Selector Gadget Experience

Following are some thoughts that I had when trying to do somethings in /View that might be common (or even helpful) to others working their way through the View experimental. ====================================================================== In the past few days I have been looking to do a simple thing - clone Allen Kamp's dropdown gadget to make it look more Windows/Internet Explorer like. The Win95 skins from DocKimbel might also be a good base as suggested on the Ally list. Specifically I want to: a. remove the slider from the dropdown b. make it show the number of items in the list not a fixed number with a scroll option This should be up to a maximum that is driven by viewable space. For now, I need only to get a short list displayed. c. use a more concise arrow that triggers the dropdown I need these for a prototype using REBOL/View where the first round of feedback focused on look-and-feel differences in a couple of key View gadgets from the brower base that the public is used to. The goal is to remove these distractions so that the product content can be discussed instead of the look and feel. Other items that I would like to see but so far do not plan to attempt until the other things get easier for me: i. add support for keyboard up down arrows as Carl implemented in his easyvid.r script so that when a dropdown that has focus gets a keyboard arrow event the dropdown reacts accordingly ii. change the dropdown selection colors to match the blue background and highlight that Windows/IE uses rather than the yellow that View defaults to. Here were the challenges: A. REBOL has some good background theory stating that communication and language are the keys. But in the gadget model deployed, some of the naming conventions seem to have varied from these tenets. For example, names such as: i. [f a e] instead of [face action event] ii. [sld sn slf lc] instead of [ slider slider-index-number (?) slider-font (?) slider-line-count (?) ] Guessing as I have done what these elements are intended for seems fundamentally different from the beliefs about language that REBOL professes. Well more accurately I think.... REBOL doesn't just profess it but displays this fundamentally in its deployment everywhere I have seen except in this instance. And I think that the model of the user interface is key to understanding and extending it, and allowing others aside from Sterling and Allen and ... to answer more of the View questions. B. The ability to probe the model is encumbered. If I probe an instance of a view gadget, REBOL goes off for a very long time. Today I probed an instance of a dropdown and found the first SLIDER token on line 4685 of the probe output. When it is scrolling by 4685 lines is a lot of wait time on my PC. Plus the first 5 times I did it, I did not wait for more than 2000 lines to go by. When I let it run longer, I must have missed that text. (You may find that the shell does not give the ability to see that many lines and drops the first ones. You can write the full probe results to a file by using: layout [b: button "Test Button"] write %yourfile.txt mold :b Then using a text editor such as Textpad, searches are possible for the keywords. That was how I got to looking for SLD today.) I find the "SLD" token (that I did not look for at all until I got the notes from Graham, Allan and Scott) on line 5381. I think that if the REBOL/View user interface is to be successful it needs more programmers able to explore the model more easily than searching text files. It needs a graphical tree representation that is "explorable". [mike--myers--cybarite--com]

Mike Myers wrote:

> I think that if the REBOL/View user interface is to be successful > it needs more programmers able to explore the model more easily > than searching text files. > > It needs a graphical tree representation that is "explorable".