boblq([EMAIL PROTECTED])@Sun, Apr 10, 2005 at 01:23:58AM -0700: > On Saturday 09 April 2005 09:12 pm, John H. Robinson, IV wrote: > > > > Of course, in a perfect world, all the tools would have CLI interfaces > > will well-defined interfaces, so the interested parties can write > > whatever GUI front ends they want. > > > > John, that is an absurdly parochial comment. Many end users > just want to do something, and that something is rarely to > "write whatever GUI front ends they want." >
I don't see the absurdity. Why does the existence of many-end- users-who-rarely-want-to-write-GUIs, invalidate a model that makes it easier and more flexible to write a GUI? Seems likely that there would be more people willing to write GUIs for a program if it were easy. And in that case, those many other users would end up with a choice of GUIs (if we assume those GUI authors are releasing programs with an open source license). I don't see that the CLI must be the _first_ available UI, as long as the logic and interface are well defined. On the other hand, having a batch/CLI interface would be the most flexible one. I would much prefer to use the same program/tool in an ssh/CLI session as I do when I have access to X. If the first UI available for the tool is graphical, it is only useful to me in one environment, and I have to be at the PC to use it (can't effectively put it into a script). My perfect world includes programs that pay attention to the standard file descriptors, and can be run in a batch mode (CLI), as a console (a la ncurses) app, and as a gui app. Configure it with one ~/.configrc regardless of the UI, and make it easier for the end user. (Or do it like vim where you can create a .gvimrc that can include your .vimrc). They should also look for an /etc/.configrc so that the SA can make the app useful if the end user doesn't want to create a custom configuration. The fact that such functionality isn't valued by most doesn't make my, or John's, preference for such things useless or parochial. It is simply a recognition that flexibility is better. It is every bit possible to take a flexible program and adapt for those who don't or can't use it as is. It is unfair to say that John is being parochial. He simply sees that a program's inflexibility prevents it from being useful to those who can and would find new, different, adaptations for it. Why should John or I be satisfied with inflexible, GUI-only programs, just because the majority of end-users wouldn't realize what was missing? I would suggest that the end user would still benefit from an easily adapted program with separated logic and interfaces, similar to the way that non-programming users benefit from open source code. Wade "What does SLB stand for, anyway?" Curry syntaxman -- [email protected] http://www.kernel-panic.org/cgi-bin/mailman/listinfo/kplug-lpsg
