[Date Prev][Date Next][Thread Prev][Thread Next]
- Subject: Re: SciTE, GUIs, desktop real estate
- From: Gavin Wraith <gavin@...>
- Date: Mon, 06 Oct 2008 12:09:29 +0100
In message <firstname.lastname@example.org> you wrote:
> On 05/10/2008 13:54, Gavin Wraith wrote:
> > But the Windows version has a fundamental drawback (I do not know
> > if versions on other platforms suffer the same): you can only view
> > one file at a time. I say "view", not "open".
> Indeed, Scintilla allows multiple views on the same document
> (some editors use this feature) but Neil never felt the need to
> implement this on SciTE (he prefers having several instances) and
> nobody else provided that either.
> Sometime I wish to have it, but I workaround this with multiple instances
> and bookmarks...
Well, I do too, now that I know how to set
> > Another gripe is the hypertrophied toolbar which takes up
> > valuable screen real estate.
> What toolbar? I haven't seen it in years... :-P
> You are kidding, right? MDI is totally out of fashion nowadays! :-D
Too bad for fashion. I often need to have four or five documents open
at once while I construct another. I like to place them where I want
on the desktop. Only for comparing related documents do I want
to have them side by side. I do not want the documents to change
'vertical' position when I make annotations on ones that are partly
> Even Microsoft dropped it. And I never liked it anyway.
> MS' choice is to multiply the instances in the task bar, which is a
> bad choice IMO.
> SciTE's choice is to use a tabbed interface (or multiple instances
> if we want, too), which is the right thing,
Not the right thing for my purposes.
> but we could have used a split window too, with views either on the
> same document (other place) or on another document.
Why 'split'? You might want to place multiple views anywhere on the screen.
Why be limited to putting them next to each other?
In the system I am used to, when you click a mouse button on an object
that is visible on the desktop the window manager sends a 'data-open'
message to the list of all running user-tasks, containing data such
as the mouse-button, pathname of the object, its filetype, position,
etc. The first task to claim the message sends back an acknowledgement,
so that no further tasks down the list are polled, and then does
what it needs to do (e.g. loading data, negotiating the opening of
windows with the window manager, responding to further messages
from the window manager on receipt of messages about user-actions etc).
It is just as easy for it to cope with multiple objects as it is
for one - it is just a question of having a table of object-data.
If there is a "main" window associated to an object one usually
takes the keys of the object-data table to be the reference numbers of the
main windows, which are returned by the window manager after a
successful request for a window.
I do not know how Windows deals with mouse-clicks, but it has to be
something along these lines, surely?
Gavin Wraith (email@example.com)
Home page: http://www.wra1th.plus.com/