I have bought wxDesigner and use it. It is inexpensive and worth having.
However, as a genuine development tool I find it to be not nearly what I
would consider "production level". What is especially odd is that for a
tool for developing GUIs, its own GUI is very crude, has the feeling of
being an "alpha release", and it often takes quite a bit of
experimentation to figure out how to do something. In addition it lacks
several fundamental featues I would expect in a commercial product. But I
won't provide a complete critque here.
The only way in which I use it is to do mockups of dialogs and it is a
tremendous help and time-saver in getting all the sizer stuff right. Once
that is done, I take the resulting code and modify it to my own tastes
(usually encapsulating things in classes, changing all the goofy
auto-generated names, adding tweaks, etc.).
In my view it is a useful (though sometimes frustrating) tool for learning
about how wxPython works and it is very valuable in doing mockups. Beyond
that, I would not attempt to push it. I think that a little more effort in
its design and useability could make it a truly marvelous tool. As it is,
if you don't expect too much, you will be very happy with what you get. If
you have high expectations, you will be disappointed.
Coupled with this I use the Wing IDE for general Python development (have
been using it for over a year) and think very highly of it. I know they
are working on a significantly improved UI for Wing and added
functionality, but I don't know when they plan to release it. For its
price (comparable to that for wxDesigner), I find Wing to be an
astoundingly good product with excellent implementation, very good
documentation, and excellent support.
···
--------------------------------------
Gary H. Merrill
Director and Principal Scientist, New Applications
Data Exploration Sciences
GlaxoSmithKline Inc.
(919) 483-8456
Which is probably to be expected, considering that it is written by a Medical Doctor in his spare time.
I definitely agree though that even with its LnF it can be a huge time-saver for building complex layouts. I think that the next time I need to do something complex I will try wxGlade and then compare the two...
···
gary.h.merrill@gsk.com wrote:
I have bought wxDesigner and use it. It is inexpensive and worth having.
However, as a genuine development tool I find it to be not nearly what I
would consider "production level". What is especially odd is that for a
tool for developing GUIs, its own GUI is very crude, has the feeling of
being an "alpha release", and it often takes quite a bit of
experimentation to figure out how to do something. In addition it lacks
several fundamental featues I would expect in a commercial product.
--
Robin Dunn
Software Craftsman http://wxPython.org Java give you jitters? Relax with wxPython!
I did and just to put in my 2 cents here I like wxDesigner better. When I
started using wxDesigner I tried to use the generated code, however as voiced
before I didn't find it suitable for larger projects. You can encapsulate in
classes etc, but in the end there is always this little bit you can't do and
have to tweak by hand. All the changes then will be lost next time you add a
dialog, since the one generated layout file is overwritten. So meanwhile I'm
just using it as a type of very comfortable xrced. As others have voiced I
think wxDesigner is worth the money and really saves a lot of time designing
dialogs. But even for XRC files it misses some tiny bit of things, i.e. one
can't override the class type generated (no "user type" controls, except the
foreign control, which usually screws up the layout). In favor of wxDesigner
I simply replaced all the XResourceHandlers and have them generate a control
based on a database table (i.e. it now creates a MaskedControl for zipcodes).
Call me old fashioned, but I still think the best "IDE" around is emacs. So I
create XRessources and do the rest by hand - faster, more relieable and
definitely more flexible than any other tool I tried - and I basically tried
all of them since I've written an application with 100k+ lines of code and
still maintain it, so a good environment to do this was my major concern. But
I finally gave up and use wxDesigner, xrced, emacs and CVS
UC
···
On Thursday 17 July 2003 11:25 am, Robin Dunn wrote:
gary.h.merrill@gsk.com wrote:
> I have bought wxDesigner and use it. It is inexpensive and worth having.
> However, as a genuine development tool I find it to be not nearly what I
> would consider "production level". What is especially odd is that for a
> tool for developing GUIs, its own GUI is very crude, has the feeling of
> being an "alpha release", and it often takes quite a bit of
> experimentation to figure out how to do something. In addition it lacks
> several fundamental featues I would expect in a commercial product.
Which is probably to be expected, considering that it is written by a
Medical Doctor in his spare time.
I definitely agree though that even with its LnF it can be a huge
time-saver for building complex layouts. I think that the next time I
need to do something complex I will try wxGlade and then compare the two...
--
Open Source Solutions 4U, LLC 2570 Fleetwood Drive
Phone: +1 650 872 2425 San Bruno, CA 94066
Cell: +1 650 302 2405 United States
Fax: +1 650 872 2417
I definitely agree though that even with its LnF it can be a huge
time-saver for building complex layouts. I think that the next time I
need to do something complex I will try wxGlade and then compare the
two...
What I'd love to see (yeah, and one of these days I need to put my
development time where my mouth is) is something similar to wxDesigner
but with an "apply the sizers after the fact" design methodology.
After experimenting with a lot of tools, I've come to appreciate the
way Qt Designer does it, and wish there were something closer to that
for wxWindows/wxPython. The initial layout of controls (including
spacers) is direct positioning, so you end up with a fixed position
set of child controls. But you can then select groups of controls and
fit them to a layout (Qt's sizer equivalent). And you can easily make
and break layouts as you experiment.
When observing our developers here, one of the biggest hurdles seems
to be the mental model you have to work within to decide how to create
the appropriate sizer hierarchy to contain your controls. wxDesigner
permits you to experience the result interactively, which can be a big
win, but it still forces you to think top down to design the sizer
hierarchy, which requires a lot of up front thought. The other sizer
based tools (or XML resource editors) all seem to work in a similar
vein. So to me, I often find myself working nearly as hard to get the
right hierarchy of sizers and inserting the controls appropriately as
when I just do it in code.
However, when initially laying out a window or dialog, it's often
easiest to just drag and drop the controls to get some perspective on
the layout, without concern for sizers. Once you have the major
components layed out, then you can consider how you might want them
interacting within a variable container window, and can apply the
layout after the fact. You'd still end up with the same sizer based
system, but you'd have gotten there in a more intuitive (at least to
me) and interactive way.