On 2010-08-10, at 19:19:46, Shane Stanley wrote:
Also it looks that the programing using libraries is not common in AppleScripters.
This is the problem we must tackle.
I don't know that it's a problem to tackle, or whether it's a reason to ask if all the effort is worthwhile.
Perhaps a good starting point would be to ask why people don't use libraries, and then see if the issues raised can be answered.
So far, the assumption seems to be that they are unused because of things to do with where they live, how they get loaded, naming, and so on. I suspect the reasons are quite different, and that extensive libraries don't offer any great advantages to a lot of scripters. In the era of AppleScriptObjC, perhaps even more so.
Whatever the reasons, tackling them would be a lot easier if they were identified.
Yes and most reasonable to address raison d'être questions (pardon my French). Here are some things I've considered.
• No 'get handlers and properties from …' command at the scripting level.
• No convenient built-in commands for library management.
• AppleScript Editor doesn't have an "Insert 'load script …'" menu item as a user convenience.
• Libraries are not too popular because they don't have much formalized support from the system.
Try to address that by designing something that can be accessed easily from all levels of scripting and coding available. Make it a "general" concept so folks with a different take on things can perceive a usage value.
• Apple seems to think that Automator is the solution for the library concept.
This list has about 100 times the activity of Automator's because there's a feeling of grass roots involvement. I'm candid with my remarks but not disrespectful — Automator is really powerful with extensive libraries — and certainly the work of an experienced developer — but it's interface is like a straight-jacket. That's what happens when your rely on the Cocoa frameworks to do your designing for you.
So, instead of an imposed interface, start with a design which has no interface and let folks add their own. After a while, a bubble effect happens and preference of the collective becomes apparent.
Try to get some consensus from those even mildly interested as to a lowest common denominator. Wish lists are welcome because sometimes a casual remark can trigger a fundamental design change which then permits more flexibility.
A small but exacting "kernel design" concept with very few restrictions. Something folks can realize is easily extensible for their purposes — with or without a UI.