If using php then something along the lines of ODBC or JDBC yes.
WebSpeed is obviously the best way, but, depends which version you're on. In OE 10.0 the WebSpeed server costs tend to get much higher (even though Progress constantly deny this - try getting the V10 for the same cost as the V9 one - they'll talk about flexibility of ownership models - blah blah blah ..., but, when it comes down to the final quote it's a nightmare).
Problem with php/ODBC route (which I admit although I've used php have never ODBC'd into progress) is the Progress license model will probably dictate that for each user who can connect you need a license - and this is difficult in that you don't always know how many users are going to use it and it can also push your costs up.
Main problem with Progress over the web is that they started off with a nice licence model and then paniced as they were worried that sites were porting applications over web (rather than, say, GUI) to save costs.
I feel that once again Progress are shooting themselves in the foot and again forcing the 'grass roots' Progress community to look at other alternatives. If I was writing a stand-alone web application I'd look at PHP (or .net asp if I had the skills) with a different database engine to Progress, e.g. My Sql or SQL Server. Most web applications are quite light so they don't require the power of Progress. If you need the Web Application(s) to interface into your main databases then how about collecting the data then having some sort of batch job (one license) which reads the data from the light weight database to Progress. The same batch job could write data into the light one, i.e. for lookups, etc.
Hope this makes sense.
P.S. we run v9 and have sued WebSpeed for about 3 years. We now face the problems of wanting to go to OE10, but, facing significant costs for upgrading WebSpeed. Cheers Progress - they still don't get it do they - grrrrrrrr !! Yes guys, another bad hair day !!