Hi Jon,
> On your discussion of recreating STATELESS session-beans for each request:
> Is the client of SessionBeans really entitled to know wheather a SessionBean is
>STATELESS or STATEFUL?
> Your strategy for STATELESS gives hard maintenance problems when a SessionBean at a
>later stage of development is required to change from STATELESS to STATEFUL. (This
>could be at a late time at for, example, optimization-time.)
> Or perhaps the design of a STATELESS sessions always requires client-knowledge of
>this property, and that this property never changes in system-lifetime? Some
>real-world experience on this?
>
My immediate and naive assumption is that a client must be aware (to
some extent)
of the nature of the session bean, if for no other reason that error
recovery. If the
bean is "stateless" then server failure and loss of a stateless bean
requires a different recovery model than in the case when a stateful
bean meets an unexpected demise?
- Larry.
begin: vcard
fn: Laurence Cable
n: Cable;Laurence
org: Java Software
adr: Sun Microsystems Inc.,;;(MS: UCUP02-201), 901 San Antonio Road,;Palo Alto;CA;94041;USA
email;internet: [EMAIL PROTECTED]
title: Senior Staff Engineer
tel;work: 408-343-1776
tel;fax: 408-863-3195
x-mozilla-cpt: ;0
x-mozilla-html: TRUE
end: vcard
