Very Computer

Greetings,

Somebody could tell me the reason of BOA's Death?
Does exist a BOA vs POA comparison page on the web?

                                        Thanks. Valentn.

Hi',

I understood that this is because the BOA was underspecified.
Consequently the ORB vendors made extensions to the BOA,
which in turn made the server-side code not cross-compatible between
various ORBs.

Therefore the OMG removed the Basic Object Adapter from the spec (in
version 2.2 of the spec I believe) and replaced it by Portable Object
Adapter. The POA, as its name suggests, allows for writing portable
server side code. The new name also allowed ORB vendors to support both
adapters for while, which is a Good Thing[TM].

I however sometimes tend to think that I may miss the BOA, if the
vendors follows the spec and remove BOA support from their products at
some point in time. What I have in mind are mainly the activation modes,
which were defined at the BOA. Yes, I know, that was one of (perhaps the
main?) source of the incompatibilies, but they were there, if one wanted
to use them. ORB specific, yes, but guaranteed. I also know that with
POA I usually am able to specify the activation modes at the IMR level,
as Steve and Michi point out in their superb book, but that's 'usually'.

I'm not a guru and do admit that what I have stated above might be a
rather loosened version of the real story, but still have the feeling
that with POA I do not have guaranteed that the ORB actually will
support activation modes, somehow.

Does anyone have an opinion on that?

But, please do not just tell me things like 'you *have* to use the POA,
it's the *standard* today'. I already know that.

regards,

michal++

Quote:> Greetings,

> Somebody could tell me the reason of BOA's Death?
> Does exist a BOA vs POA comparison page on the web?

>                                         Thanks. Valentn.

Sent via Deja.com http://www.deja.com/
Before you buy.

 The ORB Portability Enhancement RFP, the document that called for a
replacement of the BOA, lists a number of reasons. Last time I've looked,
the document was at
  ftp://ftp.omg.org/pub/docs/1995/95-06-26.ps

 You might also be interested in reading part of my thesis,
  http://www.vsb.informatik.uni-frankfurt.de/~frankp/diplom/node15.html

        Frank

--

 |                                      http://www.uni-frankfurt.de/~fp/ |
 +---- Life would be a very great deal less weird without you.  - DA ----+

 There now exists no `standard' BOA anymore. If some ORBs continue to
implement an object adapter with that name, it's entirely vendor-speci-
fic. However, that's not much of a difference to the previous state.

Quote:

> I however sometimes tend to think that I may miss the BOA,
>                     What I have in mind are mainly the activation modes,
> which were defined at the BOA.

 The BOA specs left the Implementation Repository, and any activity on it,
entirely unspecified. That's why the POA does not address an IMR -- for the
same reasons as above. Now that no IMR is mandated, vendors are left to
themselves, and can implement vendor-specific means of starting up servers
etc.
 Not addressing a feature in the POA specs does not mean it's not there, it
just means that it's vendor-specific. As the spec itself puts it, "Adminis-
trative action beyond the scope of this specification may be necessary to
inform the ORB's location service [i.e. IMR] of the creation and eventual
termination of existence of this [persistent] POA, and optionally to arrange
for on-demand activation of a process implementing this POA."

        Frank

--

 |                                      http://www.uni-frankfurt.de/~fp/ |
 +---- Life would be a very great deal less weird without you.  - DA ----+