tag:blogger.com,1999:blog-1231480044619721857.post1682894398409005844..comments2022-12-02T05:07:00.365-05:00Comments on Elastician: Does Python Scale?Mitch Garnaathttp://www.blogger.com/profile/02589240083555476561noreply@blogger.comBlogger5125tag:blogger.com,1999:blog-1231480044619721857.post-27389308829865643112011-11-04T12:10:30.564-04:002011-11-04T12:10:30.564-04:00Scalability is not a matter of languages, but a ma...Scalability is not a matter of languages, but a matter of architecture.<br /><br />Of course, there are some languages which make easy to set up some scalable architecture (Erlang comes to mind), but it's really more complex than Language A don't scale/Language B does (or SQL doesn't scale/NoSQL does)<br /><br />great postJaimehttps://www.blogger.com/profile/02966274109038870865noreply@blogger.comtag:blogger.com,1999:blog-1231480044619721857.post-49021597654239273742011-10-18T02:56:47.363-04:002011-10-18T02:56:47.363-04:00Interesting that you mention "huge sites"...Interesting that you mention "huge sites" which all have huge user bases, does that mean that they are big projects with huge code bases? No. Sure you can get the execution time to scale in python, because as Eduardo puts it; it's all about the architecture, but does it scale in development time? I think not because in the end you rely on test cases to verify that you don't break things whilst in statically typed languages, like Java, you can utilize the compiler more (javac is my best friend). Sure you still need tests to verify things but as you can rely on the compiler to give you errors when something doesn't build you don't have to run those tests as often. In big projects you often end up with a lot of test cases, the time to run all of them to verify that you haven't broken anything often increases incrementally (this may not be true if you only employ super humans that refactor often). This is why I often say; use python for small code bases because when prototyping a product you want to be fast and python is faster in the beginning but as your code base grows you'll se that your development efficiency will stagger and this is when you should consider a statically typed language instead. A feature that is critical when working in huge code bases is to see how/where code is used. When developing Java in eclipse you press ctrl+shift+g, I've tried the same feature for a python project and it doesn't work because the types in python aren't declared so the search can't reliably give you the correct answer, sure it can guess but who likes guessing? I think this is the killer feature of dynamically typed languages.Anonymoushttps://www.blogger.com/profile/08189466424462437082noreply@blogger.comtag:blogger.com,1999:blog-1231480044619721857.post-10130957520965548912011-10-14T19:27:15.504-04:002011-10-14T19:27:15.504-04:00My general answer to that question is.
No . Pytho...My general answer to that question is.<br /><br />No . Python doesn't scale.<br /><br />But the good news is, it doesn't have to. Since it plays a very small role in scalability. In the end the architecture of your app will make it scale or not. <br /><br />Node is here to prove it. Who would have thought 5 years ago that Javascript would provide such great base to build scalable web apps on?Eduardo Ceretohttp://www.cardinalpath.com/noreply@blogger.comtag:blogger.com,1999:blog-1231480044619721857.post-6491229440307133852011-10-14T17:49:38.632-04:002011-10-14T17:49:38.632-04:00Well said.Well said.Anonymousnoreply@blogger.comtag:blogger.com,1999:blog-1231480044619721857.post-14749800402387009232011-10-14T16:45:57.492-04:002011-10-14T16:45:57.492-04:00Well stated. This is the same argument (and answer...Well stated. This is the same argument (and answer) when people try to compare raw performance of one language vs. another. Raw performance, raw complexity, raw anything doesn't work when it comes to comparisons. End results are the key, and the whole IS greater than the sum of the parts.Greg L. Turnquisthttps://www.blogger.com/profile/08171029090321889292noreply@blogger.com