Organization of GUIs
I have a question about typical organization of GUIs. I will be using PyQt. I have mostly used Python and C++ in my professional life, but I just took an 8 month detour into using a functional programming language called Haskell. Haskell is pure meaning that for the most part data is not mutable, or if it is mutable, access to it is controlled very carefully. This got me thinking about how mutable data can make program behavior complicated and difficult to understood/prove- correct. I am returning to Python and PyQt for my next project, but I come away inspired to do things a little differently. For example, in GUIs I've written in the past, typically there are a lot of GUI objects (windows, data repositories, etc.) that talk to each other. One object might send messages about a change in its state to other objects that are watching it. I am now wondering if I should write a GUI so that everything is in a true hierarchy, rather than a tangle of objects with democratic relationships---and more specifically, that messages (which may cause state to be changed in the receiver of the message) should first go up the hierarchy until they reach the right level, and then down to the ultimate receiver of the message. This way a parent object in the hierarchy has complete visibility and control of messages passing between its children, leading to the possibility of consolidating the code and documentation about the children's mutability in one place (the parent). But honestly, I have never really studied how good existing GUI programs are organized. Always flew by the seat of my pants. So I'm interested in someone has thoughts about this, maybe can point me to a good book, etc. Thanks, Mike -- http://mail.python.org/mailman/listinfo/python-list
Re: Organization of GUIs
On Dec 3, 6:38 am, Lie Ryan lie.1...@gmail.com wrote: On 12/4/2009 12:44 AM, Michael Mossey wrote: I have a question about typical organization of GUIs. I will be using PyQt. Model-View-Controller (MVC) pattern. Model - all the business logic lives in the model. View - your GUI Controller - Takes input Controller notifies Model if there is user input; Model notifies View if there is an update in the model; View notifies user if there is an update in the model; User notifies controller of the changes wanted. http://en.wikipedia.org/wiki/Model%E2%80%93view%E2%80%93controller I'm aware of Model-View-Controller, but my question is broader than that for a couple reasons: View can be fine-grained. Often the View consists of a number of GUI objects. Some people write this in a democratic arrangement---they all talk to each other. This can make analyzing system behavior complicated. Hence my proposal for a hierarchy. Also, often different parts of the Model talk to different parts of the View. The question, then, is whether both the Model and View should be structured hierarchically so that all messages pass through a master or root object before going to the other Model/View, and messages within a Model or View should only move along the hierarchy. In other words, you can easily make a complete mess of MVC. -- http://mail.python.org/mailman/listinfo/python-list
Re: help with threads
On Aug 7, 5:03 pm, Piet van Oostrum p...@cs.uu.nl wrote: Michael Mossey michaelmos...@gmail.com (MM) wrote: MM Ah yes, that explains it. Some of these long computations are done in MM pure C, so I'm sure the GIL is not being released. Is that C code under your own control? Or at least the glue from Python to C? In that case, and if the C code is not manipulating Python objects or something in the Python interpreter, it could be changed to release the GIL during the computation. That's also how Numpy does it, IIRC. -- Piet van Oostrum p...@cs.uu.nl URL:http://pietvanoostrum.com[PGP 8DAE142BE17999C4] Private email: p...@vanoostrum.org I don't have any control over this code, and it's easier to solve my problem in other ways. I just put a sleep() call between calls to the C library, and that gives the network thread enough responsiveness for my particular task. I am grateful, anyway, to understand why this kind of thing happens. -- http://mail.python.org/mailman/listinfo/python-list
Re: help with threads
Ah yes, that explains it. Some of these long computations are done in pure C, so I'm sure the GIL is not being released. Thanks. -- http://mail.python.org/mailman/listinfo/python-list
help with threads
Hello, I have a simple application that needs one thread to manage networking in addition to the main thread that does the main job. It's not working right. I know hardly anything about threads, so I was hoping someone could point me in the right direction to research this. Basically, I have a program that does some computational work, and also conveys its status to a monitor program elsewhere on the network via sockets. I wanted to use a thread to manage the networking so that the main program can run without regard to networking (i.e. they would be asynchronous). So the network thread loops and calls select. My problem is that in some cases, the network thread appears to stop, while the main thread is doing a long computation. I'm hoping someone can give me a general idea what to read about. For example, under what conditions does a thread stop running? Can other threads take priority? Are there certain operations that block other threads (such as disk access)? Thanks, Mike -- http://mail.python.org/mailman/listinfo/python-list
Catching control-C
What is required in a python program to make sure it catches a control- c on the command-line? Do some i/o? The OS here is Linux. Thanks, Mike -- http://mail.python.org/mailman/listinfo/python-list
Re: Catching control-C
On Jul 6, 2:47 pm, Philip Semanchuk phi...@semanchuk.com wrote: On Jul 6, 2009, at 5:37 PM, Michael Mossey wrote: What is required in a python program to make sure it catches a control- c on the command-line? Do some i/o? The OS here is Linux. You can use a try/except to catch a KeyboardInterrupt exception, or you can trap it using the signal module:http://docs.python.org/library/signal.html You want to trap SIGINT. HTH Philip Thanks to both of you. However, my question is also about whether I need to be doing i/o or some similar operation for my program to notice in any shape or form that Control-C has been pressed. In the past, I've written Python programs that go about their business ignoring Ctrl-C. Other programs respond to it immediately by exiting. I think the difference is that the latter programs are doing i/o. But I want to understand better what the secret is to responding to a ctrl-C in any shape or form. For example, does trapping SIGINT always work, regardless of what my process is doing? Thanks, Mike -- http://mail.python.org/mailman/listinfo/python-list
