[REBOL] VID, %, dialects WAS: percent! - new datatype request
From: greggirwin:mindspring at: 13-Jun-2002 13:02
The VID vein, and what should be added to make it a better dialect for GUI
layouts, is something I've spent a little time thinking about, and I have an
opinion. (saw that coming, didn't ya? :)
VID is good, really good I think. Could it be better? Undoubtedly. Remember,
though, that VID+Layout is just *one example* of a GUI dialect. I keep
seeing requests from people for RT to add this or that to REBOL/VID/IOS but
I think the real future lies with us. Yes, RT may want or need to add
certain key functionality but they have already given us a foundation on
which we can build pretty much anything we want. Yes, a standard is a
terrific thing to have around, and they need to be aware of its importance,
but I think we're going to see many "specialized" standards emerge over
VID is a good, general, GUI dialect for doing screen layouts. As Joel
mentioned, resizable screens aren't easy to get right. Resizing individual
components isn't so hard, but maintaining a well balanced layout *is*. Just
look at how most apps handle it.
Rather than saying VID is lacking because it can't handle proportional
sizing, we can write our own flex-view and flex-layout functions based on
lessons learned writing ad-hoc resize routines. Maybe it's even a
combination. Maybe defining a flex-face style gets us most of the way there.
Is there a single heuristic we can apply to all visual layouts? Of course
not! It all depends on what, and with whom, you're trying to communicate. We
should see dialects for all kinds of specific purposes, and templates to go
with them. Some will be industry specific, some output/device related, and
some process oriented.
One of the problems I *do* see is that the stuff RT provides is more than
, so not many people in this small community are driven to
replace them, or even enhance them in a major way. Another problem is the
lack of documentation that would help us to tackle those tasks with
confidence. If things were reversed (i.e. really good docs and limited
feature implementations), what do you think would happen? Would we complain
about the poor examples and write REBOL off, or would we see the potential
and rise to the challenge?
In the early days of VB, writing VBXs with the CDK was pretty painful, so
not many people did it. The concept behind them was so powerful, though,
that it survived. Now, I'd bet not many people write VB apps using only the
11 standard controls.
The potential is there.