in reply to User Interfaces

It would be nice, but you run into the problem of some features supported on some toolkits, and not on others, and the ways to call such are all different. Sure, I know this is what a unified toolkit would try to solve , but unlike DBI and databases, where there is a nice standard SQL that all of them basically support, toolkits have no such standard so that even trying to distill a subset of GUI elements that would work would be a rather daunting task. Also, you have problems with how events are handled in the various toolkits; this is probably a bigger stumbling block for a perl interface than it might seem.

The best way to do any sort of problem that might have a graphical interface is to remove as much of the non-GUI-based logic from the code into a module, and then write GUI code that uses that module; this way, if you write a QT app and I'd like to make a GTK version, I don't have to play around with your logic, only do the necessary 'mapping' of GUI elements.

Dr. Michael K. Neylon - mneylon-pm@masemware.com || "You've left the lens cap of your mind on again, Pinky" - The Brain

Replies are listed 'Best First'.
Re: User Interfaces
by Reverend Phil (Pilgrim) on May 02, 2001 at 00:06 UTC
    Perhaps you're both right. It would be entirely convenient for there to be a standardized interface with gui generating applications or extensions. And yes, such a standardized language as SQL is required for this kind of an undertaking. Maybe that's what needs to be done. Maybe someone with some experience and three fourths of a brain should whip up an RFC for a standardized interface for GUI apps (SGL -> Standardized GUI Language). Once the community has beaten around a bunch of ideas, a standard could be developed, and then we could work on the DBI equivalents for interfacing each GUI module or application via SGL. A standard has to start as an idea and become a standard. Of course there will be functionality within some specific GUI apps that would not be supported directly through the SGL, but eventually vendors/developers would include ways of accessing this functionality via extended SGL commands.

    *shrug*

    A lack of a standard is often an indication that a standard needs to be developed, rather than a sign that all movement in the direction of a standard is hopeless.

    I think it would be rather keen.

    -=rev=-
[reply]
      If such a standard would exist, I certainly would consider writing some GUIs after all. My attempts thusfar have stranded at the superficial use of quite a variety of toolkits.

      Such a standard would really be useful IMHO.

      Jeroen
      "We are not alone"(FZ)

[reply]
Re: Re: User Interfaces
by petesmiley (Friar) on May 01, 2001 at 22:36 UTC
    Just an idea that tumbled out while I was doing something meaningless :) Take note, I'm not seriously considering this, I just thought it would be something interesting to throw out. I find that when I work on a project I spend more time working on user interfaces than anything else.

    Besides, I thought it wasn't necessarily just SQL that was important, I thought it was an issue of database drivers, which don't all communicate the same. The SQL commands between databases might be the same, but the method for connecting and communicating would be different.

    I certainly realize that user interfaces can be way more complex than database interfaces in some ways. However, I think the idea is amusing.

[reply]

In Section Meditations