Thursday, October 12, 2006

Why the Java Ecosystem is so much fun after all

I was reading these two threads on InfoQ and I was thinking that Java has had such a big success because starting from James Gosling back the old days and "ending" nowadays with Jonathan Scwartz (SIC!) the Java platform has benefited from several (lots) of the brightest minds of all times. Of course I use Jonathan more to indicate a point in time than a brilliant mind (I'm just joking, it's MY blog after all !).

No what made my day is this:

On one side we have John Crupi (of SUN) evangelizing SOA, and the whole WS* thing, telling us how they basically define, design and publish enterprise services having XML Schema as a fist class citizen: "...whether it's one mega-schema that has all your use-cases and you have to pull out the parts that have been populated or the use-cases are finer grained associated with the schemas at a lower level" etc. Here's the link.

On the other hand, up one slot (on InfoQ) we have Tim Bray (of SUN) who obviously doesn't like what vendors currently do, or trying to do with all that SOA hype, and parts of all this WS* initiative: "...and they are based on XML schema and WSDL, which are two of the ugliest, most broken and irritating specifications in the history of the universe."

So what I did is I had a really good laugh, but after all, these guys are both for real. Obviously they are deeply involved professionals, you cannot laugh at one and disconsider the other. After watching both the interviews, it's sooo easy to make your choice and choose the path you want to go.

So this "being able to choose which way to go", which at a lower level in Java manifests itself by being able to choose between technology and implementations, is available at the concept level also as proven by the difference of vision between these two guys. That's one thing where the Java ecosystem got it right. Being able to choose. But there are lots of other reasons for the success of Java of course. Mhhhm. Mmm.

Which way do you go ? :-D

As a side note, it seems that InfoQ is starting to deliver isn't it ? :-) . Well, that's happening since some time now, what Iwant to say is that now it's becoming more and more obvious. Congrats Floyd and team, and keep up the good work.

Wednesday, October 11, 2006

Gregory Benford sau Isaac Asimov ? Amandoi ?

Am terminat de citit "Teama Fundatiei", scrisa de Gregory Benford si autorizata de patrimoniul Isaac Asimov. Volumul face parte din "A doua trilogie a Fundatiei", prima (trilogie) fiind scrisa de de Asimov.

Asimov mi s-a parut mai digerabil, mai usor de citit. Benford este mai greu de citit, dar cu atat mai mare este satisfactia cand actiunea incepe sa se lege, si explorezi, pe masura posibilitatilor intelectuale, un univers creat de o gandire mult mai ampla, mai complexa, neingradita de bariere conventionale.

De fiecare data cand termin un volum SF ma intreb cum a fost posibil ca un om sa creeze un univers, sa il descrie in detaliu, sa descrie mecanismele functionarii lui si interactiunile intre indivizi in asa fel incat sa rezulte o capodopera a genului (unde e cazul bine-nteles). Ce fel de pregatire, ce gandire si ce cumul de informatii si imaginatie trebuie sa posede un om ca sa te incante cu o lectura de genul "Fundatia" (Isaac Asimov) sau "Dune" (Frank Herbert) ? In acelasi timp nu ma pot opri sa nu observ esecul calculatorului (puneti aici orice joc in mod text sau grafic piratat sau cumparat cu care v-ati pierdut vreodata timpul :-D ) in crearea unei lumi virtuale de complexitatea celor care sunt posibile prin lectura unui SF (probabil si alte genuri) de calitate. Sa fie oare o problema senzoriala ? (text+lume interioara VS surogat de lume virtuala externa, perceputa in cea mai mare parte vizual ?).

Tuesday, October 10, 2006

O(pen) S(ource) JTA Transaction Managers

Maybe not lots of you needed things like XA transactions, two phase comit, and won't need it soon but anyway if you need transactions that span more than one data source (2 databases for example) or transactions that span a database and a JMS destination (pretty common usecase), you may want to look at the following:
If anyone knows some more, post them as replies, thanks.

All of these are usable/configurable outside a full blown J2EE container (I'm not sure about JBOSS Transactions though, anyway it's highly probable that it can be done).

All of these are supported and can be easily integrated with the Spring framework, if you don't want to use a J2EE container.

For more on Spring, Jencks and Geronimo Transaction Manager see here.
For more on Spring and JOTM see here.
For more on Spring and Atomikos see here.
For more on Spring and JBOSS see here.

For more on Spring Transaction Abstraction see here. It's not related to JTA entirely but it will give you some usefull info about transactions in general.

Thanks to Guy for replying and pointing me where I was wrong about Jencks.