Purpose: develop a list of needs and wishes for someone developing stand alone or suites of applications. See also [Extension Developer's Wishlist for Tcl] and [Application User's Wishlist for Tcl]. Tk can also be covered here (we can spin it off to a new page if necessary). What constitutes '''urgent''' or '''required'''? Good question. What is critical for one person might be fluff or nonsense to someone else. The best I can do is give this kind of guideline. I would personally consider a feature urgent and needed in Tcl if it required the core to be modified, or extensive new code to be developed (as opposed to a wrapping for an existing library), to support some feature that might be useful to more than one customer. [Tcl 9.0 Wishlist] would, in my mind, be a subset of the total wishlist. ---- '''Urgent/required features missing in Tcl''' 1. Support for OO development. [Minimal OO requirements for Tcl] is the start of some work on defining this. 1. More basic functions that the Tcl programmer can invoke. What, specifically though, needs to be added? 1. Native cross-platform support for drag and drop in Tk, including suitable default bindings for some widgets (text, entry widgets in particular), and library procedures where necessary to make it easy for other widgets to be made drag'n'drop aware. '''Features that would be useful/helpful/cute but not necessary''' 1. I would put bundling itcl or other extensions into this category myself. This is because an application developer can make use of technologies like [Tclkit], Freewrap, the TclWrap, etc. to create a stand alone bundle of code. Having one or more existing extensions guaranteed to be present on the desktop if Tcl is there becomes, in my mind, a helpful, but not critical/necessary wish. ---- Joe's Current Wish List... 1. A "file truncate" command. --> see [ftruncate] in [TclX] NOTE: I meant a truncate command that is in the core, is fully cross platform, and operates on an open file_id (I have part of an implementation for Win32 done already). 2. An "lsearch -nocase" option, for uniformity with string match. 3. Bug fixes to make loading/unloading/reloading Tcl/Tk work properly. 4. Blue Sky #1: all "known issues" fixed. 5. Blue Sky #2: fully thread-safe/bulletproof Tk. JJM ---- 1. a built in debugger ala perl that lets you single step and examine variables. TclPro, RamDebugger, idebug, etc. all fall short [AK]: What is missing from [TclPro] to make it fall short in your eyes ? [PSW]: 1. It is not built in 2. [TclPro] is impossible (at least very difficult) to build 3. The newer version tcldev is not free [LV]: active support ---- 1. [BLT]s bgexec. It makes control of both stdout and stderr of an exec'd program very easy. This feature is requested every once in a while in the newsgroup. [US] Is it really bgexec that you want, or just easier control of stdout and stderr? bgexec has most of the limitations of Tcl's [exec] and it seems a shame to continue those if something new were being added. ---- '''Breadcrust's Wish List''' 1. ''Fileevent errorreadable'' to allow the programs to take care of stderr instead of having to redirect it or using catch. (nice) 2. Ability to read more hardware ports on all platforms (eg - Parallel, USB, IR). (required) 3. Something like TclX's ''filelist'' built-in. (nice) 4. Tk image rotate (required) 5. Toplevel -title -geometry & -icon options (nice) 6. ''wm iconbitmap -defualt'' on all platforms, not just windoze (nice) 7. Command to change window icon to image stored in variable (nice) 8. Statusbar & Iconbar/Toolbar widgets (nice) 9. Linux kernal level Tcl execution (MISC plug-in driver) (nice) ---- [LV] a free tool that provides reports and possibly interactive access to variable and proc cross referencing, ala C's cscope program. ---- [Category Suggestions]