I'm jumping the gun a little bit here, in that Ty is still (supposedly) reviewing the documentation and procedures for the A-Shell/SQL Version 1.2 update (which supports ODBC), but I thought I would give a heads-up to those who have been waiting for some way to connect to databases other than MySQL. Besides giving you a chance to get started downloading/installing or just thinking about your accessing your database of choice, I'm also looking for some more feedback about preferences.

When we started this, we were mainly thinking about "real" database engines - MySQL (which some purists might not even consider "real"), SQL Server, Oracle, DB2. And, we imagined separate, native connectors for each one, with SQL Server being the next in line after MySQL. But it turns out that Microsoft has deprecated the native TDS interface and advises developers to stick with ODBC or one of its wrappers (such as OLEDB or ADO). So that's what I did, and so we now (or will soon) have an ODBC connector between A-Shell and SQL Server.

However, once you can talk to one ODBC connector, you can (at least theoretically) talk to most or even all of them. Which means that we are potentially opening the door to connecting to everything from DB2 to Excel via this same connector. Is there a problem with that? Actually yes.

The problem is that the ODBC standard, in trying to be all things to all databases, applications, and operating systems, allows for such a wide range of degrees of conformance that really making it work for a particular OS/database combination is almost like developing a new interface. That may be a bit of an exaggeration since the really common stuff, like SELECT * FROM , will probably work similarly across all cases, if we aren't too picky about the precise details of the returned column attributes. But given that the A-Shell developer crowd is willing to shed gallons of bloody ink over issues like millicolumns and the proper use of the Control+Home key, I can easily imagine an extended period of chaotic struggles as we sort out the lines of responsibility between A-Shell, your application, the data source manager, the drivers, and the databases themselves.

Which leads to the question motivating the long-winded intro:

Given the wide range of possibilities presented by ODBC, what do A-Shell developers want or expect to do with it? Is there any consensus about which target database(s) we should focus on first? Is is just A-Shell/Windows developers who care about ODBC, or do Linux/AIX developers also want to use it (perhaps to get at SQL Server from UNIX, or to get at Oracle or DB2)?

While on the subject of ODBC/UNIX, I'm curious if anyone has any experience there. ODBC was essentially invented by Microsoft but not surprisingly they didn't extend it beyond Windows, leaving it to third parties to do so. The main UNIX implementation is unixODBC , which is free, but the connectors generally are not. For example, if you want to connect to SQL Server from UNIX (and a number of people have suggested to me that this is exactly what they want to do), the least expensive connector I've found is from EasySoft at around $1000. There is another one from DataDirect Technologies which appears to start at around $4000. For obvious reasons we're not going to go out and buy one of each just to test them and tweak our interface to work well with them, just on the chance that someone might want to go that route. But we may have to, in order to properly support it, if there really is such a demand. However, if our past experience with the ODBC drivers for ISAM is any indication, the degree of interest withers away pretty quickly when people encounter driver/connector prices upwards of $1000. Which makes me wonder whether the majority of interest in the ODBC connector ends up being for connecting to something very inexpensive, like Access.