Saved (on disk) Opt Results Silently Destroyed

Questions about MultiCharts and user contributed studies.
User avatar
MC_Prog
Posts: 330
Joined: 28 Feb 2007
Has thanked: 64 times
Been thanked: 35 times

Saved (on disk) Opt Results Silently Destroyed

Postby MC_Prog » 06 Jan 2013

Re: MC 8.5 Beta2 -

Consider: a user runs an optimization and then saves the workspace to disk with the opt results.

Later, the user opens that workspace (opt results still there, good), then goes thru Format Signal ... -> Optimize -> Exhautive to reach the Exhaustive Search Properties. At this point the user examines the setup, perhaps takes a screenshot, etc., then clicks on Cancel, then Close.

No new optimization is performed (or intended to be performed). Nevertheless,the previous optimization results are silently destroyed. This is terrible, and should never happen.

Even worse -

Let's say our observant user luckily notices that his previously saved opt results have been silently destroyed. In an attempt to escape this unintended situation, the user now Closes the workspace (F4) without saving it. However, the copy on disk is altered and no longer has the opt results in it. They have been lost forever.

This is horrible, and defeats a primary purpose of saving opt results (and workspaces) in the first place.

I strongly believe a workspace saved on disk should NEVER be altered unless the user CHOOSES to save over it.

Please correct both the silent destruction of the saved opt results in the open workspace (when no new opt has been performed to write over them), and the silent alteration of the workspace file-on-disk when the user has not chosen to save the open workspace over them.

Thank you.

P.S. This has now been entered into PM as: MC-1209 - Saved Opt Results Silently Destroyed (on Disk too!)

User avatar
Andrew MultiCharts
Posts: 1587
Joined: 11 Oct 2011
Has thanked: 931 times
Been thanked: 559 times

Re: Saved (on disk) Opt Results Silently Destroyed

Postby Andrew MultiCharts » 07 Jan 2013

Hello MC_Prog,

Please go to File -> Preferences -> "Workspaces on Start" tab. Make a screenshot of the opened window and attach here, please.

User avatar
MC_Prog
Posts: 330
Joined: 28 Feb 2007
Has thanked: 64 times
Been thanked: 35 times

Re: Saved (on disk) Opt Results Silently Destroyed

Postby MC_Prog » 07 Jan 2013

Hi.

I've attached the picture, and I think it does represent a situation that I or any user needs to be aware of. In this case my MC 8.5 Beta2 settings were configured to save the workspaces on exit without raising a dialog about it. This is unlike my production MC setup. I will change that, and I appreciate having my attention called to it.

Image

However, the above is a different "silent overwrite" scenario than what I have reported. Please notice that in my reported steps I used F4 to explicitly close and dismiss the workspace prior to exiting the program. IOW, I explicitly closed and dismissed the workspace without electing to save it. (Had I elected to save it, that would have been via Ctrl-S, not via F4.)

Thus, for the problem scenario I am reporting, at the time of program closure, there is no workspace open and present for the system to silently save per the Preference choice.

It's a subtle but key distinction! If perhaps I'm still not understanding the situation correctly, then of course I want to rectify that, but if I am understanding it correctly, then I think a code fix will be required to prevent unasked-for destruction of the user's work (opt results).

Thanks.
Attachments
DoNotShow_AndSave_02.png
Config for Silent Save at MC Exit
(39.79 KiB) Downloaded 1951 times
Last edited by MC_Prog on 07 Jan 2013, edited 1 time in total.

User avatar
furytrader
Posts: 354
Joined: 30 Jul 2010
Location: Chicago, IL
Has thanked: 155 times
Been thanked: 217 times

Re: Saved (on disk) Opt Results Silently Destroyed

Postby furytrader » 07 Jan 2013

I have always thought that it would be nice to be able to save optimization results and re-load them at some point in the future.

User avatar
Andrew MultiCharts
Posts: 1587
Joined: 11 Oct 2011
Has thanked: 931 times
Been thanked: 559 times

Re: Saved (on disk) Opt Results Silently Destroyed

Postby Andrew MultiCharts » 08 Jan 2013

However, the above is a different "silent overwrite" scenario than what I have reported. Please notice that in my reported steps I used F4 to explicitly close and dismiss the workspace prior to exiting the program. IOW, I explicitly closed and dismissed the workspace without electing to save it. (Had I elected to save it, that would have been via Ctrl-S, not via F4.)
If it was set up as it is shown on your screenshot, every closed workspace is automatically saved (no matter, if the workspace was closed manually or when the software was shut down).
Consider: a user runs an optimization and then saves the workspace to disk with the opt results.

Later, the user opens that workspace (opt results still there, good), then goes thru Format Signal ... -> Optimize -> Exhautive to reach the Exhaustive Search Properties. At this point the user examines the setup, perhaps takes a screenshot, etc., then clicks on Cancel, then Close.

No new optimization is performed (or intended to be performed). Nevertheless,the previous optimization results are silently destroyed. This is terrible, and should never happen.
We will study this scenario and i will ket you know.

User avatar
MC_Prog
Posts: 330
Joined: 28 Feb 2007
Has thanked: 64 times
Been thanked: 35 times

Re: Saved (on disk) Opt Results Silently Destroyed

Postby MC_Prog » 08 Jan 2013

If it was set up as it is shown on your screenshot, every closed workspace is automatically saved (no matter, if the workspace was closed manually or when the software was shut down).
Thanks for your reply.

I'm not perfectly clear about the statement above. Do you mean that (with this setup) the workspace is saved in its current condition immediately at the time F4 is pressed? (If so, I would consider this to be incorrect behavior, as pressing F4 explicitly means DO NOT SAVE TO DISK).

Or do you mean that when I press F4 (again, with this setup) the workspace is dismissed without saving (as one would expect, since one did not do Ctrl-S) but is somehow "kept around" in memory and then later saved in it's current condition at program exit?

Or something else?

IMO, it is very necessary for saving/not-saving behavior to be both fail-safe designed by the engineers and exactly understood by the users. If either is not the case, the consequences in lost work are severe. My interest, both personally and on behalf of all the users, is to see that those negative consequences are avoided always - because they are in fact avoidable with the proper care and attention to detail.

User avatar
MC_Prog
Posts: 330
Joined: 28 Feb 2007
Has thanked: 64 times
Been thanked: 35 times

Re: Saved (on disk) Opt Results Silently Destroyed

Postby MC_Prog » 08 Jan 2013

Consider: a user runs an optimization and then saves the workspace to disk with the opt results.

Later, the user opens that workspace (opt results still there, good), then goes thru Format Signal ... -> Optimize -> Exhautive to reach the Exhaustive Search Properties. At this point the user examines the setup, perhaps takes a screenshot, etc., then clicks on Cancel, then Close.

No new optimization is performed (or intended to be performed). Nevertheless,the previous optimization results are silently destroyed. This is terrible, and should never happen.
We will study this scenario and i will ket you know.
Thanks!

User avatar
Dave Masalov
Posts: 1712
Joined: 16 Apr 2010
Has thanked: 51 times
Been thanked: 489 times

Re: Saved (on disk) Opt Results Silently Destroyed

Postby Dave Masalov » 09 Jan 2013

If it was set up as it is shown on your screenshot, every closed workspace is automatically saved (no matter, if the workspace was closed manually or when the software was shut down).
Thanks for your reply.

I'm not perfectly clear about the statement above. Do you mean that (with this setup) the workspace is saved in its current condition immediately at the time F4 is pressed? (If so, I would consider this to be incorrect behavior, as pressing F4 explicitly means DO NOT SAVE TO DISK).

Or do you mean that when I press F4 (again, with this setup) the workspace is dismissed without saving (as one would expect, since one did not do Ctrl-S) but is somehow "kept around" in memory and then later saved in it's current condition at program exit?

Or something else?

IMO, it is very necessary for saving/not-saving behavior to be both fail-safe designed by the engineers and exactly understood by the users. If either is not the case, the consequences in lost work are severe. My interest, both personally and on behalf of all the users, is to see that those negative consequences are avoided always - because they are in fact avoidable with the proper care and attention to detail.
With your setup workspace will be saved if you close it (press F4) or close Multicharts. You can select the "Do not save workspaces" option or uncheck the "Do not show "Sawe Workspace" dialog on Exit to choose to save or not to save every workspace manually.

As for the optimization issue that you have reported, it has been confirmed and added to our fix list. Thank you for reporting.

User avatar
MC_Prog
Posts: 330
Joined: 28 Feb 2007
Has thanked: 64 times
Been thanked: 35 times

Re: Saved (on disk) Opt Results Silently Destroyed

Postby MC_Prog » 09 Jan 2013

With your setup workspace will be saved if you close it (press F4) or close Multicharts. You can select the "Do not save workspaces" option or uncheck the "Do not show "Save Workspace" dialog on Exit to choose to save or not to save every workspace manually.
Dave - Thanks much for verifying the facts and getting back to me here.

While I understand from this discussion that we can configure the Prefs to prevent the worst case (and I have done so), I do want to say that IMO, logically and functionally, the acts of a) doing/not-doing the save, and b) displaying/not-displaying a dialog about it are (or should be) two separate things.

IOW, F4 and Ctrl-S (aka File -> Close Workspace, and File -> Save Workspace) should not behave the same, ever. The former should close the workspace without saving it and the latter should close and save the workspace in all cases.

So, I accept that the current facts have been established, and I hope that the logic of what I'm saying here can be accepted and implemented as an improvement for the future.

Again, thanks!

P.S. Issue reported into PM: MC-1214 - File -> Close Workspace sometimes saves the workspace (but should not)
As for the optimization issue that you have reported, it has been confirmed and added to our fix list. Thank you for reporting.
And thanks for this too!


Return to “MultiCharts”