Opened 8 years ago

Closed 8 years ago

#328 closed defect (fixed)

vHW combined distribution plot legend

Reported by: demeler Owned by: gegorbet
Priority: normal Milestone: future
Component: ultrascan3 Version:
Keywords: review Cc:

Description

on the vHW combined distribution plot legend the entries are not what I would expect. The legend should show the triple description (solution/cell contents). Attached is an example of run 205603 from Julian Klosowiak (Northwestern University) where the descriptions should be: Red: MiroL + Ca Solution, green: MiroL Solution, blue: MiroS Solution

Attachments (4)

vhw.png (94.8 KB) - added by demeler 8 years ago.
vhw-2.png (49.9 KB) - added by demeler 8 years ago.
vhw-config-01.png (72.9 KB) - added by gegorbet 8 years ago.
us_plot config dialog
vhw-c-plot02.png (137.2 KB) - added by gegorbet 8 years ago.

Download all attachments as: .zip

Change History (20)

Changed 8 years ago by demeler

comment:1 Changed 8 years ago by gegorbet

  • Status changed from new to assigned

I have seen some really long description strings used. Using the description instead of the triple could lengthen the legend part considerably. Also, there is no requirement that the description be unique from triple to triple; so changing to description could lead to a legend where all colors have the same annotation. Wouldn't that confuse things?

The user chooses which triples to include in the plot by triple string. It seems to me that a legend that ties to the list to the left from which choices are made is more intuitive.

comment:2 Changed 8 years ago by demeler

I think the Beckman program allows 80 characters for a string. This is very reasonable, but I think we allow descriptions to be extended in US3 (as we did in US2). This has ALWAYS worked fine. I don't see a reason to change this. If the legend becomes large, so be it. As long as all the info is shown it will be useful to the user. What we have right now is not useful, especially not in a report. If you want to present a vHW combined plot by itself the legend doesn't make any sense. And why does it always say "DPoint" at the beginning? Where does that come from? I really need the description for each triple in the legend, nothing else, except of course the triple. I suggest: color dot - triple - description

comment:3 Changed 8 years ago by demeler

If the description for each cell is the same it could be intentional (repeat experiments). If they are and shouldn't be then it is the fault of the user and not our problem.

comment:4 Changed 8 years ago by gegorbet

  • Keywords review added

The vHW_enhanced and vHW_combined_plot programs have been modified to include the description string that goes with each triple. This shows up in the combined plot legend.

The desired behavior is in the vHW_combined program, but meaningful description strings will only be available from data files saved by the latest vHW_enhanced program.

This ticket is review-ready.

comment:5 follow-up: Changed 8 years ago by demeler

Yes, the description string now shows up - however, the DPoint/ELine still shows up - that is confusing to the user and should be removed.

comment:6 in reply to: ↑ 5 ; follow-up: Changed 8 years ago by gegorbet

Replying to demeler:

Yes, the description string now shows up - however, the DPoint/ELine still shows up - that is confusing to the user and should be removed.

I do not understand why this would be confusing. The legend is an explanation of plot elements. Some are Distribution Points (DPoint), some are Envelope Lines (ELine). This is extra information about each element, but I cannot see how it would lead to confusion.

comment:7 in reply to: ↑ 6 ; follow-up: Changed 8 years ago by demeler

Replying to gegorbet:

Replying to demeler:

Yes, the description string now shows up - however, the DPoint/ELine still shows up - that is confusing to the user and should be removed.

I do not understand why this would be confusing. The legend is an explanation of plot elements. Some are Distribution Points (DPoint), some are Envelope Lines (ELine). This is extra information about each element, but I cannot see how it would lead to confusion.

Because DPoint/ELine aren't real words. The information is already shown in the symbol clearly enough - the symbol shows if it is a point or a line, and which color it is - that's more than sufficient. We can save space in the legend by removing the superflous/confusing DPoint/ELine.

comment:8 in reply to: ↑ 7 ; follow-up: Changed 8 years ago by gegorbet

Replying to demeler:

Replying to gegorbet:

Replying to demeler:

Yes, the description string now shows up - however, the DPoint/ELine still shows up - that is confusing to the user and should be removed.

I do not understand why this would be confusing. The legend is an explanation of plot elements. Some are Distribution Points (DPoint), some are Envelope Lines (ELine). This is extra information about each element, but I cannot see how it would lead to confusion.

Because DPoint/ELine aren't real words. The information is already shown in the symbol clearly enough - the symbol shows if it is a point or a line, and which color it is - that's more than sufficient. We can save space in the legend by removing the superflous/confusing DPoint/ELine.

DPoint/ELine are not superfluous because they provide a way to distinguish two different plot elements that otherwise would have identical tags. This comes into play, for example, when choosing plot elements in the plot Config dialog. If real words are needed, how about "(Point)" and "(Line)"?

comment:9 in reply to: ↑ 8 ; follow-up: Changed 8 years ago by demeler

Replying to gegorbet:

DPoint/ELine are not superfluous because they provide a way to distinguish two different plot elements that otherwise would have identical tags. This comes into play, for example, when choosing plot elements in the plot Config dialog. If real words are needed, how about "(Point)" and "(Line)"?

Take a look at the attached picture - there is a very clear distinction in the symbol - one is a line, one is a point (ellipse). There is no need to spell it out - the symbol shown next to text is sufficient to distinguish the plot elements. I would like to get rid of the point/line text element.

Changed 8 years ago by demeler

Changed 8 years ago by gegorbet

us_plot config dialog

comment:10 in reply to: ↑ 9 ; follow-up: Changed 8 years ago by gegorbet

Replying to demeler:

Replying to gegorbet:

DPoint/ELine are not superfluous because they provide a way to distinguish two different plot elements that otherwise would have identical tags. This comes into play, for example, when choosing plot elements in the plot Config dialog. If real words are needed, how about "(Point)" and "(Line)"?

Take a look at the attached picture - there is a very clear distinction in the symbol - one is a line, one is a point (ellipse). There is no need to spell it out - the symbol shown next to text is sufficient to distinguish the plot elements. I would like to get rid of the point/line text element.

Although the addition of a "(Point)" prefix is not strictly necessary in the plot legend, it is needed for the plot Config dialog, Look at the new attachment. Tests show that with no distinguishing prefix, selection of line 3 actually presents line 1, the first plot element of three with identical tags.

comment:11 in reply to: ↑ 10 ; follow-up: Changed 8 years ago by demeler

Replying to gegorbet:

Replying to demeler:

Replying to gegorbet:

DPoint/ELine are not superfluous because they provide a way to distinguish two different plot elements that otherwise would have identical tags. This comes into play, for example, when choosing plot elements in the plot Config dialog. If real words are needed, how about "(Point)" and "(Line)"?

Take a look at the attached picture - there is a very clear distinction in the symbol - one is a line, one is a point (ellipse). There is no need to spell it out - the symbol shown next to text is sufficient to distinguish the plot elements. I would like to get rid of the point/line text element.

Although the addition of a "(Point)" prefix is not strictly necessary in the plot legend, it is needed for the plot Config dialog, Look at the new attachment. Tests show that with no distinguishing prefix, selection of line 3 actually presents line 1, the first plot element of three with identical tags.

What you show in the attachment looks good to me - just in the plot legend I find it superfluous. Is it possible to hide the info in the plot legend and just show it in the config list?

comment:12 in reply to: ↑ 11 ; follow-up: Changed 8 years ago by gegorbet

Replying to demeler:

Replying to gegorbet:

Replying to demeler:

Replying to gegorbet:

DPoint/ELine are not superfluous because they provide a way to distinguish two different plot elements that otherwise would have identical tags. This comes into play, for example, when choosing plot elements in the plot Config dialog. If real words are needed, how about "(Point)" and "(Line)"?

Take a look at the attached picture - there is a very clear distinction in the symbol - one is a line, one is a point (ellipse). There is no need to spell it out - the symbol shown next to text is sufficient to distinguish the plot elements. I would like to get rid of the point/line text element.

Although the addition of a "(Point)" prefix is not strictly necessary in the plot legend, it is needed for the plot Config dialog, Look at the new attachment. Tests show that with no distinguishing prefix, selection of line 3 actually presents line 1, the first plot element of three with identical tags.

What you show in the attachment looks good to me - just in the plot legend I find it superfluous. Is it possible to hide the info in the plot legend and just show it in the config list?

It might be possible to do what you suggest, but it would require a lot of digging into us_plot and QwtPlotCurve code. Right now, the ID string is set once in the us_curve constructor and then automatically appears in the legend and config list. It seems like it would take a lot of effort to avoid what may be somewhat superfluous, but is not harmful or confusing, especially if we change prefixes to "(Point)" and "(Line)".

comment:13 in reply to: ↑ 12 Changed 8 years ago by demeler

It might be possible to do what you suggest, but it would require a lot of digging into us_plot and QwtPlotCurve code. Right now, the ID string is set once in the us_curve constructor and then automatically appears in the legend and config list. It seems like it would take a lot of effort to avoid what may be somewhat superfluous, but is not harmful or confusing, especially if we change prefixes to "(Point)" and "(Line)".

In that case, let append the words (diff.) and (int.) for differential and integral distribution, which is what they really are, at the end of the comment string. That is at least somewhat informative.

Changed 8 years ago by gegorbet

comment:14 follow-up: Changed 8 years ago by gegorbet

Are "(diff.)" and "(integ.)" suffixes as shown in the latest attached screenshot acceptable?

comment:15 in reply to: ↑ 14 Changed 8 years ago by demeler

Replying to gegorbet:

Are "(diff.)" and "(integ.)" suffixes as shown in the latest attached screenshot acceptable?

Looks good! I am going to close this ticket, please commit changes.

comment:16 Changed 8 years ago by demeler

  • Resolution set to fixed
  • Status changed from assigned to closed
Note: See TracTickets for help on using tickets.