Tweets for 2010-02-20
- @shon_ likely to convert json to object n then use jinja2 in reply to shon_ #
- The entire clojure sequence library is almost immediately intuitive to someone who is comfortable with python especially its itertools #
- Having said that (ie. the last tweet) pythonistas frown on some identical constructs instead preferring the readable alternatives #
- So intuitively I find myself really (as in really really) liking clojure though I shudder at the thought of having to train a team to use it #
- And while I find myself really bored at learning the myriad details of scala, it seems like a better language for java folks 2 migrate 2 #
- facebook's hiphop-php now at GitHub http://ff.im/-gggwQ #
- jQuery 1.4.2 Released. 2x faster than 1.4.1 and 3x than 1.3.2 http://ff.im/-gggGh #
- High Scalability – High Scalability – Twitter’s Plan to Analyze 100 Billion Tweets http://ff.im/-gghcl #
- Gears API Blog: Goodbye Gears, Hello HTML5 http://ff.im/-ggheT #
- Migrating to the Twitter Streaming API: A Primer http://ff.im/-gghoq #
- @jneira but the planet doesn't have enough non code monkeys. And you forget the project manager monkeys as well
in reply to jneira # - A small clojure reference webapp: cow-blog http://github.com/briancarper/cow-blog Older version using RDBMS http://is.gd/8NtwQ #
- A legacy buff who had used the function incorrectly in the first place takes on Rasmus (of PHP fame) http://is.gd/8NIbq #
- re: last tweet Legacy should not be an excuse to stop improvement. Last tweet was via @jamesiry @michaelg @lsegal #
- Wondering: Isn't there far more philosophical consistency between schemaless (#nosql) databases & dynamically type langs than static ? #
- @codemonkeyism I'm looking for reasons and haven't found them. Both treat schemas/types as flexible. in reply to codemonkeyism #
- RT: @NeilRobbins @dnene seesm to me there is Static – Dynamic impedence mismatch when negotaiting between schemaless & static objects #nosql #
- @codemonkeyism Consistency across treatment of objects n data is perhaps relevant (irrespective of what Lispists may opine) (1/2) in reply to codemonkeyism #
- @codemonkeyism If objects in memory also are the objects on disk then I wonder how you can argue objects & data are different (2/2) in reply to codemonkeyism #
- @offbytwo Was referring to the challenge of having a large no. of mixed quality developers trained in thinking n writing prefix notation. in reply to offbytwo #
- @offbytwo I argued in earlier tweets clojure syntax is not consistent with human linguistic patterns so its much harder to train the avg dev in reply to offbytwo #
- @kaleidic Indifference is not so much an issue. Ability to learn enough to create something without blind ^C-^V and maintain it cleanly is in reply to kaleidic #
- @kaleidic Many developers still struggle with db access and web forms issues. I know many who found Procedural infinitely easier than OO in reply to kaleidic #
- @kaleidic We need languages that are easy not sexy – clojure is the latter. in reply to kaleidic #
- @kaleidic Of course I meant sexy as defined by a geek. in reply to kaleidic #
- @kaleidic After due thought I correct myself .. perhaps indifference will drive down effectiveness and could be an important factor. in reply to kaleidic #
- @codemonkeyism Wasn't arguing everything is bits .. My hypothesis flexibility of attribute lists in both types n tables is more consistent in reply to codemonkeyism #
- @codemonkeyism .. (contd from last tweet) as opposed to flexibility in type system and inflexibility in schema or the other way around in reply to codemonkeyism #
- @fogus English n three other Indian langs I know. "noun verb noun" is more natural to "verb noun noun" eg. "I ate eggs" vs. "ate I eggs" in reply to fogus #
- @fogus On a different note, most Indian langs are phonetically scripted; learning to write them takes far less effort than english spellings in reply to fogus #
- @fogus I am not arguing what is more efficient or more precise (clojure is) – simply what is more natural (clojure isn't) in reply to fogus #
- @fogus Don't know japanese perhaps that hypothesis deserves some study:) I was guilty of stating "human" instead of "english + indian langs" in reply to fogus #
- @fogus I am really liking clojure/ I love python, n I generally like to see the better parts of all langs rather than be a lang lawyer (1/2) in reply to fogus #
- @fogus but imo its infinitely (slight exaggeration) easier to teach and comprehend python than clojure /cc @kaledic (2/2) in reply to fogus #
- @fogus here's a post I wrote a little earlier on python contrasting another which was based on clojure http://is.gd/8OvdF in reply to fogus #
- @fogus I like to believe the more popular languages should reach out to the "masses" of "programmers" but not everyone wants 2b more popular in reply to fogus #
- @fogus re:contentious I really like clojure .. its just that I do not believe it reaches out to the masses I referred to. http://is.gd/8OwoE in reply to fogus #
- @psnively Its a different goal .. mathematics and logic do not take precedence over economics .. software is a means not a goal @fogus in reply to psnively #
- +1 RT: @psnively @dnene (This goes a long way towards explaining my impatience with complaints of Scala's complexity, for example.) #
- @jneira I concur .. but I also am interested in the practicality of what makes something easy for the proletariat not bourgeoisie alone in reply to jneira #
- I RT'ed @psnively's tweet without grokking it (sorry:( ) . But scala is at the end of the day less expensive to train in for java folks #
- @psnively in an earlier life I was a little responsible for making them do that.
in reply to psnively # - @psnively +1 .. and that applies to java too (not the type system per se) (which probably made it so popular) in reply to psnively #
- @jneira Convince the lumpenproletariats (especially the lumpen ones) X on their CV is good n they will be the best X fanboys till the next X in reply to jneira #
- @codemonkeyism Not entirely but not independent of. If Types r not about attributes, objects are not data, ORM would've never been required. in reply to codemonkeyism #
- @codemonkeyism Quoting this blog post "I’ve been a long fan of ORMs – especially Hibernate – but I’m no longer" http://is.gd/8OOSU
in reply to codemonkeyism # - @offbytwo I would be more than happy to read up your experiences of training such a team esp. how 50%tile devs adapted to the prefix syntax in reply to offbytwo #
- @codemonkeyism I am uncomfortable with that .. to me storage is persistent objects and object persistence requires storage (1/2) in reply to codemonkeyism #
- @codemonkeyism even if they are not the same, there's probably only some constrained distance separating them, hence consistency (2/2) in reply to codemonkeyism #
- Here's what I further thought about my consistency argument between static/dyn types n schemaless dbs http://is.gd/8OVOY /cc @codemonkeyism #
- Conventional (and my implicit) assumption was data schema is colocated with the data – in the database (driven by RDBMS paradigm of course) #
- But when static typed langs use schemaless storage .. the schema is now encoded in the language and not the db (a point @davetron5000 made) #
- Thus static langs use schemaless storage as actually a fixed schema non relational storage while dyn langs can use it for flexible schema. #
- @psnively +1. But I have been involved in many discussions arguing business targets *cannot* wait on that. So do the best with what we can in reply to psnively #
- @psnively @davetron5000 because business targets cannot wait on dev competence catching up .. we need to make things as easy as we can in reply to psnively #
- @Daneel3001 @NeilRobbins You had a nice discussion My summary tweets on the topic http://is.gd/8OZKO http://is.gd/8OZUa http://is.gd/8OZqn #
- @kaleidic I completely agree .. current team size = 3. But I feel like I escaped the inconvenient challenges leaving them 4 others to solve. in reply to kaleidic #
- @ryansroberts For some apps dynamic schema is really important and for others the non relational structure – its very app domain specific in reply to ryansroberts #