Jan 20, 2006
8:41 AM EDT
I'm curious about some aspects of what you're proposing. To my way of thinking, perhaps, with enough work, you could implement something like your framework for a word processor, but how would you propose that a "light weight framework" be built for, say, a spreadsheet? Or a desktop database? Or especially a presentation application? Also, should this be built on QT, or GTK? Or should the framework include enough code to allow it to compile using either (these are GUI applications, after all)? Or should it be built using some other not-yet-invented widget tool set? Or do you mean that these processes/plugins would not get a terminal, but would instead act more like background processes with the GUI application piece built by the vendor accessing them via pipes or sockets or something? Also, given how complex this backend infrastructure sounds, are you proposing that average desktop users should have to edit these configuration files in order to create something fairly complex, such as a newsletter layout? Or is this something that should be done programatically?
Jan 20, 2006
9:17 AM EDT
|rob: I didn't write this one. Should you address it to Herschel?|
Jan 20, 2006
3:29 PM EDT
|oops... Saw the posted by and didn't notice the author. Ah well, same questions apply if Herschel would care to reply.|
Jan 22, 2006
10:49 AM EDT
|rob - sorry for being so late in giving any sort of reply, been busy trying to research the next installment of this line of thought.
Since I am having problems finding some material I read years ago I am going to cheat just a bit and use items I can just cite and some of my thoughts on this topic that went out as part of an email to someone else.
First look at the comments in another comment thread on this same story by Tux_Wonderdog http://lxer.com/module/forums/t/21319/ because he is thinking in a quite similar manner as I.
Next, you are stressing the gui part too much - I delayed responding in part because I thought there were a class of applications that might rightly be nearly all graphically based. Upon reflection I see this group can be probably smaller than I first thought. Let me stress I see this first as a data problem, where graphics should be introduced only when they play irreplaceable role. Now let me steal some of my thoughts:
" ... I see most of it as a data storage problems where existing data types and scopes should suffice to describe even "graphical data" as mostly differing types of numerics. Think of digital photos in pixels: location (values along set axis, 2 to 3 values at most as large integer or floats), density (large integer or float), color (three sets of integers in sequence), reflection (light source angle) integer (polar angle) or float, etc. If it were a display or drawing light color washes could be described numerically (color and density) if dynamic one could use an equation ,with trig. type functions to describe density fluctuation of color density and transparency. Once this data has been processed it could then be thrown at a graphical display program."
Think of needing a snap shot that you wish to print, you might want less than the full data set, i.e a cropped photo and due to printing the resolution will be less than the stored potential in the data. Why make the graphics tool pretend to be active when it is extracting and number crunching that is really using the processing resources. Here at best the cropping could be graphics and a preview of the likely appearance of the final print could be displayed. In the preview the density printed is a weighted average as is the color for every new, bloated pixel.
Here is where I think it might be done as I have in mind, but that is only my mind extrapolating from some discussions I have read, I could be completely wrong, but it's a reasonable path:
" I think a lot of astronomy is now done by similar means. A lot of the current earth bound telescopes have to remover distortions due to the atmosphere and structural displacement of large and separate mirrors whose total capture has to be processed prior to producing a picture."
You should be able to infer I am a database person with some scientific training - both would be correct.
Let me know if this suffices for now - HC.
Posting in this forum is limited to members of the group: [ForumMods, SITEADMINS, MEMBERS.]
Becoming a member of LXer is easy and free. Join Us!