Markku.Savela@tel.vtt.fi (Markku Savela) (02/20/90)
In article <TQV1CHCxds13@ficc.uu.net> peter@ficc.uu.net (Peter da Silva) writes: >Is there any way other than NeWS, then, to move the UI where it belongs... >in the display the user is working on? [ NOTE: This is posted to comp.windows.misc and comp.mail.multi-media. I really don't know in which group this kind of discussions should be placed, so there is no followup-line in header... ] All this talk about user interface seems to assume that the application is on the same machine (or behind a moderately fast connections) and controls directly the layout of the screen. What if an application is a server residing on a network (like Internet) and is meant to be accessed from all over the globe from all kinds of environments? Currently the server has no other option than to use simple terminal protocol. If it wants to give graphical information, proprietary and other special protocols are used--tough luck for all others who would have the physical capability to show the graphics but don't have the right protocol. [ What non-text services are currently existing, btw?] The service could support OpenLook or Motif interface directly, but doing both of them would be rather heavy on the server. Besides, there will always be other user interfaces. I say the service shouldn't care about details of user interface, like presentation. It should only concentrate on providing the service and let the user decide his/her user interface. In terms of P1201 I see the division of tasks between server and user environment as follows: 6) Application 5) Dialogue 4) Presentation .... Server <--- communication ---> User Program Agent (for each - OpenLook | interface) - Motif | USER - vt100 ;) | - ... | <- wide area -> <- local/fast net -> Now comes the hard question: is it possible to define an abstract information exchange protocol at Dialogue level that would fill the need of all or most imaginable services? What are the concepts this protocol operates on? *If* this was possible, it would have the following consequences: - service need to support only one interaction style with clients, regardless of their local requirements - each user environment need only one User Agent program for all services it needs to access (replacement for the terminal emulation--instead of activating xterm, user activates xIR or something ;) There are requirements for the Dialogue level protocol: - it must be able to pass text and graphics - it must be extensible for new data types without breaking existing "xIR" programs (old versions would just treat unknown information types as byte streams--still capable of passing them on without knowing what it was..) - information in processable form is important, most of the information is cut/pasted to the other applications anyway, edited and reformatted. (unrestricted, raw POSTSCRIPT will not fill this requirement) There are many services that could profit from this universal access method. Converting the current text based services into this format would enable them to move gracefully into multi-media applications. The current trend seems to be to design application specific pairs (server,user agent), each being fixed to some specific user interface. This leads to multitude of access programs and to a chaos--I want a single program to access most of the information (like xterm, I activate an instance for each access). And I wan't that program to have the user interface I like, other people may use different interface, but the same service (e.g. there would be different user agent for each environment, but one user would usually use only his/her favorite version). Some examples of services - internet name services (either DNS or X.500) - mailbox service (X.400 has this P7, that would probably be obsolete after this? ;) - traditional information services (instead of text only, I could finally get full documents with pictures, later even animations and what not... ;) - real time conferencing services, and other group communication services. In short: I wan't a replacement for the old terminal emulation straight jacket. I want to break out. I'm looking for the universal information access tool for the next millenium... :-) -- Markku Savela | savela@tel.vtt.fi Technical Research Centre of Finland | Telecommunications Laboratory | Markku.Savela@vtt.fi Otakaari 7 B, SF-02150 ESPOO, Finland | savela%vtttel@router.funet.fi
Markku.Savela@tel.vtt.fi (Markku Savela) (02/20/90)
In article <4538@hemuli.tik.vtt.fi> Markku.Savela@tel.vtt.fi (Markku Savela) writes: > so there is no followup-line in header... ] Sigh, just shows I don't know all about posting. There appears to be a "Followup-To: comp.windows.misc" in headers after I posted it. So, be warned... -- Markku Savela | savela@tel.vtt.fi Technical Research Centre of Finland | Telecommunications Laboratory | Markku.Savela@vtt.fi Otakaari 7 B, SF-02150 ESPOO, Finland | savela%vtttel@router.funet.fi
sac@Apple.COM (Steve Cisler) (02/21/90)
Savela says: I'm looking for the universal information access tool for the next millenium... the NISO Z39.50 protocol might be a good place to start. I think this American protocol is being discussed in ISO meetings now. It was written by the library community and is being experimented with by a number of commercial firms and academic library systems. Steve Cisler Apple Library sac@apple.com .