Gregory wrote: >--- Jim Self <[EMAIL PROTECTED]> wrote: > >> Kevin, I think you greatly underestimate the difficulty of providing >> a general search >> functionality that is easy to use and fast enough to be useful on >> large and complex "live" >> data bases. Then again, it is said that an advantage of youth is in >> not knowing that a >> thing is impossible until after you have done it. ;) >> >> The difficulty is in satisfying all three objectives at the same time >> (or up to seven >> depending on how you count [general, easy, fast, useful, large, >> complex, live]). > >I don't know that prospects for building a general search feature are >THAT bleak, but it's certainly a non-trivial problem (consider how much >has been written on query optimization). I also agree that a brute >force solution is not good enough.
I didn't mean to say that the prospects are bleak,... just not trivial as Kevin seemed to be saying. Actually, I think the prospects are quite good for building such a general search feature on the query tools in M2Web. >> Brute >> force searches can be painfully slow, so I think that the general >> philosophy is (has been) >> that routine queries should always be based on cross references for >> ease of use, >> efficiency, and speed. > >Even writing a query mechanism that could make use of existing >cross-references would be a big improvement. Your wording here suggests that I wasn't very clear in the latter part of my previous message. That is exactly the function of the tool that I have been calling "iterator". >But there is more that >could be done: rewriting queries, lazy building of cross-reference data >for frequently used fields, caching of intermediate values, etc. But >this won't work if people use direct global sets to update the files >because updates need to be able to invalidate the cache. I think it is important to consider why there is this recurrent complaint about programmers avoiding the use of API's and what can be done about it constructively. The question that started this thread of discussion pointed out some flaws or gaps in the current set of VistA API's that essentially force programmers to work at the low level of fiddling with the pieces of global nodes instead of working with mnemonic or numeric field identifiers. --------------------------------------- Jim Self Systems Architect, Lead Developer VMTH Computer Services, UC Davis (http://www.vmth.ucdavis.edu/us/jaself) (http://vista.vmth.ucdavis.edu/) (http://openforum.worldvista.org/~forum/m.cgi) ------------------------------------------------------- Using Tomcat but need to do more? Need to support web services, security? Get stuff done quickly with pre-integrated technology to make your job easier Download IBM WebSphere Application Server v.1.0.1 based on Apache Geronimo http://sel.as-us.falkag.net/sel?cmd=lnk&kid=120709&bid=263057&dat=121642 _______________________________________________ Hardhats-members mailing list Hardhats-members@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/hardhats-members