adam@bsw.UUCP (Adam de Boor) (06/16/89)
Is it just my warped perceptions, or is there a noticeable lack of understanding as far as the power of the "translations" resource goes? In each of the applications I've looked at closely (well, both of them: xpostit and xman), I find things have been hard-coded in that could have been much better done via the translation manager. xpostit, while using a widget for its 3M plaid icon, has a single action to which it binds all button events. The action just calls a callback that says "left button? raise all notes; middle button? lower them all; right button? bring up a menu". There are only three sizes of note you can create, although you can change the size with the window manager (a cumbersome way to provide functionality, I must say). A much better way to perform the task at hand is for the Plaid widget to provide the actions RaiseAll, LowerAll, SaveAll, and MakeNote (which takes two parameters: the height and width in inches). This allows me to bind "Shift<Btn1Down>" to make myself a 1x1.5 inch note (not one provided by the menu, but just about the right size for me), or "<Key>s" to "SaveAll", rather than having to bring up the menu. xman, on the other hand, has a particularly annoying "feature" where it brings up a menu whenever the mouse crosses its title bar. I had hoped to be able to just rebind that to "<BtnDown>: MenuPopup(menu)" (that's another thing lacking in xpostit -- the menu is brought up by hand rather than with a MenuPopup action, so there's no way to rebind the event used to bring the thing up), but no. It doesn't even pretend to be flexible (as xpostit does), it just calls XtAddEventHandler. I don't mean this to be a personal attack on the authors of these programs -- xpostit is quite a nice little program now that I've been able to customize it and fix the saving of the note positions so they come up where I so carefully placed them before logging out. Rather, this is a plea for people to recognize the power of the translation and action mechanism and the flexibility it provides. It gave me a warm fuzzy feeling when I was able to change all scrollbars to have what I consider the "right" button bindings (left goes up, right goes down) in all Xaw-using clients by adding three lines to my resources file. Please, folks, let's not make applications overly restrictive, shall we? a
rws@EXPO.LCS.MIT.EDU (06/16/89)
xman, on the other hand, has a particularly annoying "feature" xman for R4 has been almost completely overhauled.
swick@ATHENA.MIT.EDU (Ralph R. Swick) (06/16/89)
Your observations and plea are certainly welcomed here. We couldn't agree more! My only real comment is in defense of at least one of the two applications you cite, and others of their generation; it's only a (mere :-) matter of time. Chris had the less than pleasant experience of developing xman concurrently with the development of Xt and Xaw (pre R1 days, folks). For some odd reason, getting it to work and keeping it working under each baselevel of the tools got more attention than did making its implementation a perfect coding example. You'll be thrilled with the cleanup Chris has done in xman for the next release, I promise :-) We're slowly headed towards the nirvana of having the core clients be useful programming examples as well as useful tools...
deboor@buddy.Berkeley.EDU (Adam R de Boor) (06/17/89)
I dunno, all these promises about R4 are making me mighty impatient :) Does the R4 xman geometry management still wipe out the displayed man page part if you resize one of its windows with both the directory and the man page showing? That one's also annoying... Hoping "this yearish" is soon, but understanding if it's not, a
rws@EXPO.LCS.MIT.EDU (06/17/89)
I dunno, all these promises about R4 are making me mighty impatient :) Gee, and we haven't even told you the GOOD stuff yet ...
kit@EXPO.LCS.MIT.EDU (Chris D. Peterson) (06/20/89)
> In each > of the applications I've looked at closely (well, both of them: xpostit and > xman), I find things have been hard-coded in that could have been much better > done via the translation manager. Xman will have this problem "fixed" in R4, I have made almost everything an action, this means that you can pretty much change everything around. Chris D. Peterson MIT X Consortium Net: kit@expo.lcs.mit.edu Phone: (617) 253 - 9608 Address: MIT - Room NE43-213
kit@EXPO.LCS.MIT.EDU (Chris D. Peterson) (06/21/89)
> Does the R4 xman geometry management still wipe out the displayed man page > part if you resize one of its windows with both the directory and the man > page showing? That one's also annoying... Nope, it took a major rewrite of the VPaned widget to fix it, but it has been fixed. Chris
earle@MAHENDO.JPL.NASA.GOV (Greg Earle) (07/23/89)
In article <8906161233.AA20184@LYRE.MIT.EDU> Ralph Swick writes: >Your observations and plea are certainly welcomed here. >We couldn't agree more! >... >My only real comment is in defense of at least one of the two >applications you cite, and others of their generation; it's only >a (mere :-) matter of time. In a (vaguely) similar vein, let me add a plea to those of you who have written contributed software for X11 - please read the new ICCCM and if you find that your programs are not setting all the `proper' hints for the Window Manager to allow correct usage of your program, then please consider adding the necessary code to do so, and re-submit it. The OPEN LOOK Window Manager (olwm), supplied as the default WM with Sun's X11/NeWS, is fiercely ICCCM-compliant, and I have found quite a few contrib programs that don't set the hints for getting keyboard focus from the WM, which renders them rather difficult to use (^: Two immediate examples that come to mind are the X versions of `mazewar' and `wanderer'. Thanks, - Greg Earle Disclaimer: Anything I post Sun Microsystems, Inc. from this account JPL on-site Software Support represents *my* earle@Sun.COM opinions *only*, earle@mahendo.JPL.NASA.GOV (Guest) and *not* Sun's.