Table fields... oh boy.

Troy Rollins troy at rpsystems.net
Sun Aug 15 13:08:17 EDT 2004


On Aug 15, 2004, at 11:17 AM, Richard Gaskin wrote:

> Spreadsheets and iTunes are very different creatures.
>
> Therein lies the difficulty in satisfying the request:  so many 
> options,
> so little specification.

OK, fair enough. Either. Both. A flexible solution of multi-column, 
single line or multi-line data display.

> I'm not sure we're talking about the same things.  Not that what you're
> looking for is necessarily trivial, but rather on the contrary: there
> are MANY types of multi-column lists, if you observe them closely.
> There is no one-size-fits-all.
>
> Some specifics about what you're looking for would be helpful in
> providing a solution, but you've already ruled that out: you'll accept
> nothing that isn't built into the engine, and no one can build it into
> the engine until its details have been specified.

Specifically, for this upcoming project, I need single-line, resizable 
by headers, with contained checkboxes, possible progress bars, and 
small icons. Multi-line would be nice, but that is its own batch of 
headaches, I realize.

> Given the world of difference between the two I hope you'll forgive the
> confusion and, if you'll reconsider the requirement that it be fully
> engine-based we might be able to provide what you need once we
> understand it.

Yes, it IS confusing. I don't require an engine modification 
specifically, but #670 pretty clearly identifies the nature of tool I'm 
looking for. While all higher-level languages cause developers to 
"make-do" with the collection of resources available, and design and 
build projects based on the tools which they have at hand, this 
particular feature is particularly complex to "fake" well.

A data grid, with whatever features which could be dragged out of the 
toolbar into a project would be ideal, because like all the other 
tools, it would have a nice API, and be easy to integrate into any 
project. Just another incarnation of a field would be less than 
desirable unless it was very carefully worked up. The current table 
object hardly behaves like a proper table. (e.g. tabbing past the right 
edge, cell editing, etc.)

> These are not subtle differences; data grids are inherently complex
> objects.  This is probably why OS vendors haven't bothered to make an
> API for one.

True. And no doubt the reason I really don't want to hack together a 
one-off custom version every time one is needed.

The existing tab-stopped based field, is simply not good enough. But, 
if there were some decent generic version, with whatever feature set, 
like a good high-level language developer, I'd try to design my project 
around whatever it could do.

A RunRev solution would be best, since it would be supported, etc. But 
I also have no doubt that it is possible to make something generic 
through transcript alone... I've never seen one, but that doesn't mean 
it isn't possible.

>
> So there are three options at hand:
>
> - Specify what you need and let's see if someone has what
>   you're looking for in scripted form.

A working table / data grid object, which allows column resizing and 
sorting via headers, cell editing, proper treatment of the tab and 
return keys, lockable horizontal size - with a reasonable API (addRow, 
delRow, rowClicked,cellClicked, etc.) The rest of it is bonus features, 
I guess.

>
> - Specify what you need and let's see if it could be added
>   to the engine.

#670 does a good job, though admittedly vague and broad. I didn't write 
it, but based on the signers-on, I'm not the only one who would like to 
see such a tool.

>
> - Send us a postcard from El Dorado
>   (ride, bodly ride, the shade replied...)

Yes, Mr. Poe, it could be that I "seek for Eldorado." I have a tendency 
to want high-level tools to offer the same unrestricted features as 
low-level tools, and low-level tools to offer the same productivity and 
freedom from gibberish syntax that high-level tools do. There really is 
no Eldorado, is there?  Darn. I guess with high-level languages, you 
pick the one with the best features/compromises ratio, and make it work 
from there. Doesn't mean I won't keep up the search, or trying to get 
Revolution to continue transforming in the direction of Eldorado.

--
Troy
RPSystems, Ltd.
http://www.rpsystems.net



More information about the Use-livecode mailing list