Subject | RE: [IBO] Restructure minor problems and some requests |
---|---|
Author | Kaputnik |
Post date | 2002-03-03T19:41:40Z |
> >O.K. Will be changed......Its not too hard anyway and so I'll go teh way of
> > > b) If you move an item to the top or the bottom of the list using
> > > the "go to top" and "go to bottom" button, the item looses focus
> >
> > yes, this is intended, but can be changed easily. The reason is
> simply, that
> > by moving an item to the top or bottom, you're normally done
> with your work
> > and select new columns anyway. Can be changed though
>
> Ok, but it can also be used like a "fast forward" and then refine
> the position.
>
least resistance
> >I'll see If I can find some nice way to do it. Both ways are confusing. The
> > > c) if you have item 1...9 and hide 2,3,4,5, if you then take the
> > > first item and click the down arrow to move it down, you have to
> > > click 6 times to have it move to the second position. It's like
> > > the code is stepping through the list without jumping the hidden
> > > (so not more present) elements of the visible columns.
> >
> > Yes, this is intended. I am moving the items through all fields
> order, not
> > all visible fields order. Doing the latter would have introduced some
> > complexity I did not have enough time to cope with. You can see
> the behavoir
> > at full work, if you move one column up while having several invisible
> > columns over it and set these columns visible again. I prefer
> this that way
> > but I think there is a possiblility to change the behavior in the future
>
> It's confusing... you don't know if you clicked the button or
> not, or what is happening. As a programmer, I can understand such
> behaviour, but a simple user will not.
>
now present one will not move the fields, if there are invisible, and the
other one will move a field not only 1 position up, but could move it 4 or
10 positions up in one click.
Perhaps there is a way to combine both into it...
>ANd requiring lots of coding. If anybody wants to, he's welcome.....perhaps
> The
> > default delphi splitter behavior (I am using a plain TSplitter
> here) has no
> > support for proportional splitting, so I would need to code it
> by hand, and
> > I don't see too much use for this right now...
>
> It's just making thing more useful
>
I can tinker around with anchors a bit....
> >Cooper is an authority, but not always right. Only god is always right, and
> > > e) I HATE flat buttons, they break one of the basic interface
> > > design principles, so I always set the bar Flat to false.
> > > Unfortunately, your window ignores this setup... ;)
> > >
> >
> > Ouch, you mean the flat buttons for moving the columns? Well,
> flat or not
> > flat is personal preference and not design principle and is not
> a subject of
> > discussion for me.
> > I like them, you not. Period.
>
> Ok, but IT IS A DESIGN PRINCIPLE. If you put somebody in fron of
> something that has the look of being pushable, he will try to
> push it. If flat, who can imagin that can be pressed? IE is one
> of the worse interface I've seen. You can't understand what
> button is active if you don't move the mouse on it, and you have
> to imagin that those stragne pictures are buttons and not
> decorative effects. Of course, you can learn it, but you can also
> leark WordStar key combinations...
> (have a look to www.cooper.com).
> But, of course, what Microsoft does, everybody likes...
>
Cooper is not him.
Most people I know prefer flat, and none of my targeted users is confused by
flat, as any targeted user knows IE better than Netscape :-)
Lets take it really to preference, not design principle. A glyph itself
imposes enough hint to push it.
> >Don't forget: The restructure dialog work dataset-centric, and not
> > > What about having some more property to customize? Like colors...
> >
> > What do you mean with this? You mean an interface for TIB_SessionProps?
> No, please, not. Just grid properties (fonts, selected color,
> etc.... but maybe it's a wrong idea, some user could make a mess
> of the grid :(
>
grid-centric. The dialog does not know anything of the grid. It only knows
the Datasource and its corresponding Dataset. I can only change
Dataset-Properties, not grid ones. THe grid is only adapting the changes as
they happen :-)
> > I think this should be fairly simple provided I can find a niceFormstorage works both, with the registry and with Ini-Files. I prefer
> colorpicker
> > which is Public Domain (and NOT Freeware) to inlcude the units
> directly into
> > IBO, and not to require third-party stuff to be installed along
> with IBO.
> >
> > > add a PageControl... but ask Jason first. What I would like to
> > > have, and is a general consideration, is a component that can
> > > save forms (I mean, components, grids, etc.) state to a blob
> > > field of a table with form_name+user_id fields, so every time a
> > > user reopens the form, the values he used the last time are
> > > presented back (very useful for forms asking parameters...).
> >
> > I use TFormStorage from RXLib to save that stuff into the
> registry....I do
> > not want to bloat the Database with such stuff. I have Formstorages in
> > Datamodules with a custom onRestorePlacement and
> onSavePlacement handler,
> > where all Queries on a Datamodule are looped through and the properties
> > FieldsIndex, FieldsVisible, OrderingItemNo and
> FieldsDisplayWidth are saved
> > and restored. Works like a charm, and RXLib is free....
> > When I create a form in any kind of app, the first thing I do
> is to drop the
> > TFormPlacement onto it, and set the registry-key. Saving
> formposition, size
> > and all kinds of properties gathered trhough RTTI is such a
> nice thing, I
> > would NEVER try to code this stuff myself. Use what's already there, and
> > RXLIb is definitely a must-be to consider for every
> delphi-installation :-)
>
> Ok, but I don't like put things in registry. If you reinstall
> Windows, you loose your setup. But thanks for the suggestion. I
> will have a look at Rx components, and adapt them to IBO
> properties and definitions :: I think it will a good starting
> point (btw, si RxLib still alive? Their site is... something
> different now! http://www.rxlib.com/)
>
registry, other prefer ini-files. Again, a matter of preference.
Cu, Nick