Re: : Ada Utility Library 1.4.0 is available

[Georg Bauhaus <rm.dash-bauhaus@futureapps.de>]

On 16.01.12 09:10, AdaMagica wrote:
> Hm, I looked at your beans, and I do not see the value of this.

TTBOMK, beans at the most basic level will usually exist so that
there is a way for some kind of configurable framework to automatically
write ("serialize") plain old Ada objects to external storage, and restore
them as Ada objects, as needed. (There is more, like transactions,
or events sent to beans.)  The external storage typically
provides for storing conventional database types; the storage can
be configured without touching the programs, and can be anything
that meets some simple requirements, not just RDMSs. Could be
JSON, or XML, too.  JSON is particularly hip if you want
communication with Google terminals, such as Android devices
or browser based virtual machines, JSON being  a Javascript
format (assoc lists).

The easiest way to map plain old objects to external storage
is via symbolic names understood at both the Ada end
and at the data storage end. The resulting objects can be
used by programs written in other languages, too.
As long as the data items are complete (no references, say),
the data storage needs only know, for each data item to be handled,
a pair consisting of the name of data item, and the predefined
type of item, perhaps in some data nest fashion. Hence, I think,

   subtype Polytype is Util.Beans.Objects.Object;
     -- for illustrating the multi-type nature

   overriding
   function Get_Value (From : Compute_Bean;
                       Name : String) return Polytype;

together with the definition of a mapping between Polytype,
names, and program types.


<Side-note related="somewhat">Rumor has it that not all projects
are entirely happy after having moved away from plain old
relational SQL (such as JDBC) to object storage (such as Hibernate).</>