Andrew Ho <[EMAIL PROTECTED]> wrote: > > On Sun, 15 Mar 2004, Tim Churches wrote: > > > On Mon, 2004-03-15 at 06:18, Andrew Ho wrote: > > > On Sun, 14 Mar 2004, Elpidio Latorilla wrote: > > > ... > > > > But my original query was based on my wish to have a > > > > "behind-the-scene-app-to-app" dialog. > > > > > > Elpidio, > > > What's the difference between an application talking to OIO vs. > a human > > > user? :-) > > > > Applications (and computers in general) are very stupid, and thus > there > > needs to be a means of spelling out to them what they can and can't > do. > ... > > Tim, > You did not address how a human user is any different?
Sorry, I assumed most of us were familiar with these funny creatures called humans. > I think it is similarly helpful to clearly spell out what an user > can > and can't do through the user interface. Sure, but the mechanisms are different. Typically a combination of natural language (eg 'Enter date of birth here'), implicit visual clues and prompts and other encodings (eg a graphic of a big WARNING sign) are used. All of these are gibberish as far as the computer is concerned, unless you teach it to recognise the specific gibberish - but that is fragile and leads to a high maintenance overhead. Web service protocols and interfaces aim to make as much information about the Web service as explicit as possible using standardardise metadata structures. That's the difference. Both are needed, but they address different audiences. Tim C
