Ada Component Registry proposal

["Robert I. Eachus" <rieachus@comcast.net>]

  I  put together an XML DOCTYPE for a Ada Component Registry, and tried 
it out on GNADE. If a couple of people who understand XML would like to 
look at it and criticize, I'll send you a copy. But right now I consider 
it too preliminary to let loose in the wild. I'll probably have a much 
better worked out version by the end of next week. (Oops! Mental note to 
self,  the registry itself needs a version number. ;-)

I picked GNADE as a test case though, expecting it to be a decent stress 
 test. Boy was it! For me as well as the XML grammer. I'd like to 
comment on a few things that I found and what they mean for a formal 
registry, or for the Common Ada Library.

One of the things that the registry does is makes the zip or tar or 
 whatever file that you can download more accessable. Much more 
accessable. Amazingly more accessable. GNADE as delivered has over a 
 dozen directories and if you actually compile it, creates a few more. 
(And just to confuse you puts some stuff into existing directories 
outside the GNADE structure.) With the registry, you can look and see 
that most of the source files are children of GNU.DB, and that package 
declaration can be found in ../gnade-src-1.4.3a/support. The problem 
though can be found in that word most.  There is also a separate OCI  
root for an Oracle binding, and over a dozen packages and programs at 
the library level that can conflict with other bindings.  These include 
such wonderful names as Parser, Scanner, SQL, ODBC,  and Tools.

Am I saying that GNADE is horribly organized?  No.  Just that GNADE is a 
excellent example of why a registry is needed.  If there were a standard 
place for a database interface packages--I would prefer 
Interfaces.Xxxxx, but that is a detail--then GNADE could provide some of 
those packages and mix and match with implementations of such packages 
from elsewhere.  And GNU.DB.Support would disappear as GNADE moved to 
using more standard implementations of tables and lists.  Finally, the 
scanner and parser that GNADE provides to support embedded SQL could be 
reused by other projects that have nothing to do with databases.

But all of this has to be understood in terms of encouragement and 
process.  GNADE is not going to adopt major changes just because I say 
they should.  But the existance of a useable registry and "standard" 
Common Ada Library would encourage the GNADE authors to evolve their 
code in directions that will make it easier to reuse.

-- 

                                                    Robert I. Eachus

"Quality is the Buddha. Quality is scientific reality. Quality is the goal of Art. It remains to work these concepts into a practical, down-to-earth context, and for this there is nothing more practical or down-to-earth than what I have been talking about all along...the repair of an old motorcycle."  -- from Zen and the Art of Motorcycle Maintenance by Robert Pirsig