ajy@cel.co.uk (andrew yeomans) (02/21/91)
It seems to me that the debate on which is better, Motif / OpenLook / Windows / Apple / Presentation Manager / etc. is sadly too similar to the flames on which programming language is better, when the user wanted an application, not a language. (flame half on) I believe that the current window system API's are FAR too complex for effective use, at least by the majority of programmers. For example, my copy of the Motif Programmers' Reference has almost 1000 pages - compare this with the 272 pages of the K+R C Programming Language (2nd ed), which includes both tutorial and reference. It seems to me that we should instead be debating which Interface Builder we should be using, as these offer the chance of: a) Simplifying use of the window system, so that most of the available functionality is easy to understand and use; b) Building the same user interface to run with either Motif or OpenLook - without preventing use of 'special' functions in each. (flame off) I'd be interested on hearing opinions as to whether use of (graphical) Interface Builders is the 'right' way to go in the Motif/OpenLook debate. -- Andrew Yeomans | UUCP: ajy@cel.uucp or ..!{ukc,mcsun,uunet}!cel!ajy Crosfield Electronics Ltd | PSTN: +44 442 230000 X 3371 Fax: +44 442 232301 Three Cherry Trees Lane | These opinions are MINE, all MINE! Hemel Hempstead | Ps 66: "Thou broughtest us into the net; Herts, HP2 7RH, England | thou laidst affliction upon our loins."
david@jpl-devvax.JPL.NASA.GOV (David E. Smyth) (02/22/91)
ajy@cel.co.uk (andrew yeomans) writes: > >I believe that the current window system API's are FAR too complex for >effective use ... > >It seems to me that we should instead be debating which Interface Builder we >should be using, as these offer the chance of: > > a) Simplifying use of the window system, so that most of the available > functionality is easy to understand and use; > b) Building the same user interface to run with either Motif or > OpenLook - without preventing use of 'special' functions in each. OK, class, guess what I'm going to recommend ;^) Use Wcl. Then the documents your programmers will mostly have to use will be reduced to one: the widget set man pages. And the way they need to learn about widget programming is nicer : learn about the widgets, then Xt, then X11 (many will never need to know anything beyond the widgets - they can use Wcl provided callbacks as examples, and just edit, instead of compose - a much easier job). And your code will be pretty darn free of widget set dependencies. In an application I've been developing, out of 10027 lines of C, include files, and resource files, there are 287 references to Motif (the widget set I'm using). That seems like a pretty widget set independent application to me. And what's better, the count of lines dependent upon Motif is actually decreasing over time! ------------------------------------------------------------------------- David Smyth david@jpl-devvax.jpl.nasa.gov Senior Software Engineer, seismo!cit-vax!jpl-devvax!david X and Object Guru. (818)393-0983 Jet Propulsion Lab, M/S 230-103, 4800 Oak Grove Drive, Pasadena, CA 91109 ------------------------------------------------------------------------- "There was a time when man would attribute everything he didn't understand to God. Now, with our heightened understanding of science, we attribute them to UFO's." -Gary Friedman -------------------------------------------------------------------------
jeenglis@alcor.usc.edu (A mutable language) (02/22/91)
david@jpl-devvax.JPL.NASA.GOV (David E. Smyth) writes: >ajy@cel.co.uk (andrew yeomans) writes: >> >>I believe that the current window system API's are FAR too complex for >>effective use ... >> >>It seems to me that we should instead be debating which Interface Builder we >>should be using, as these offer the chance of: >> >> a) Simplifying use of the window system, so that most of the available >> functionality is easy to understand and use; >> b) Building the same user interface to run with either Motif or >> OpenLook - without preventing use of 'special' functions in each. >OK, class, guess what I'm going to recommend ;^) >Use Wcl. Then the documents your programmers will mostly have to use >will be reduced to one: the widget set man pages. And the way they >need to learn about widget programming is nicer : learn about the >widgets, then Xt, then X11 (many will never need to know anything >beyond the widgets - they can use Wcl provided callbacks as examples, >and just edit, instead of compose - a much easier job). I don't know about this... Personally, I couldn't understand the Athena widgets until I had read all about Xt, and much of Xt didn't make any sense until I had read more about Xlib and the X protocol. If I hadn't known a little bit about Xlib/X to start with, probably NONE of Xt would have made any sense at first. The learning curve for X programming is outrageously steep no matter how you go about learning it. Tools like Wcl make programming easier, but there's still a lot to learn before you can start writing code. Not that I'm knocking Wcl -- it's definitely a step in the right direction. -- But the fish won't be looking back at me and whimpering. -- St. Terri jeenglis@alcor.usc.edu // ignore this posting