Hi, again, Pete. > > on a practical level, seriously needs further research. In the ideal > > world, of course, all SQL would port gracefully between database > Nice but not always possible. For instance in some databases like Access (or > is that a pseudo-database ;-] ) you need to quote datetime values with #
<grin> Access is what you do when you get information from a real database. > > If you really need to do something different, and you want to abstract > > it so that it's all in one place, you can wrap it in a function > Only viable if your application is generating the IDs. This is not always the > case. Many (most?) databases have some sort of "generator", "sequence" or > "unique" column. In this case you want the database to manage the key With Oracle, you can do a SELECT SEQCOLUMN.NEXTVAL FROM DUAL; store it in a variable and then use it at whim. I'd have thought most databases provide a similar mechanism to generate a unique key. Definitely should be a standard, since it's just so common. Alas! I digress. > Persistence layers, OR tools and the like. Turbine Torques an example of a > relatively good layer - nicest to use I have found in opensource world. Then > you have more highend solutions like EJBs entity beans. Staying away from particular architecture ties is one of the goals for BOX. EJBs would be out of the question; but Turbine Torques I'm certain is worth a deep reading. > The person everyone thinks highly of in persistence world is "Scot Ambler" - > never read his work but it would be worth looking at his stuff. Noted; thank you. > impossible to say without doing direct comparisons with comparable complexity > etc of pages. However it sounds like similar performance to Cocoon1 based > apps (Do you use DOM?). I use DOM. Would I get better performance out of SAX? I'd gain performance using SAXON instead of Xalan. I actually tried plugging SAXON into the architecture, but it blew up when I tried to create a TreeWalker, or some such useful object. I may have been mistaken about the hits/sec. I wrote a small shell script which was basically: while true; do wget -q http://server/path/Servlet > content.html; done I launched this script as ten processes (in a Cygwin shell, over the 100MBit network) and then browsed the site manually. I also launched the script thirty times, but at such a point I don't think it'd make a bit of difference. So the 10 hits/sec. in my previous message was probably significantly lower than actuality. Thanks for the tips, Dave Jarvis -- To unsubscribe, e-mail: <mailto:[EMAIL PROTECTED]> For additional commands, e-mail: <mailto:[EMAIL PROTECTED]>