What improvements would you like to see in Vensim?
-
- Super Administrator
- Posts: 4667
- Joined: Wed Mar 05, 2003 3:10 am
What improvements would you like to see in Vensim?
If anyone has any ideas on what they would like to see in Vensim, please let us know.
Also if there is something that you feel could be made better, also reply to this and let us know your thoughts.
If you have a question, please start a separate thread - we can't easily respond here.
Thanks in advance,
Tony.
UPDATE:
Help shape the roadmap for future Vensim versions - take our user survey:
http://vensim.com/survey.html?utm_sourc ... n=VenForum
Also if there is something that you feel could be made better, also reply to this and let us know your thoughts.
If you have a question, please start a separate thread - we can't easily respond here.
Thanks in advance,
Tony.
UPDATE:
Help shape the roadmap for future Vensim versions - take our user survey:
http://vensim.com/survey.html?utm_sourc ... n=VenForum
Advice to posters seeking help (it really helps us to help you)
http://www.ventanasystems.co.uk/forum/v ... f=2&t=4391
Units are important!
http://www.bbc.co.uk/news/magazine-27509559
http://www.ventanasystems.co.uk/forum/v ... f=2&t=4391
Units are important!
http://www.bbc.co.uk/news/magazine-27509559
-
- Senior Member
- Posts: 107
- Joined: Wed Nov 26, 2008 6:12 am
Re: What improvements would you like to see in Vensim?
When viewing a graph, it would be helpful to be able to tell the exact value of the variable of concern as one scrolls along the independent axis. This would save an intermediary step of having to open up a table.
Re: What improvements would you like to see in Vensim?
Hi everybody
Karan has a good idea about the graph showing the exact values.
I have too the same kind of problem using reality checks.
The RC control panel gives the possibility to view graphs but unfortunately it would be often useful to have the exact value of the graphs.
In particular when one chooses to test a specific RC with the sim active option, the results of the run is saved, but unfortunately not the value at the right side of the contraints. The only thing that one sees about that right side is on a graph. It is possible to get the value but one must define a new RC with the right side of the constraint as a test input of the new RC which is very cumbersome. The best solution would be to have available a table option right in the RC control panel or at least to have the numerical values of the graph available.
I have totally changed the way I build, use and analyse Vensim models very recently. With the old diagram based methods, I think that Vensim is a very mature product and I had no technical problems. With the new method which is based on reality checks, I am forced to use Vensim very differently, use new tools that I did not use precedently or use old tools differently. It is possible that in the future it may generate new unsatisfied needs about the software.
The important thing for me is that the reality checks tools remain available in new future releases, as it seems surprisingly very seldom used.
Best regards.
JJ
Karan has a good idea about the graph showing the exact values.
I have too the same kind of problem using reality checks.
The RC control panel gives the possibility to view graphs but unfortunately it would be often useful to have the exact value of the graphs.
In particular when one chooses to test a specific RC with the sim active option, the results of the run is saved, but unfortunately not the value at the right side of the contraints. The only thing that one sees about that right side is on a graph. It is possible to get the value but one must define a new RC with the right side of the constraint as a test input of the new RC which is very cumbersome. The best solution would be to have available a table option right in the RC control panel or at least to have the numerical values of the graph available.
I have totally changed the way I build, use and analyse Vensim models very recently. With the old diagram based methods, I think that Vensim is a very mature product and I had no technical problems. With the new method which is based on reality checks, I am forced to use Vensim very differently, use new tools that I did not use precedently or use old tools differently. It is possible that in the future it may generate new unsatisfied needs about the software.
The important thing for me is that the reality checks tools remain available in new future releases, as it seems surprisingly very seldom used.
Best regards.
JJ
Re: What improvements would you like to see in Vensim?
Hi everybody
Another point:
When on uses the systhesim and a slider, the only way possible to change the value of the parameter is to move the cursor.
If instead of using a slider one uses a shadow variable of the parameter whose value one wants to change, it is possible to change the value with the slider underneath the shadow variable using the left and right arrow so as to change the value incremetally which is often necessary. It is impossible to change the value little step by little step with the current input slider. Practically I use an input slider and I put underneath the shadow variable to be able to move the constant's value step by step if I want to. Would it be possible to do the same with the input slider only?
Regards.
JJ
Another point:
When on uses the systhesim and a slider, the only way possible to change the value of the parameter is to move the cursor.
If instead of using a slider one uses a shadow variable of the parameter whose value one wants to change, it is possible to change the value with the slider underneath the shadow variable using the left and right arrow so as to change the value incremetally which is often necessary. It is impossible to change the value little step by little step with the current input slider. Practically I use an input slider and I put underneath the shadow variable to be able to move the constant's value step by step if I want to. Would it be possible to do the same with the input slider only?
Regards.
JJ
-
- Senior Member
- Posts: 107
- Joined: Wed Nov 26, 2008 6:12 am
Re: What improvements would you like to see in Vensim?
Greetings!
I would also like to see a Stock "non-negative" feature, specifically for Vensim PLE; often, in my short introductory workshops and courses, students at this early stage of learning get lost with MIN, MAX, non-linear formulations for flow control.
My Best,
Karan
I would also like to see a Stock "non-negative" feature, specifically for Vensim PLE; often, in my short introductory workshops and courses, students at this early stage of learning get lost with MIN, MAX, non-linear formulations for flow control.
My Best,
Karan
Re: What improvements would you like to see in Vensim?
Hi Karan
What do you mean by stock non negative feature?
Forcing the stock to be equal to 0, when it is calculated to be negative? It will generate wrong calculations.
Forcing an output rate to - stock / time step? But what if there is an input and an output? Which rate will be modified, the input or the output?
Generally the output, but in some cases it might be both.
Or simply a notification by the software that a stock has become negative?
In that case, there should be an option telling the software, whether it accepts to run the model with a warning, or best option stops the calculations when a stock becomes negative as with an overflow, showing a run time error.
You can presently use the the reality check : Time >= 0 :implies: stock >= 0 but it is not automatic, students having to build an equation for each stock and press Ctrl-U.
Reality checks are available with the PLE version.
Regards.
JJ
What do you mean by stock non negative feature?
Forcing the stock to be equal to 0, when it is calculated to be negative? It will generate wrong calculations.
Forcing an output rate to - stock / time step? But what if there is an input and an output? Which rate will be modified, the input or the output?
Generally the output, but in some cases it might be both.
Or simply a notification by the software that a stock has become negative?
In that case, there should be an option telling the software, whether it accepts to run the model with a warning, or best option stops the calculations when a stock becomes negative as with an overflow, showing a run time error.
You can presently use the the reality check : Time >= 0 :implies: stock >= 0 but it is not automatic, students having to build an equation for each stock and press Ctrl-U.
Reality checks are available with the PLE version.
Regards.
JJ
Re: What improvements would you like to see in Vensim?
Hi everybody
I have two other problems that I would like to adress if possible.
First: if one forces a constant to a determined behavior in a test input of the reality checks by example: constant = RC decay(constant,decay time,rc start time) with decay time = 3 months and rc start time = 10 months for instance and if you run the check highlighting only the check you want to test to save the run, the value that takes the constant is not saved and generally you get only one value instead of a value for each period. This is rather annoying if you want to know exactly the values taken by the constant.
It would be better that in this case, all the values taken by the constant be shown as it is done with the synthesim if one overrides a constant with a specific non constant value. In synthesim if one save the run all the values taken by the constant are saved and shown.
I generally solve the problem by creating a new variable constant_name_var = constant_name.
If one changes the value of the constant, the variable constant_name_var is kept and all the values are show on a graph or table. But it is rather cumbersome even if one hides all the alias variables.
Second problem still with reality checks. If one uses the same test input with different constraints and if one wants to test one specific RC, even if one highlights the sim active button, all the constraints concerned that have the same test inputs are checked at the same time.
I avoid the problem by not using a test input and writing the same equation in the test input of each constraints. This way only the specific constraint gets checked and one is sure that the run saved concerns that specific check.
Best regards.
JJ
I have two other problems that I would like to adress if possible.
First: if one forces a constant to a determined behavior in a test input of the reality checks by example: constant = RC decay(constant,decay time,rc start time) with decay time = 3 months and rc start time = 10 months for instance and if you run the check highlighting only the check you want to test to save the run, the value that takes the constant is not saved and generally you get only one value instead of a value for each period. This is rather annoying if you want to know exactly the values taken by the constant.
It would be better that in this case, all the values taken by the constant be shown as it is done with the synthesim if one overrides a constant with a specific non constant value. In synthesim if one save the run all the values taken by the constant are saved and shown.
I generally solve the problem by creating a new variable constant_name_var = constant_name.
If one changes the value of the constant, the variable constant_name_var is kept and all the values are show on a graph or table. But it is rather cumbersome even if one hides all the alias variables.
Second problem still with reality checks. If one uses the same test input with different constraints and if one wants to test one specific RC, even if one highlights the sim active button, all the constraints concerned that have the same test inputs are checked at the same time.
I avoid the problem by not using a test input and writing the same equation in the test input of each constraints. This way only the specific constraint gets checked and one is sure that the run saved concerns that specific check.
Best regards.
JJ
Re: What improvements would you like to see in Vensim?
It would be helpful for me if external function integration was simplified - particularly for cases where loops (such as 'do while') are needed
Phil
Phil
Re: What improvements would you like to see in Vensim?
I think there is too much hidden in the 'Control Panel'. In particular, both datasets and graphs could be separate icons. When I first use vensim after not using it for a while, it always takes me some time to remember where to find these functions.
Re: What improvements would you like to see in Vensim?
As Tom adviced it, it would be nice to have a get vdf data and get vdf constants that deliver a provided value if nothing is available, whatever the reason: no existing file, or no existing variable in the file. This should need an additional parameter to both functions.
Regards.
JJ
Regards.
JJ
Re: What improvements would you like to see in Vensim?
It would be interesting if it was possible to run reality checks from a script file or from an external program like Excel VBA. I know that RC are not compiled, but it should be possible at least from a script file.
I have to launch reality checks to generate a base run to be compared with a second run and if I change the parameters of my model, I have to change both the base run and the base run with the first base modification and this is rather cumbersome, if it is not possible to make it automatic.
Regards.
JJ
I have to launch reality checks to generate a base run to be compared with a second run and if I change the parameters of my model, I have to change both the base run and the base run with the first base modification and this is rather cumbersome, if it is not possible to make it automatic.
Regards.
JJ
Re: What improvements would you like to see in Vensim?
There is place in the simulation control where you can enter a comment for a specific run.
The window cannot be scrolled vertically only horizontally and you can enter as much characters than you want. But once you press the set button it only saves approximately the first 500 characters.
There should be a scrollable window and impossible to enter more than the number of characters to be saved, and preferably extend the number of 500. Commenting a run is important.
Regards.
JJ
The window cannot be scrolled vertically only horizontally and you can enter as much characters than you want. But once you press the set button it only saves approximately the first 500 characters.
There should be a scrollable window and impossible to enter more than the number of characters to be saved, and preferably extend the number of 500. Commenting a run is important.
Regards.
JJ
Re: What improvements would you like to see in Vensim?
It is impossible to change the value of a string variable to ‘’ by using the set up a simulation tool or using synthesim and changing the value of the string variable. If one erases the string value, the software shows an error ‘unable to read as a number’. If one changes to ‘’ and one launches the run it says:
ERROR: Unable to open the file ''.vdf.
ERROR: Unable to open '' to get data.
Fortunately if one saves a .cin file having changed the value of the string to something and one edits the .cin file with a text editor, and one erases the string value in the file, if one runs the model loading the .cin file modified it works!
But this obliges to make diverse manipulations using .cin file instead of simply using the set up a simulation button.
I have to change string variable to ‘’ to make it possible to use get vdf data with a ‘’ string variable as file.
Regards.
JJ
ERROR: Unable to open the file ''.vdf.
ERROR: Unable to open '' to get data.
Fortunately if one saves a .cin file having changed the value of the string to something and one edits the .cin file with a text editor, and one erases the string value in the file, if one runs the model loading the .cin file modified it works!
But this obliges to make diverse manipulations using .cin file instead of simply using the set up a simulation button.
I have to change string variable to ‘’ to make it possible to use get vdf data with a ‘’ string variable as file.
Regards.
JJ
Re: What improvements would you like to see in Vensim?
Hi
If one uses a long name for a reality check, and the test is negative for it, the time when the check was not valid or besomes valid again is not shown in the window because it does not truncate the phrase to put the end on the next line.
This is sometimes annoying, especially if one uses expressions on the right side of the equations that are not calculated if one chooses to test only one RC with the sim active button on the reality check control window and save the results.
Allowing the result to be shown entirely in the window would be helpful.
Regards.
JJ
If one uses a long name for a reality check, and the test is negative for it, the time when the check was not valid or besomes valid again is not shown in the window because it does not truncate the phrase to put the end on the next line.
This is sometimes annoying, especially if one uses expressions on the right side of the equations that are not calculated if one chooses to test only one RC with the sim active button on the reality check control window and save the results.
Allowing the result to be shown entirely in the window would be helpful.
Regards.
JJ
-
- Senior Member
- Posts: 107
- Joined: Wed Nov 26, 2008 6:12 am
Re: What improvements would you like to see in Vensim?
It would be nice if Ghost Variables could accept inputs as well, this is particularly useful when you are working across multiple views!
Re: What improvements would you like to see in Vensim?
Hi Karan
Logically if you want to see the input of a shadow variable, the solution is to change the shadow variable into a model variable which will show like any normal variable and with its inputs.
Regards.
JJ
Logically if you want to see the input of a shadow variable, the solution is to change the shadow variable into a model variable which will show like any normal variable and with its inputs.
Regards.
JJ
-
- Senior Member
- Posts: 107
- Joined: Wed Nov 26, 2008 6:12 am
Re: What improvements would you like to see in Vensim?
Hi JJ, Thanks, but this is still one step removed; would like ghost variables (except stocks) to accept inputs without additional steps.
Re: What improvements would you like to see in Vensim?
Hi Karan
If all the ghost variables accept inputs, do you want the added ghost variables to have their own inputs too? That may add a considerable overhead to all the views. I do not see what additionnal work it takes just to click on the shadow variable you want to convert into a model one. With this solution you may chose exactly up to want point you want to trace back the influences at work.
For what reasons do you want this feature implemented?
Regards.
JJ
If all the ghost variables accept inputs, do you want the added ghost variables to have their own inputs too? That may add a considerable overhead to all the views. I do not see what additionnal work it takes just to click on the shadow variable you want to convert into a model one. With this solution you may chose exactly up to want point you want to trace back the influences at work.
For what reasons do you want this feature implemented?
Regards.
JJ
-
- Junior Member
- Posts: 4
- Joined: Tue May 10, 2011 2:42 am
Re: What improvements would you like to see in Vensim?
Dear Admin,
With the free license , Vensim missed much function as ' CUMULATE and NPV..."
With student , Can you have any support for purchasing ( do researching Master ) ??
With the free license , Vensim missed much function as ' CUMULATE and NPV..."
With student , Can you have any support for purchasing ( do researching Master ) ??
Re: What improvements would you like to see in Vensim?
/*
Advice to posters (it really helps us to help you)
http://www.ventanasystems.co.uk/forum/v ... f=2&t=4391
Blog: http://blog.metasd.com
Model library: http://models.metasd.com
Bookmarks: http://delicious.com/tomfid/SystemDynamics
*/
Advice to posters (it really helps us to help you)
http://www.ventanasystems.co.uk/forum/v ... f=2&t=4391
Blog: http://blog.metasd.com
Model library: http://models.metasd.com
Bookmarks: http://delicious.com/tomfid/SystemDynamics
*/
Re: What improvements would you like to see in Vensim?
In the equation editor, I would like to see colour-coding of parentheses in some format as per other programming languages, preferably activated by the live cursor position, e.g. for passages such as the following:
IF THEN ELSE
(Lorem Ipsum < Dolor Sit Amet*Fraction of Consectetur Adipisicing Elit,
Eiusmod Tempor Switch*(Lorem Ipsum -(Lorem Ipsum *Nostrud Generation Rate)) ,
Eiusmod Tempor Switch*(Dolor Sit Amet*Fraction of Consectetur Adipisicing Elit- Dolor Sit Amet* Fraction of Consectetur Adipisicing Elit *Nostrud Generation Rate)
)
IF THEN ELSE
(Lorem Ipsum < Dolor Sit Amet*Fraction of Consectetur Adipisicing Elit,
Eiusmod Tempor Switch*(Lorem Ipsum -(Lorem Ipsum *Nostrud Generation Rate)) ,
Eiusmod Tempor Switch*(Dolor Sit Amet*Fraction of Consectetur Adipisicing Elit- Dolor Sit Amet* Fraction of Consectetur Adipisicing Elit *Nostrud Generation Rate)
)
Re: What improvements would you like to see in Vensim?
- Ability to import an image as a valve icon, for example, to elicit a shape such as the following:
- Attachments
-
- example.jpg (11.84 KiB) Viewed 3103674 times
Re: What improvements would you like to see in Vensim?
Many of the improvements that have been mentioned here are right on the spot and would make life with Vensim more enjoyable: e.g. directly reading numerical values from graphs, improvements with reality checks (JJ: they are indeed very worthy to keep) and so on.
I would like to propose a much more mundane improvement: Go with the time and make Vensim a 64bit application. What has started out as a way to make complexity tractable on computers in the 60s and 70s (aggregated stock and flow simulation) tends to become a liability nowadays when one builds bigger models using subscripts. While Vensim already surpases the restrictions of say PowerSim Studio (6 subscripts and only 500.000 elements per variable, as far as I know) models in some areas can become very complex. For example one of the more recent issues of the SD review dealt with SD in the domain of transportation modeling which naturally introduces spatial complexity next to that defined in time.
If SD is to survive in business consulting -- it seems to still remain a mostly academic endeavor -- it should not flinch from detail rich problems due to memory limitations. We all love SD for the clarity it brings with small to medium scale models sometimes refered to as 'insight models'. The day to day problems of managers go beyond this. They want to have a tool that will also adress the more intricate problems with lots of detail.
While agent-based modelling has advantages due to the way data are handled (cf. Keenan et al., Modeling General Motors and the North American Automobile Market, in: Proceedigs of the SD conference, 2004
http://www.systemdynamics.org/conferenc ... 4KEENA.pdf
-- the paper is right on target as regarding the limitations of SD and of Vensim). SD still IMHO has great advantages for telling the story and showing the problem structure. It often means that one needs to use the elegant formulations using subscripts and vector functions.
But what sense is there in having up to 8 dimensional arrays if at 8 Bit per element the model will soon break down as 32 Bit applications are limited to 2-3 GB of RAM? Such restrictions should be outdated by now and I would strongly pledge to have them removed in Vensim. I know that PowerSim will offer a 64Bit Version of its software by the end of 2011. So, have Vensim not fall back here.
Kind regards,
Guido
PS: The tools mentioned by Keenan (eg. a Vensim Profiler - how much memory will the equations need, where are equation bottlenecks... - and a Vensim Syntax Coloring Tool) would be on my personal wish list as well...
--
BSL MANAGEMENT SUPPORT
Blomenburg 1
24238 Selent
Germany
I would like to propose a much more mundane improvement: Go with the time and make Vensim a 64bit application. What has started out as a way to make complexity tractable on computers in the 60s and 70s (aggregated stock and flow simulation) tends to become a liability nowadays when one builds bigger models using subscripts. While Vensim already surpases the restrictions of say PowerSim Studio (6 subscripts and only 500.000 elements per variable, as far as I know) models in some areas can become very complex. For example one of the more recent issues of the SD review dealt with SD in the domain of transportation modeling which naturally introduces spatial complexity next to that defined in time.
If SD is to survive in business consulting -- it seems to still remain a mostly academic endeavor -- it should not flinch from detail rich problems due to memory limitations. We all love SD for the clarity it brings with small to medium scale models sometimes refered to as 'insight models'. The day to day problems of managers go beyond this. They want to have a tool that will also adress the more intricate problems with lots of detail.
While agent-based modelling has advantages due to the way data are handled (cf. Keenan et al., Modeling General Motors and the North American Automobile Market, in: Proceedigs of the SD conference, 2004
http://www.systemdynamics.org/conferenc ... 4KEENA.pdf
-- the paper is right on target as regarding the limitations of SD and of Vensim). SD still IMHO has great advantages for telling the story and showing the problem structure. It often means that one needs to use the elegant formulations using subscripts and vector functions.
But what sense is there in having up to 8 dimensional arrays if at 8 Bit per element the model will soon break down as 32 Bit applications are limited to 2-3 GB of RAM? Such restrictions should be outdated by now and I would strongly pledge to have them removed in Vensim. I know that PowerSim will offer a 64Bit Version of its software by the end of 2011. So, have Vensim not fall back here.
Kind regards,
Guido
PS: The tools mentioned by Keenan (eg. a Vensim Profiler - how much memory will the equations need, where are equation bottlenecks... - and a Vensim Syntax Coloring Tool) would be on my personal wish list as well...
--
BSL MANAGEMENT SUPPORT
Blomenburg 1
24238 Selent
Germany
Re: What improvements would you like to see in Vensim?
After having started a fruitful discussion about 'dealing with bigger models' I was recently guided to Nathaniel Osgood's papers dealing with the comparison between ABM and SD. In this paper he dares to give some outlook for the future of both modeling worlds and notes, that both worlds will probably approach each other (cf. AnyLogic):
http://www.systemdynamics.org/conferenc ... GOO488.pdf
I would really like to see some of his suggestions regarding the further development of the 'mature' tools for SD like Vensim implemented:
Guido
http://www.systemdynamics.org/conferenc ... GOO488.pdf
I would really like to see some of his suggestions regarding the further development of the 'mature' tools for SD like Vensim implemented:
- dynamical subscripting ranges giving flexibility in modeling
- maybe some support for ABM-like-modeling in Vensim (hybrid modeling)
Guido
-
- Junior Member
- Posts: 13
- Joined: Tue Sep 11, 2007 4:23 am
Re: What improvements would you like to see in Vensim?
At my company, we have a problem presenting models. Our models have a lot of "spaghetti" and variable names that don't mean much to our clients. We use layering a lot, but this still doesn't allow us to hide variable names in an important loop that are irrelevant to our clients. It would be nice to be able to create a view that shows arrows in a causal chain that skip intermediate structure, including stocks and flows even (maybe allowing us to show a delay sign in that link if we want). We'd like this view to be just another view of the simulatable model, and not a separate file. Not sure how to do this, but we would love to work with you on this idea. It needs more thought.