From: "Geir Magnusson Jr." <[EMAIL PROTECTED]> > On 11/2/01 3:13 AM, "Jon Stevens" <[EMAIL PROTECTED]> wrote: > > > on 11/1/01 11:59 PM, "Matt Egyhazy" <[EMAIL PROTECTED]> wrote: > > > >> perhaps sun should make it more clear that petshop is not a benchmark and is > >> instead a multi-faceted example of the possibilities offered by j2ee. i > >> suppose they could rework it and create a benchmark out of it... > >> > >> microsoft is obviously misusing it...and that is possibly what they do best > >> (and are 'doing right'), spread FUD. > >> > >> matt > > > > I don't get it though. Why shouldn't a fully functional demo also be able to > > be used as a real application (or the basis for one)? A pet store shopping > > cart isn't rocket science. Even if it is just an educational science > > project, why does it have to be that much more complex and overly done than > > an equivalent application in another language/system/os. > > > > The point being is that M$ claims that their version of PetStore is that > > much smaller and easier to understand. Well, if their version also has all > > of the same showcase of features, then how come it is still that much > > smaller/simpler/faster? > > Except that (someone noted) that the MSFT version doesn't have all the same > features, so it's not a valid comparison. > > This is an opportunity for us to build a PetStore example w/o a middle tier, > and see how that compares to the .NET version.
Agreed. It could also be a useful comparison to see how simple Java of PetStore, w/o a middle tier compares to the full-monty, multi-tiered EJB PetStore implementation with data access objects, session beans, entity beans, state beans and all those other patterns. i.e. a way for developers to evaluate what the time & performance differences are (both in development, maintenance and runtime) from doing things in a simple way, just at the web tier with a seperation from business logic, persistence & presentation or using all the various EJB technologies & related patterns. Then developers could see how it affects the amount of code, what different deployment topologies are open to them etc. It could help developers decide when EJB is right for them and when its not. I admit I'm an EJB-cynic myself after being burnt on several projects - however it would be good if the EJB-petstore could clearly demonstrate the benefits they offer over using just regular beans in a single-tier petstore. All that extra money we need to pay to the EJB vendors and all that extra code & complexity must have some clearly demonstrable benefits right? :-) James _________________________________________________________ Do You Yahoo!? Get your free @yahoo.com address at http://mail.yahoo.com -- To unsubscribe, e-mail: <mailto:[EMAIL PROTECTED]> For additional commands, e-mail: <mailto:[EMAIL PROTECTED]>
