A must article for tcl/tk'ers ...
call_me_anything
suchenwi
> http://freshmeat.net/articles/view/928/
where it says:
"Please note that I won't cover Motif/Lesstif in this article because
it is similar to Xlib in terms of age and complexity. Tcl/Tk isn't
much better."
Age: need not be a bad thing (maybe he means the pre-Tile looks of
Tk).
Complexity: is any of the reviewed toolkits simpler to use (at same
power) than Tk?
Uwe Klein
> On 5 Feb., 12:39, "call_me_anything" <s...@yahoo.com> wrote:
>
>>http://freshmeat.net/articles/view/928/
>
>
> where it says:
> "Please note that I won't cover Motif/Lesstif in this article because
> it is similar to Xlib in terms of age and complexity. Tcl/Tk isn't
> much better."
>
> Age: need not be a bad thing (maybe he means the pre-Tile looks of
> Tk).
most people don't get that there is a difference between esthetics and
functionality.
>
> Complexity: is any of the reviewed toolkits simpler to use (at same
> power) than Tk?
>
uwe
Keith Nash
>> "Please note that I won't cover Motif/Lesstif in this article because
>> it is similar to Xlib in terms of age and complexity. Tcl/Tk isn't
>> much better."
>>
>> Age: need not be a bad thing (maybe he means the pre-Tile looks of
>> Tk).
>
> functionality.
>>
>> Complexity: is any of the reviewed toolkits simpler to use (at same
>> power) than Tk?
It would be good to write some counter-propaganda for release when 8.5.0 is
ready.
Keith.
Ian Bell
> http://freshmeat.net/articles/view/928/
Personally I was unimpressed. It is written by a student and simply reflects
his prejudices and lack of knowledge/experience. Not surprisingly, the
usual pundits for each toolkit follow up with their own.
Ian
sleb...@gmail.com
Funny thing is that, back in 2003 I did indeed read the article and
decided to try the mentioned toolkits. WxWidgets looked the most
promising with a cheap (free) commercial-release license.
After weeks of playing with different toolkits I finally found the
perfect one: Tk. On Windows Tk looks nice and native. In fact,
wxWidgets had some very un-Windows looking widgets compared to Tk. So
I started using Tk with C (back then I didn't take Tcl seriously).
A couple of weeks later I realised that my Tcl prototypes had more
features, was easier to maintain and had fewer bugs than my "final" C
code. I dumped the buggy C code, cleaned up the Tcl version and
shipped it as a freewrapped exe.
Ian Bell
Strangely enough, at about the same time I had been using gtk for a couple
of years, and becoming increasingly frustrated, when I found Tcl/Tk. I
couldn't believe the its sheer elegance of the language not to mention the
incredibly easy GUI building facilities.
A year ago I 'fell out' with Tcl/Tk when it looked like it was going to be
OO'ed to death. I took some time out to look OOP in the eye (again), read
several C++ books and reviewed a whole bunch of C++ widget sets. Needless
to say I am back with Tcl/Tk.
Ian
Kevin Walzer
toolkits and language combinations (Python, Tcl, Apple's Cocoa
frameworks, AppleScript), and have found that Tcl/Tk is far and away the
best combination for ease-of-use and power. Tk does need to be extended
beyond the base widget set for a really nice GUI, but with Tile,
BWidgets, Tablelist, and others, everything you need is there.
--
Kevin Walzer
Code by Kevin
http://www.codebykevin.com
Shin
> http://freshmeat.net/articles/view/928/
Call me Granpa! - tk_bisque rulez!
Actually I don't give a dime on all this eye-candy fuzz. I'd rather
like to see the main stream recognizing standards instead - Menubar on
top, scrollbars on the right, context menus on mousebutton 3 and
scalable fonts.
Shin
Uwe Klein
TkSwitchers!
in 1992 i bought from Motorola a complete 88.000 RISC unix System
( very nice VME Desktop System with a MVME187 CPU Board and
lavish 512MB of SCSI Disk space )
heavily pushed into the market at that time and thus _a lot_
cheaper than the 68040 MVME167 based one i would have preferred.
With the X Addon package came the complete X Documentation,
about 3 feet of bookshelf whose perusal ascertained me
that graphics programming was not for me.
better stay with data aquisition, distributed computing
and realtime stuff.
Years later i found Osterhouts book on tcltk.
Instant graphix!, Horray!!
After another ten years most of the utilities i write
are in tcl, only heavy bitmangling and the realtime stuff
is still in C.
uwe
Dav...@gmail.com
> http://freshmeat.net/articles/view/928/
This makes me wonder. Is it possible to setup a tk app such that an
end user can drag the widgets around while the program is running?
Robert Heller
Actually, yes. That is what xf is/was.
>
>
--
Robert Heller -- 978-544-6933
Deepwoods Software -- Linux Installation and Administration
http://www.deepsoft.com/ -- Web Hosting, with CGI and Database
hel...@deepsoft.com -- Contract Programming: C/C++, Tcl/Tk
Bryan Oakley
Ian Bell
Yes, my experience pretty much parallels this. You have summarised it well.
ian
sp...@controlq.com
On Mon, 5 Feb 2007, Ian Bell wrote:
> Yes, my experience pretty much parallels this. You have summarised it well.
>
> ian
<.02$>
It seems that everyone and their brother has an urge to develop a
C++ based portable GUI toolkit. (Oh and it occurs that one notable
ommission from the OP's article on C++ tools, was Ultimate++.)
http://www.ultimatepp.org/www$uppweb2$overview$en-us.html
In general, I too would agree about the merits of Tk (portability, ability
to use ANSI C, simplicity etc), and given the Tile approach which
separates form (look and feel) from function, it gets even better!!! But
at the risk of being too defensive, and while Tk is known and loved by us
all 8-), we should not wear blinders ... particularly when it comes to
modern GUI development.
Indeed, one GUI builder/approach which I think which deserves some
mention, is the Glade tool, and the libglade library. Glade is an
interface editor, which allows you to graphically build an interface on
top of GTK+, and then generates the widget description in XML. The
libglade library then dynamically creates the interface at runtime from
the .xml file, and can be linked to ANSI C, and therefore just about any
language or scripting tool. This is IMHO a very good approach to modern
development techniques, and bears some scrutiny. I have not played with
glade yet, but I plan to, and of particular note will be how well they
have separated function and appearance, as well as callbacks/signals.
In general, developers care more about function, whereas
users/suits/marketing types like the eyecandy. In an ideal world, one
would simply program up an app, and then have the marketing types add the
frosting to the eyecandy in order to appeal the unwashed masses ...
(theme-ing? also relates to the .CSS thread on this forum)
Having said this ... tk/Tile style widgets are *VERY* nice/useful and
conceptually easy to both understand and use effectively. Having 3
(count'em) geometry managers (and frames as well) is also much appreciated
when you look at the approach taken other systems (simple frame based
geometries). While it is abundantly clear that Tile widgets seem to be a
step in the right direction ... have we gone far enough??
One final note ... while Win/Linux/Unix/OSX has been the lingua franca for
target platforms to date ... more and more PDA's (Palm/Linux/WINCE) and
cell phones are viable application platforms, and it would be very nice to
have 1 platform to take everywhere (but now I'm ranting).
</.02$>
Cheers,
Rob.
----== Posted via Newsfeeds.Com - Unlimited-Unrestricted-Secure Usenet News==----
http://www.newsfeeds.com The #1 Newsgroup Service in the World! >100,000 Newsgroups
---= East/West-Coast Server Farms - Total Privacy via Encryption =---
EKB
I completely agree with everything you say! Except... bisque. No, I
have seen enough bisque to last me a lifetime. :-) And actually, color
is another standard that I try to stick to -- I try to get the color
set by the user for their environment, and use transparent backgrounds
on images. But color is always tricky, and I've still got room to
improve.
sleb...@gmail.com
Download Visual Tcl (http://vtcl.sourceforge.net/) and place and drag
widgets live at runtime! Visual Tcl is a self-contained, pure-tcl
application without dependencies on external packages. So yes, it can/
has been done.
sleb...@gmail.com
> On Feb 6, 3:07 am, "Davo...@gmail.com" <Davo...@gmail.com> wrote:
>
> > On Feb 5, 6:39 am, "call_me_anything" <s...@yahoo.com> wrote:
>
> > >http://freshmeat.net/articles/view/928/
>
> > This makes me wonder. Is it possible to setup a tk app such that an
> > end user can drag the widgets around while the program is running?
>
# In fact, here's a simple example:
place [label .l -text "This is a label"] -x 0 -y 0
place [entry .e] -x 0 -y 20
set px 0
set py 0
proc handleClick {win} {
global px py
set px [winfo pointerx .]
set py [winfo pointery .]
bind . <Motion> [list handleMotion $win]
}
proc handleMotion {win} {
global px py
set x [winfo pointerx .]
set y [winfo pointery .]
place configure $win \
-x [expr {[winfo x $win]+$x-$px}] \
-y [expr {[winfo y $win]+$y-$py}]
set px $x
set py $y
}
proc handleLetGo {win} {
bind . <Motion> {}
focus $win
}
foreach x {.l .e} {
bind $x <Button-1> [list handleClick $x]
bind $x <ButtonRelease> [list handleLetGo $x]
}
# notice that the entry widget is actually editable!
Donal K. Fellows
> A year ago I 'fell out' with Tcl/Tk when it looked like it was going to be
> OO'ed to death. I took some time out to look OOP in the eye (again), read
> several C++ books and reviewed a whole bunch of C++ widget sets. Needless
> to say I am back with Tcl/Tk.
Heh. I know the feeling well. It's why I for one don't want people to
have to use any OO unless it is helping people make their code better.
Guess that marks me as an anti-purist...
Donal.
Dan Smart
>
> Indeed, one GUI builder/approach which I think which deserves some
> mention, is the Glade tool, and the libglade library. Glade is an
> interface editor, which allows you to graphically build an interface on
> top of GTK+, and then generates the widget description in XML. The
> libglade library then dynamically creates the interface at runtime from
> the .xml file, and can be linked to ANSI C, and therefore just about
> any language or scripting tool. This is IMHO a very good approach to
> modern development techniques, and bears some scrutiny. I have not
> played with glade yet, but I plan to, and of particular note will be
> how well they have separated function and appearance, as well as
> callbacks/signals.
This is one of many examples of "scripting the creation of your GUI in
XML", which if you are stuck using a compiled language for your user
interface must seem like a realLy good idea. But a *much* better idea
is to script the creation of your GUI in something that *isn't* xml[1],
like for instance, TCL/Tk. Then you see TCL/Tk dynamically creates the
interface at runtime, and it can be linked to ANSI C (and numerous
other languages as well). :-)
[prunage]
>
> One final note ... while Win/Linux/Unix/OSX has been the lingua franca
> for target platforms to date ... more and more PDA's (Palm/Linux/WINCE)
> and cell phones are viable application platforms, and it would be very
> nice to have 1 platform to take everywhere (but now I'm ranting).
Hmm...
Well TCL/Tk doesn't run on the iPhone yet (as far as I know), but it's
certainly running on some Linux PDA's, I know it's runnning on
WinCE/"Windows Mobile", but I'm not sure of the status of the Palm port
any more. So, not quite total coverage, but getting pretty close.
Dan "Just passing by" Smart
[1] Well actually pretty much anything that *isn't* XML is probably a
better idea than just about anything that is :-)
sleb...@gmail.com
> Hmm...
> Well TCL/Tk doesn't run on the iPhone yet (as far as I know), but it's
> certainly running on some Linux PDA's, I know it's runnning on
> WinCE/"Windows Mobile", but I'm not sure of the status of the Palm port
> any more. So, not quite total coverage, but getting pretty close.
>
Jacl is very Tcl-like and runs on my mobile phone (actually, anywhere
where Java runs). Unfortunately it uses horrid Java forms for UI, not
Tk :-(
Ian Bell
I'd say it marks you as one of the few who realise OO is not a panancea.
Ian
Don Porter
> But at the risk of being too defensive, and while Tk is known
> and loved by us all 8-), we should not wear blinders ... particularly
> when it comes to modern GUI development.
> While it is abundantly clear that Tile widgets
> seem to be a step in the right direction ... have we gone far enough??
Who is "we" ?
From where I sit, being as generous as I can, there are no more than
five people developing Tk. You're not one of them. Netiher is the
author of the "must read" article. Neither is the originator of this
thread who called it to our attention. Neither is anyone replying to
this thread, including me.
Maybe those who actually do the development work have more patience than
I do, but I would find all these pronouncements about whether their
efforts have gone far enough to be quite annoying.
--
| Don Porter Mathematical and Computational Sciences Division |
| donald...@nist.gov Information Technology Laboratory |
| http://math.nist.gov/~DPorter/ NIST |
|______________________________________________________________________|
sp...@controlq.com
> sp...@controlq.com wrote:
>> But at the risk of being too defensive, and while Tk is known and loved by
>> us all 8-), we should not wear blinders ... particularly when it comes to
>> modern GUI development.
> ...
>> While it is abundantly clear that Tile widgets seem to be a step in the
>> right direction ... have we gone far enough??
>
> Who is "we" ?
>
> From where I sit, being as generous as I can, there are no more than
> five people developing Tk. You're not one of them. Netiher is the author of
> the "must read" article. Neither is the originator of this thread who called
> it to our attention. Neither is anyone replying to this thread, including me.
>
> Maybe those who actually do the development work have more patience than I
> do, but I would find all these pronouncements about whether their efforts
> have gone far enough to be quite annoying.
No offense intended to the core group, the spirit of my post was not to be
critical, but rather to provoke thought. Far from impugning the efforts
of those who work tirelessly on Tcl/Tk, much credit is due them, I am only
asking a question or two. Where you suggest I was being critical, I
sought only to recognize what is happening in the area of *OTHER*
development tools, and where Tcl/Tk could be improved, not so much by
effort, but as by approach.
Further when I said "we", I consider myself as a frequent user of tcl/tk
on various platforms, to be part of the greater community of tcl users,
and so in spite of not having the ability to spend time on open source
development (owing to the scarce time available to a contract consultant),
I feel that I do indeed have a stake in the future of the tool. You imply
that I cannot consider myself a part of the Tcl community, and that my
opinions do not matter. I must respectfully disagree.
As it happens, the OP struck a chord as I *AM* just now looking ahead at
my own needs for a OS independent development platform, and of late, I've
been reviewing some of those same alternatives reviewed in the article,
and more (Tile/tk, wxWindows, Ultimate++, QT, Fltk, GTK+/Glade and various
other libraries both commercial and open source). This is healthy, IMHO,
for my future, and in the process, I might just learn a thing or two about
how different approaches might improve my own efforts.
As it turns out, I did not design and implement Tile/tk, but I'm very glad
that some very gifted and dedicated individuals did. I am using it to
good effect, and likely will for the forseeable future. While I am indeed
grateful for the opportunity to use it, every language and indeed every
software application has a natural evolution, and should change over time
to reflect the community it serves, and their changing needs.
When I said that it was "abundantly clear that Tile widgets seem to be a
step in the right direction" I meant just that. It is a great first step
towards modernizing Tk. Hopefully it will not be the last one "we" take.
So, to answer your question, "we" is me, and you and *ALL* of those who
use Tcl/Tk on a regular basis to do real work.
Keith Nash
>> "Please note that I won't cover Motif/Lesstif in this article because
>> it is similar to Xlib in terms of age and complexity. Tcl/Tk isn't
>> much better."
>>
>> Age: need not be a bad thing (maybe he means the pre-Tile looks of
>> Tk).
>
> functionality.
>>
>> Complexity: is any of the reviewed toolkits simpler to use (at same
>> power) than Tk?
It would be good to write some counter-propaganda for release when 8.5.0 is
ready.
Keith.
Bryan Oakley
Robert Heller
Actually, yes. That is what xf is/was.
Donal K. Fellows
> A year ago I 'fell out' with Tcl/Tk when it looked like it was going to be
> OO'ed to death. I took some time out to look OOP in the eye (again), read
> several C++ books and reviewed a whole bunch of C++ widget sets. Needless
> to say I am back with Tcl/Tk.
Heh. I know the feeling well. It's why I for one don't want people to