Wgy isn't there a good RAD Gui tool fo python

[Andrew Berg]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: RIPEMD160

Rather than taking advantage of that convenience, commercial vendors
put barriers in the way and try to carve out little walled gardens.
Did they not learn anything from AOL?

DRM and activation schemes will /always/ make things harder, but that is
the cost of doing business, at least in the minds of commercial software
vendors.

There are actually a lot of good freeware (proprietary, but zero cost)
apps out there. Some even better than open-source alternatives. I avoid
commercial apps, though, since they tend to be far inferior to the
alternatives (inconvenience aside).

Where is the Windows equivalent of yum or apt-get? Why isn't there a
central repository of independent and third party Windows software?

If Microsoft made such a repository, how much of the repository would be
high-quality open-source software, and how much would be commercial
shovelware?

Attempts at independent repos have been made, but they all fail because
there's no effort among developers (especially developers of proprietary
software), to package their software this way. These attempts also fail
because they fail to gain support from users (a catch-22 where users
don't bother because there's not much in the repo and there's not much
in the repo because users don't bother).

It seems clear to me that it is the major open source communities
that aim for convenience, at the cost of the opportunity to sell
licences.

The developers of open-source projects often aim to please the user
rather than make money. You'd think pleasing the user and making money
would go hand-in-hand, but history has shown that the latter can be
achieved with little thought of the former.

That might have been true, oh, 20 years ago, but today, that's far
less of a rule. Linux distros make interoperability far simpler. Some
level of savvy is needed, but it is remarkable how much Linux
software Just Works.

At first, Linux had to learn how to crawl and then walk. Now it's doing
gymnastics.

In my experience, two categories of Linux software are generally hard
to deal with: one-man projects (usually stuck on version 0.2b for the
last seven years), and big, popular projects that have been taken
over by developers from the Windows world (I'm looking at you,
Firefox). YMMV.

Firefox (and Thunderbird with it) are falling into the same trap that
many fall into when they become popular. This is more prevalent among
commercial apps, but it's not too surprising considering Firefox's
popularity. The trap is making things shiny. That is, using UI designs
(and to a lesser extent adding neat, but generally useless features)
that appeal to the computer-illiterate masses who cling to something
that looks neat, regardless of how useful it ultimately is. AFAICT,
Mozilla's problem isn't that incompetent Windows-centric devs took over,
but rather that Google and MS were stepping up their game with their
respective browsers and is desperately trying not to lose market share.

- --
CPython 3.2 | Windows NT 6.1.7601.17592 | Thunderbird 5.0
PGP/GPG Public Key ID: 0xF88E034060A78FCB
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.11 (MingW32)
Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org/

iQEcBAEBAwAGBQJOHC4EAAoJEPiOA0Bgp4/Lgm0IAOT+/LQNalPHm5pvt4ilF1yt
RM9fPBSgAF5k9U8jWBuQy/V6QJ/a1Sfkzu8ulZ8TyAYS64quucIqTwMJugdTUmct
KsGbDsyXg0FObMxNiKKFuZblVYOtnULkYtYZOxeE33qy+85X6NMuFUv7ARHaLi/3
1Bdmnsj43hRrzJ1Rwb8x+xbOmiq+fJ7199loPQ+unSu7s37NJoL1e1vFNnsmGz8A
Jg58Q0MbGiwettPdM9ZySYWgTJhiawtEX4SF6YiQqf22e04OyPWyxUfejixnZNoQ
7vbksr9k8PQzuTlG2y3G1pJx6XGrxgOQuEoVjInMGbZW0tx43paJLEWCOcd38FI=
=3FGv
-----END PGP SIGNATURE-----

 

[Chris Angelico]

Where is the Windows equivalent of yum or apt-get? Why isn't there a central
repository of independent and third party Windows software? It seems clear
to me that it is the major open source communities that aim for
convenience, at the cost of the opportunity to sell licences.

The nearest commercial equivalent is probably Apple's iTunes store. It
manages to be the "one place to go" for iphone apps, many of which
cost money. Upside: Developers know where to host their stuff if they
want it to sell. Downside: Developers have to host it there if they
want it to sell - and Apple snag 30% on the way through.

I've not seen a Windows equivalent, but Microsoft could make one if
they wanted to. All they need is for the next version of Windows to
recommend that all software be signed, and make it somewhat awkward to
install unsigned software, and that would be that. It would probably
be the knell of Windows, but it could be done.

ChrisA

 

[CM]

I use a GUI builder because I'd rather click less than type more. I
just
tried that in Boa Constructor; with ~10 mouse clicks I produced 964
characters of Python code. Now, sure, depending on how I wrote the
code I could do better than that, but for me, I just find it more
intuitive
and easier to use a GUI to make a GUI.

Often a GUI builder is used as a bad replacement for sketch-pad and
pencil.

I would use a sketch-pad and pencil and *then* use the GUI builder.
What's nice about a builder is one can move things around quickly
and see the results in the real application, which one can never
really
see well on a paper sketch. You could use a mock-up program of
course, but I feel you might as well do it in the builder because when
you're satisfied with it you have a real runnable application instead
of just a picture.

Using a GUI builder with layout managers might actually
feel awkward.

It takes some getting used to in Boa, in my experience, but then
it feels intuitive and I really like using sizers with Boa. It helps
if
you give your sizers descriptive names.

Che

 

[rantingrick]

I use a GUI builder because I'd rather click less than
type more. I just tried that in Boa Constructor; with ~10
mouse clicks I produced 964 characters of Python code.

Remember, it's NOT the length of the code that matters, no, it's the
motion of the "sources" ocean. Did it produce rough seas full of
spaghetti monsters? Or tranquil fjords worth pining over (sadly to
death apparently?)?

1. Never judge the quality of code simply by it's length. Because if
you do, some folks might suffer from "source envy"!

Also, you MAY have created 964 chars of code with your ten or so
clicks HOWEVER that is just template code. You'll need to set many
attributes for the widgets before they are ready for prime time. Your
"supposed" ten or so click estimate is very naive. It takes MUCH more
to create even a simple GUI, because, we have NOT even discussed logic
yet!

Now, sure, depending on how I wrote the code I could do
better than that, but for me, I just find it more
intuitive and easier to use a GUI to make a GUI.

Personal opinions should always be respected, and as such i respect
yours but later i would outline my GUI design workflow so pay close
attention.

I would use a sketch-pad and pencil and *then* use the GUI builder.

But do you really? Your following statements lead me to believe that
you don't.

What's nice about a builder is one can move things around
quickly and see the results in the real application, which
one can never really see well on a paper sketch. 

I prefer to skip any pencil and paper completely myself. I just use my
imagination. UNLESS the GUI is EXTREMELY complicated. For me the
design of a GUI starts in my brain. No pencil, no paper, no three
hours using Auto Cad GUI designer. Next i start creating widgets and
laying them out using geometry managers (in CODE). Finally i run a few
tests, make a few changes, and design phase is over. Time for logic.

---------------------------------------------
My argument against GUI builders is two fold.
---------------------------------------------

1. GUI builders remove us from the initial "mental design phase" and
temp us to let our inner "click-ity-click" and "drag-ity-drag" child
loose. This inner child likes to play but he hates to plan. Very soon
he has the play room floor (source code) overflowing with toys (code)
arranged in a completely haphazard way. Unlike the child however,
there is no code mommy to spank this bad little boy when he is a
programmer. So he just keeps messing up play room after play room
making a complete fool of himself along the way.

2. GUI builders remove us from the source code. When you are playing
"clicky-click" with yourself you could be in the trenches fighting the
spaghetti code monster. Instead you are losing mental focus. Remember,
playing with yourself makes you lazy!
----------------------------------------------

What happens is... you get lost "playing" and fail to keep your mental
focus. A programmers metal focus is his most valuable weapon in the
fight against the spaghetti code monster. I am a programmer. I love my
source code more than i love most people in this world. I do not want
to be away from my source. I am jealous of my source! And so too
should you be.

Kevin made the argument earlier that Tkinter (and others) are so easy
to use that they render needing a GUI builder useless -- and he is
correct! But did you know that there are GUI libraries EVEN more
highly abstracted than Tkinter? Oh yes! So your "OMG, this typing and
using my imagination is so difficult" *crap* is really making me
laugh.

That is my argument people. Opinions may vary. Keep watch for the
spaghetti code monster!
Cheers folks.

PS: if you don't like to type, programming IS NOT the best career (or
hobby) choice for you.

 

[CM]

Remember, it's NOT the length of the code that matters, no, it's the
motion of the "sources" ocean. Did it produce rough seas full of
spaghetti monsters? Or tranquil fjords worth pining over (sadly to
death apparently?)?

In my experience, the GUI builder I use creates reasonable code
that deals with the GUI in a separate portion of the code. It does
not strike me as spaghetti-ish (though it's not perfect).

Also, you MAY have created 964 chars of code with your ten or so
clicks HOWEVER that is just template code. You'll need to set many
attributes for the widgets before they are ready for prime time. Your
"supposed" ten or so click estimate is very naive. It takes MUCH more
to create even a simple GUI, because, we have NOT even discussed logic
yet!

Sure. But my point was just that to even get as far as I did (which
was
just a frame and two unspecified widgets) takes 964+ keystrokes, but
only ~10 clicks. So the pacing of keystrokes:clicks is favorable.
If I built a small functioning GUI application, it might take 100
clicks
and 9,640 keystrokes (very roughly). But it is the same point.

But do you really? Your following statements lead me to believe that
you don't.

I just meant that though I might start on paper, once it is on
the screen I sometimes will shift things around a bit at that point
to see how it looks. This is easily done with sizers and a sizer
collection manager and an up/down arrow, so it is worth an
extra minute to just see how it looks.

 1. GUI builders remove us from the initial "mental design phase" and
temp us to let our inner "click-ity-click" and "drag-ity-drag" child
loose. This inner child likes to play but he hates to plan. Very soon
he has the play room floor (source code) overflowing with toys (code)
arranged in a completely haphazard way. Unlike the child however,
there is no code mommy to spank this bad little boy when he is a
programmer. So he just keeps messing up play room after play room
making a complete fool of himself along the way.

 2. GUI builders remove us from the source code. When you are playing
"clicky-click" with yourself you could be in the trenches fighting the
spaghetti code monster. Instead you are losing mental focus. Remember,
playing with yourself makes you lazy!

I've certainly heard of others who feel that working with only code
is "cleaner" for them, mentally speaking. I can understand that. I
think it just depends on what one is used to. I don't find the GUI
builder disrupts my ability to plan or keep things orderly. In fact,
most of my disorder and spaghetti problems have been in the logic
side of the applications, the part which the GUI builder doesn't have
anything to do with. (That's my own issue to keep working on).

Kevin made the argument earlier that Tkinter (and others) are so easy
to use that they render needing a GUI builder useless -- and he is
correct! But did you know that there are GUI libraries EVEN more
highly abstracted than Tkinter? Oh yes! So your "OMG, this typing and
using my imagination is so difficult" *crap* is really making me
laugh.

My attitude is, if I could speak in English to an AI to tell it what
I'd like
the program to do, I'd do it. Yes, since I can't do that, I inevitably
do
sometimes enjoy puzzling things out, but only because I have to.

PS: if you don't like to type, programming IS NOT the best career (or
hobby) choice for you.

I guess it is not so much that I dislike typing, as I dislike having
to
switch from visual mode to code mode, remember the keywords and
such for the widgets, rather than quickly clicking around. The
keystroke count is really just a proxy for that sort of effort.

CM

 

[rusi]

My attitude is, if I could speak in English to an AI to tell it what I'd like the
program to do, I'd do it. Yes, since I can't do that, I inevitably
do sometimes enjoy puzzling things out, but only because I have to.


I guess it is not so much that I dislike typing, as I dislike having
to
switch from visual mode to code mode, remember the keywords and
such for the widgets, rather than quickly clicking around.  The
keystroke count is really just a proxy for that sort of effort.

Yes. This is what is called the semantic gap.

Say you were a programmer who had to write software for numerical
analysis.
Would you write it in assembly even if, say, you knew assembly very
well?
I contend that most sane programmers would choose an algebraic
language because they understand (formally or intuitively it does not
matter) that minimizing semantic gaps are best for programming.

Writing text that indirectly describes a gui rather than directly
drawing it is analogous to writing assembly that implies an algebraic
operation instead of writing the algebra directly.

As for Kevin's point:

One reason there hasn't been much demand for a GUI builder is that, in
many cases, it's just as simpler or simpler to code a GUI by hand.

I am not sure how to interpret that. If you are saying that most of
today's gui builders are too close to suxware to be worth the time, I
guess its true (and eminently practical)

If on the other hand the claim is that the very idea of gui-builders
is a flawed one I think the jury is still out on that. And the
history of computer science repeatedly shows that very high level
ideas take decades to enter the mainstream.

Think of garbage collection in 1960 in the esoteric lisp finally
getting mainlined in Java in 1995.