Re: Engine Performance Requirements - Help !

In article <473moo$9hj@cssun.mathcs.emory.edu> rob@dssmktg.com (Robert Minter) writes: >From: rob@dssmktg.com (Robert Minter) >Subject: Re: Engine Performance Requirements - Help ! >Date: 30 Oct 1995 18:23:04 -0500 >In article <ian.goddard.31.0000C09B@geo2.poptel.org.uk> > ian.goddard@geo2.poptel.org.uk "Ian Goddard" writes: >> > "tech@compsys.demon.co.uk" <158.152.33.120> writes: >> > We are currently developing an Informix 4GL application for one of >> > our customers. >> >> > The customer wants data retrieval time of < 1 sec from queries >> >> Informix will wait for some time after you press ESC to enter your query. >> This is done in order to distinguish a single ESC from an escape sequence such >> as the ones the arrow keys generate. This uses up most of your sec. :-( >If you can, change the default ACCEPT KEY to something else other than the >default ESCAPE key. This should speed up the time between hitting the key >and starting the actual query. The point I'm making is that if you accept a user spec that requires sub-second response you need to think about the possibility that ESCAPE processing will eat up that specified response time before the request gets passed to the engine. Changing ACCEPT to something else is one way of doing it. But if the solution mix includes Perform you could be in trouble; I don't recall being able to change ACCEPT on Perform. It may also cause trouble if your users have been brought up to use ESCAPE. Personally I would try to re-write the spec to specify "sub-second response by the engine to queries which return < x rows". Ian