I'm launching into a project to move a large set of applications from working against a vfp back end to working against a sql server back end.. and have been looking at cursor adapters as a possible step in that direction.

There are a lot of things I like about CA's: they combine some of the advantages of remote views and SPT. I like the consistent interface to vfp, sql server, and xml data, the fact that I don't need to have dbc's for remote views, the fact that they can call stored procedures, their performance (which seems comparable to SPT)

On the other hand, they are a pain to construct.. even aside from the 255 character property limitation.
We have a LOT of tables and remote views.
I've used Mark McCasland's CA builder to easilty create CA objects for the tables..
But creating them for views is a pain.. and there doesn't seem to be any simple way to adapt Mark's code to handle remote views (which are usually parametrized and sitting in a dbc separate from the tables themselves)..

The CA bulder that comes with VFP seems useless for anything but the most straightforward views.

Thus for example, trying to move even a simple view such as this one to a CA via the builder doesn't work:

SELECT Sh.shshipno, Sh.shshipweek, Sh.dypk, Sh.shpk, Sh.sgpk, Sh.destpk,;
Sh.capk, Sh.ohpk, Sh.shpo, Sh.shcarrierp, Sh.shmethodp, Sh.typestatus, Sh.pspk,;
Sh.whpk, Wh.whdescript, Sh.shfreight, Sh.shservice, Sh.sifobcode,;
Sh.sistatus, Sh.invoicenr,;
Ty.tydescript AS cfobcode, Sh.adduser, Sh.adddtime, Sh.edtdtime,;
Sh.edtuser, Gr.grdescript AS destination,
Sh.recstatus ;
FROM ;
ShipmentHeader Sh ;
INNER JOIN Warehouse Wh ;
ON Wh.whpk = Sh.whpk ;
INNER JOIN Type Ty ;
ON Ty.typk = Sh.sifobcode ;
INNER JOIN GeoRegion Gr ;
ON Gr.grpk = Sh.destpk;
WHERE Sh.ohpk = ( ?tOhPk );
ORDER BY Sh.shshipno

The builder seems to assume that join clauses will always be on fields with the same name?

I would love to see a program such as Mark's which would also handle parameterized remote views..

Am I missing something here? What do I really gain by going the CA route rather than sticking with SPT? Especially when going the CA route involves all this hassle (note that because of the 255 limit, I can't just simply populate the schema, update field list, etc properties.. but would need to do it programatically..)

Ilmar