded topological sorting « Otaku, Cedric's weblog http://ff.im/-erLCq #

January 18th, 2010 in Tweets | Comments Off

Tweets for 2010-01-17

@codemonkeyism Not sure who has a good clue on what exactly happens with the GIL. I sure hope it goes away, but keeping my fingers crossed. in reply to codemonkeyism #
RT: @davetron5000 RT @unclebobmartin New Blog on Dependency Injection Inversion: http://bit.ly/6JUc6T <- couldn't have said it better myself #
@unclebobmartin interestingly the old spring IOC implementation using XML config does maintain centralised one place dependencies in reply to unclebobmartin #
@_sunil_ first link not working re: modeling trees in reply to _sunil_ #

January 17th, 2010 in Tweets | Comments Off

Tweets for 2010-01-16

@codemonkeyism why do you have to use KDE to run Hadoop in a VM ? in reply to codemonkeyism #
@fmarinelli I think it was 16K with available upgrades http://en.wikipedia.org/wiki/Zx_spectrum#ZX_Spectrum_16K.2F48K in reply to fmarinelli #

January 16th, 2010 in Tweets | Comments Off

Software development is about the middle and not just the end

View Blog Post

January 16th, 2010 in Blog | Comments Off

links for 2010-01-15

January 16th, 2010 in Bookmarks | Comments Off

Tweets for 2010-01-15

Nice read #design #patterns #practices Patterns in Practice: A Retrospective http://bit.ly/73Eqw0 by @codebetter via @az1 #
Giggle is a nice git client. Seems to be a little less featured but a much cleaner look. (And has gravatar integration for the mugshots) #

January 15th, 2010 in Tweets | Comments Off