Discussions

News: Google cloud languages: Python pedals to peak? Java on the way?

  1. According to a recent survey of more than 500 developers from Evans Data, Python use has risen 45% since the spring of 2009! The company correlates these results with the release of the Google App Engine cloud development platform. The App engine only supported Python when released in April of 2008, though it has recently come to support Java. What do you think? Will Python peak in the Google cloud? Will Java take over sooner or later? What about GO? http://www.businesswire.com/portal/site/home/permalink/?ndmViewId=news_view&newsId=20091110006763

    Threaded Messages (11)

  2. OLPC[ Go to top ]

    ┬┐What about OLPC and Sugar? Python is the main language there and the project is pretty recent.
  3. Re: OLPC[ Go to top ]

    good is 110 % tahnk you is survey
  4. Re: OLPC[ Go to top ]

    is ood is and to go jeson
  5. "A recent survey from Evans Data claims Python use has risen 45% since the spring of 2009! Mmmm. Kind of non-linear." Is it? Here's an example of linear growth: # increase ----------- 1 --- 2 100% 3 50% 4 33% 5 25% 6 20% 7 16.7% 8 14.3& 9 12.5% 10 11.1% 11 10% ... 100 1% 45% probably is a big jump but without previous numbers, there's no way to tell if that is the case. A yearly increase of 1% is non-linear. I can't really say for sure but this seems kind of shaky. 380+ respondents? Maybe that's statistically significant but could easily be non-representative. 45% across the set of all developers seems a bit unlikely. Have Python book sales gone up similarly? What about hits to Python websites or posts on Python forums?
  6. Java will overtake in time[ Go to top ]

    java will overtake in time. They release Java support recently so obviously it takes a lot for community to adopt. Also many Java apps are in production and not easy to move on GAE right away. GAE has to enable SQL support to Java apps and see the jump right away but that won't happen in near future. Also JDO or JPA as datastore is not very popular or say easy to migrate from existing SQL base backend. I think one question every developer or company will ask when get into cloud "how easy to migrate and costly to port away from GAE"
  7. python sucks[ Go to top ]

    Quote from Google enginner "Python will still be slower than C and Java, use more memory and have inferior threading " http://groups.google.com/group/unladen-swallow/browse_thread/thread/4edbc406f544643e
  8. And your point is????[ Go to top ]

    The same points were made about Java - and if those were the only grounds on which to choose a language, we'd all be using assembler. There were people who thought the move from punched cards to interactive VDU terminals was a waste of precious computing resource. Overall, though, the trend since the 1950s has been to higher-level, more abstract languages - like Java - balanced against machine resources. Java is just occupies the current 'sweet spot'.
  9. Oops[ Go to top ]

    Shame my reply wasn't in line with the comment!
  10. Re: And your point is????[ Go to top ]

    The same points were made about Java - and if those were the only grounds on which to choose a language, we'd all be using assembler.
    Actually, the situation is quite different. If you check the linked thread, there's a lot of talk about how application developers rewrote hot spots either in cython or plain C. This was never the case with Java. The closest we got was native parts of VM, linking to capabilities of native OS (IO, DirectX...) and a few esoteric things like high performance encryption. "Going native" was never a serious option for application developers.
  11. Re: And your point is????[ Go to top ]

    The same points were made about Java - and if those were the only grounds on which to choose a language, we'd all be using assembler.
    Agree completely
    There were people who thought the move from punched cards to interactive VDU terminals was a waste of precious computing resource.

    Overall, though, the trend since the 1950s has been to higher-level, more abstract languages - like Java - balanced against machine resources.
    I am not convinced of this. I think there has always been a place for high level languages, like shell and JCL in certain domains.


    Java is just occupies the current 'sweet spot'.
    Java's sweet spot does not seem to be in a performance/abstraction dimension. Slow interpreted languages, like PHP, work fine for web applications from a performance perspective. And Java is not the best choice for high performance, although Java's performance issues in many specific domain seem to be solvable if that was a priority: http://marc.info/?l=git&m=124111702609723&w=2 So Java is not the best choice for writing something like git, from a performance perspective.
  12. Re: And your point is????[ Go to top ]

    And Java is not the best choice for high performance
    For problems where hot-spot operations needs access to some kind of low level primitives Java is going to be slow, however, to the best of my knowledge, there is no such platform to develop applications for high performance business application computing as Azul systems (http://www.azulsystems.com/). These guys are running JVMs on 800+ CPUs, multigigabyte heaps and what not. (their JVM is based on OpenJDK if I am not mistaken) If you ever wondered on which platform Doug Lea is doing research for next generation of concurrent/parallel Java facilities, check Azul out. PS I'm in no way associated with Azul etc...