Re: jndi database connection pool to slow ?
gert wrote:
So here i [sic] was selling my self to a j2ee [sic] consulting company hoping they
would hire me. And while i [sic] was lucky to hit the jackpot, meaning
getting a chance to talk to some one that actually knows j2ee, i [sic]
starting to discus about why ejb 2.x struts [sic] jsf [sic] or any non pure xhtml [sic]
css [sic] javascript [sic] (ajax [sic]) solution was a bad idea and that i [sic] really like
sql [sic] more then using a orm [sic] layer. You can imagine this was starting to
look more like a battlefield then a job interview :)
I agree with some of what you say, with reservations.
I could defend my self on all grounds except for two things
1) everybody is using struts and hibernate
Not everybody, and not always well, and never without difficulty.
The trouble is that people use these tools *without understanding the
paradigm* behind them. For Struts, it's the Model-View-Controller (MVC)
pattern. I've worked on teams where people knew the Struts API cold but
didn't understand the MVC architecture. Sad.
ORM layers are for people ign'ant of SQL. The new stuff, OTOH, with the Java
Persistence API (JPA) annotations, is actually useful. Try writing a good
data-access layer yourself; it's not trivial.
2) he said making a jndi [sic] connecting pool was slow.
Can he prove that? Has he ever measured it?
Somehow I doubt it.
Can any one explain to me why jndi [sic]/jdbc [sic] is slow and whats the
alternative ?
I'd be curious to know *if* it's "slow", what "slow" means in that context,
and compared to what. Somehow I doubt it.
It is not always good interview technique to inform the interviewer that he
has his head up his ass, particularly if you prove it.
--
Lew