Problem running examples...
System dynamics program with additional features for economics
Brought to you by:
hpcoder,
profstevekeen
When i try and load an example in the Win1D9 version i do not see any diagram of the model appearing, nothing, blank canvas. I thought it would show something?
If i click on Run it merrily consumes CPU (i let it run a model for an hour while i watched TV :) but there's nothing to show for it!
How do i retrieve what i seem to be missing: e.g. the diagram of the model being run, its output etc. ?
There should be a diagram showing. If the diagram is empty, then clicking run will just be running an empty model - using lots of CPU doing nothing, as you say.
The question is why the model didn't load. Was this one of the examples from the 1.D9 distribution, or did it come from somewhere else?
That's what I thought, that a diagram should be showing. Yes, it is one of the examples that came with the distribution, i have tried several of them, same problem. Perhaps if i ran the .exe from the command line it might produce some messages....no, didn't help.
I did the install on a different machine, same problem, no diagram. (Win 7, version 1.D9)
If i create a model by dragging a couple of items onto the canvas and then save it, when i reload it it comes up blank so the issue is consistent.
The mystery continues...
I have verified that 1.D9 behaves as you say on Win 7. This is bizarre, as I'm sure Steve Keen checked this version on Win 7 - when I see him next, I'll ask about his set up. It definitely would have been checked on Windows XP Home.
However, the latest beta build seems to be fine, so you can at least try this out - see http://www.hpcoders.com.au/Minsky.1.2.D34-win-dist.tar.gz
You still need to run install.bat as administrator from the command line.
I'll discuss with Steve about going to an early release of Aquinas.
I tried the beta version, first uninstalled 1.D9 and then installed 1.2D34, but it has the same problem, no diagram appears.
Also in this beta build the canvas seems a bit buggy: when you add items to the canvas they appear but when you attempt to drag them they disappear (Godley table and graph are ok in this regard).
It would be good to add an "About" on the main window to show what version is being run.
Not sure what's going on with your computer.
I am now adding stuff to the Window distribution build target to create an msi file, hopefully this will resolve all the Win 7 stuff.
I always meant to do this, the install.bat idea was just to get something working (works fine on XP :).
I'll let you know when its done, and you can take it for a spin on your system.
Proper msi installers are now available for download as Minsky.1.D11. This is still a version of the Aristotle release, so is feature identical to 1.D9, other than having an installer.
Please let me know how this works. I tested it on both XP Home and Win 7, and it all looks good.
If you still have blank canvases, please let me know what else you might have on your system, and I'll try to isolate the problem.
I also just installed latest version 1.D011 (using msi installer) on my Win 7 system and have the same problem - still a blank screen when loading examples. Als,o no zoom buttons?
Can create a model and save it, but can't load and run it though ???
Zooming is a feature of the Aquinas release, whereas 1.D11 is an Aristotle release. Aquinas is very nearly ready for release.
Were you running the examples that came with the release, or examples from somewhere else? There is a major schema change with Aquinas. The way to tell is to open the .mky file in a text editor. If the file starts with the line "<root>", then is is a version 0 schema, and Aristotle should be able to read it. If it starts with "<Minsky", then it is version 1, and Aristotle (1.D11) will not be able to read it.
If you'd like a copy of the latest Aquinas build for Windows, I'm happy to send it to you
I'm trying the examples that came with Minsky.1.D11-win-dist.zip.
I tried looking at an .mky file with notebook, its full of things like
<root>
<stockVars>
<double>0</double>"
etc. (this is the start of the file "UWS05.mky")
- there seems to be no "" or "<Minsky" at the start.
[ I note that the 'things' (characters within "<" and ">" symbols) I mention above don't appear in this listing of my comments - but they do when you edit this message. I guess its a Markdown feature.]
Last edit: Brian S. 2013-02-11
That is correct. 1.D11 only supports Minsky schema version "0", which
starts with "<root>", but is otherwise only defined by the Minsky
source code. Minsky schema 1, which is documented in
http://minsky.sf.net/minsky, is due to arrive with the Aquinas
release. Aquinas, will (cross my heart) be able to read schema 0 files
for backward compatibility, as well as read/write schema 1 files.
Cheers
On Mon, Feb 11, 2013 at 09:54:54PM +0000, Brian S. wrote:
--
Prof Russell Standish Phone 0425 253119 (mobile)
Principal, High Performance Coders
Visiting Professor of Mathematics hpcoder@hpcoders.com.au
University of New South Wales http://www.hpcoders.com.au
I solved my 'problem'. I think it's a Windows feature - I had copied the examples to my Minsky directory (C:/Program Files/Minsky) and was trying to run them from there.
I guess Windows doesn't like that - when I tranfsferred the 'Examples' directory to another disk (i.e. NOT the C:/ drive) it works fine.
Maybe running as administrator would have helped?
Thanks for your help.
I had version D9 on my Win7 desktop computer and it worked fine (from a clean install--i.e. no previous versions on the computer ever). I just upgraded to D11 and it seems to work without error (models show up without fault, including the new "group" feature).
That said, I do have a problem getting any version above 1.D26 working on my laptop (IIRC...the really old one. I don't see it listed anymore; it may have come from Prof. Keen's website). I suspect something might be in the registry from the old version that is preventing D9 or D11 from opening properly. But I don't know for sure. I've tried deleting the old one and even poking through the registry, but can't find what is preventing the newer versions from installing/opening properly.
As an aside, I see Prof. Keen is able to input trigonometric functions into the models, but D11 doesn't seem to have that feature. Is the trig input a part of "Aquinas"?
I would appreciate receiving it if possible.
As a follow-up, I just downloaded and installed the "binaries" (1.2.D088) from Prof. Keen's website (for windows) on my desktop computer. Now I have the trig functions... is this the Aquinas version? Maybe this version (1.2.D088) will fix my laptop problems?
1.2.D88 is a release candidate for Aquinas, release to beta testers for testing purposes. You're welcome to try it and put it through its paces. Obviously, it is very nearly ready, as Steve has been making it available from his website.
You're also welcome to join the beta testers programme - sign up at https://lists.sourceforge.net/lists/listinfo/minsky-betatesters
I'm curious to know what you can find out about the version not working on your laptop. Are you seeing the same symptoms as kaiteriteri?
I have joined the beta testers programme ;-)
As for the laptop (Win7), no, I am not seeing the same symptoms. It appears that since I had that D26 version (which worked), any updated version is prevented from opening. They seem to install without hassle, but when I go to open them (from desktop icon) all I get is a quick flash of a window trying to open and then it immediately closes. The D88 Beta is also doing the same thing (so there goes that theory).
No error messages or anything. It just seems to crash. That's why I was wondering if there may have been some conflict in the registry. I did see some "Minsky" entries that had no apparent targets or values. I was thinking I should just go through the entire registry and delete everything that is associated with it, and try reinstalling from there.
On Mon, Feb 11, 2013 at 01:37:19AM +0000, heckler73 wrote:
Be careful to make sure you have a valid Windows checkpoint before
doing that. It is to easy to trash your entire system, if you just go
willy-nilly deleting Registry entries.
Another backup option if you have the Windows media disk, is to repair
an operating system (available under the install menu). AT least that
true with XP - I haven't tried that with Win7.
Another thing you can do - I'll create a special debug version of
Minsky for you to try and install. The debug version has a console, so
that error messages can be visible. Its a pity that Windows is so
brain dead that you can't get console I/O by running from the
command line - instead you must compile it with console support,
meaning there is an ugly black window around all the time.
Cheers
--
Prof Russell Standish Phone 0425 253119 (mobile)
Principal, High Performance Coders
Visiting Professor of Mathematics hpcoder@hpcoders.com.au
University of New South Wales http://www.hpcoders.com.au
Hi heckler73,
I have create a "DEBUG" version of Minsky.1.2.D88, which you should be able to install. The advantage is, that this one will display error messages, eg from missing DLLs.
Its available at
http://www.hpcoders.com.au/Minsky.1.2.D88-win-dist-DEBUG.tar.gz
You may need to run it from the command line - ie from cmd.exe, if you have nothing better, as if you double click on it, the app may well close before you can read the message.
You can also run the exe in place (ie without installing it), by running the executable located in the Minsky folder of the distribution.
I'm keen to here what find out!
Cheers
That is what comes out... I installed the version (after unpacking it), opened a cmd console, went to its location and ran it. Running it from the location on E:\ (where I unpacked it) gives the same result--with the exception of the last line.
This is very helpful.
In the meantime, you should be good to go if you simply delete the file
C:\Users\Finn/.minskyrc
and then try not top open any examples from a directory containing "$"
in its path.
It looks like I have to be careful about the possibility of people
putting $ in their filenames.
Cheers
On Mon, Feb 11, 2013 at 05:57:06AM +0000, heckler73 wrote:
--
Prof Russell Standish Phone 0425 253119 (mobile)
Principal, High Performance Coders
Visiting Professor of Mathematics hpcoder@hpcoders.com.au
University of New South Wales http://www.hpcoders.com.au
AHA! It worked like a charm...
And I have one of the example models from Steve open and it's working just fine...although it does start to slow down as the model runs. I presume that is because I am on a Netbook, so resources are going to become limited in a hurry. Nevertheless, it seems to be working under a loaded system (browser open with videos playing, etc) and no crashes!
Thanks again, now I can go learn how to use this while watching Prof. Keen's tutorials at the same time in comfort.
Cheers to you, good sir!