LabVIEW Idea Exchange

Community Browser
cancel
Showing results for 
Search instead for 
Did you mean: 
Post an idea

Many or most VIs that ship with LabVIEW have their protection set to Unlocked (no password). The screenshot below shows a selection of such VIs.

1.png

 

It would be much better if the protection of vi.lib VIs was set to "Locked (no password)", to prevent accidental modification.

2.png

 

It seems very risky for built-in VIs to be open to modification, especially to accidental modification.

 

Scenario 1: Developer A is developing an application on their machine which contains modified vi.lib VIs which were accidentally modified as part of work on previous projects. They build an application which passes validation and starts to be used in production. All of the developer's source code is committed to a source code repository. Developer A leaves the company. Six months later Developer B is asked to pull the code from the repository and add a minor improvement. The application behaves very differently when Developer B builds it on their machine. A long and complicated troubleshooting session later, Developer B concludes that the different behaviour was likely caused by modified vi.lib VIs on Developer A's machine. Developer B cannot be sure, because Developer A's machine was wiped when they left, so there is no way to unequivocally prove the conclusion.

 

Scenario 2: A team of developers builds a test system for a defence application. The code is completed, and the test system is put through a thorough  commissioning and validation process that involves testing dozens of known good units and known bad units. The validation process takes three weeks to complete. Management plans to not have to run the whole validation process for future minor changes. Instead they will ask the development team to perform code reviews and record notes for each minor change. Revalidation is not necessary if the code reviewers agree that the changes are non-functional, for example, the wording was changed in a dialogue message, or a logo was added to the UI. This sounds like a great plan, but is technically unsafe. Strictly speaking the whole revalidation process would have to be rerun, even for minor changes, due to the fact that not all of the source code is visible in the repository (there is uncontrolled source code in vi.lib that could have been modified in between builds).

Essentially I don't think it's safe for an app to contain source code that is not visible or tracked in a repository.

I can't think of a simple, quick solution to the concerns above, but having all vi.lib VIs set to "Locked (no password)" could be a quick first step towards reducing the likelihood of this issue. Developers would at least have to consciously edit vi.lib VIs, rather than doing it accidentally which can happen now. Of course, a malicious actor could still wreak havoc by editing a few inconspicuous vi.lib VIs.

The risk would be reduced further if the vi.lib VIs were password-protected. This would come at the expense of not being able to view the source code of native VIs, something which I find useful. Therefore, I personally would prefer "Locked (no password)" to password-protected, but I might prefer password-protected to unlocked.

Similar concerns apply to non-NI third-party libraries that install in vi.lib, user.lib or instr.lib, for example the extremely useful OpenG libraries. These too are examples of uncontrolled source code. For this reason some developers I worked with preferred to copy the OpenG and other libraries into the project repository (this involves a tedious job of opening each library VI and relinking it to the other library VIs in their new location).


This idea is similar but potentially easier to implement than the following idea: Make the VI's from the "vi.lib" Read-only - NI Community

 

Thanks

We've all seen the demo that shogg did where he set the color of his splitter bars to the panel color so they disappear at run time, but what if the splitter goes over background controls that you want to persist between panels?

 

22583iA7F722F2038D69E4

 

It look slike the smallest i can have my splitter bars is 2 pixels.  I totally want them to show up in edit mode, but I'd like them to be 0 pixels wide in run mode (selectable, of course).

Unlike the scales of numeric controls, graph scales don't support text labels (wouldn't that be cool! :smileywink: ) *(see footnote)

 

It could be handled very similar to the way text labels are handled for the scales of numeric controls, so most of the code is already there.

 

This would come in very handy for e.g. histograms or bar graphs, where each bar needs a text label, or for cases where we have arbitrary units.

 

Examples for integer scales: 

  • "January", "February", ...
  • "LabVIEW users", "CVI Users" ...
  • "Europe", "Asia", ...

 

Examples for floating point scales (x, or y):

  • "Too cold", "cold", "warm", "hot", "too hot"...
  • "small", "medium", "large", ...
  • "min", "max"...
  • "high frequrency", "low frequency"...

 

*My quote from this old discussion . See also Ben's example further down.

It is time to put a dent in the floating point "problems" encountered by many in LV.  Due to the (not so?) well-known limitations of floating point representations, comparisons can often lead to surprising results.  I propose a new configuration for the comparison functions when floats are involved, call it "Compare Floats" or otherwise.  When selected, I suggest that Equals? becomes "Almost Equal?" and the icon changes to the approximately equal sign.  EqualToZero could be AlmostEqualToZero, again with appropriate icon changes.  GreaterThanorAlmostEqual, etc.

 

AlmostEqual.png

 

 

I do not think these need to be new functions on the palette, just a configuration option (Comparison Mode).  They should expose a couple of terminals for options so we can control what close means (# of sig figs, # digits, absolute difference, etc.) with reasonable defaults so most cases we do not have to worry about it.  We get all of the ease and polymorphism that comes with the built-in functions.

 

There are many ways to do this, I won't be so bold as to specify which way to go.  I am confident that any reasonable method would be a vast improvement over the current method which is hope that you are never bitten by Equals?.

I often use LabVIEW projects.  Projects may contain libraries, classes, etc.

 

Changing the contents of any component (such as a library) within a project causes the project file to change -- even though the project definition itself shouldn't have actually changed. For instance, if this is my project:

_carl_0-1656622304965.png

and I delete a VI that is inside an .lvlib, the project file itself claims to have had an attribute change:

_carl_2-1656622426038.png

_carl_3-1656622515498.png

In theory, this top-level project change isn't necessary -- since the project should just reference the library, and only the library should reflect a change.

 

These unnecessary file changes make it trickier to deal with source control, merging code, tracking changes, etc, along with adding one more file that you get prompted to save every time you close a project.

Even though I use the tools selection in the automatic mode, many times I need some specific tool not accessible in the automatic mode or force one like "Position/Size/Select". Sometimes, mostly designing User Interfaces (UI) that takes all screen, we got the Tools Palette blocking access to something. Then we move the Tools Palette until it get in front of something else.

 

The idea is to have an option to have all buttons in the toolbar. Some programs allows you "dock" just dragging the over the tool bar.

 

Many times we have plenty of space to have all buttons there, if not, we can have a second row of buttons. That will be nice to add more buttons as we wish.

 

tools palette buttons in the toolbar.png

 

PS: I added a swap button between the foreground and background colors. The idea is explained in this thread Swap Colors in Tools Palette.

The request:

The thickness of splitter bars can be resized through the front panel at edit time, but this property is not exposed through property nodes. Why not expose this?

 

Background:

I use splitter bars all the time for resizable GUIs. They're also cumbersome. Many of the feature improvements I'd love to see have already been captured on this forum. In the mean time, I've got tools for automating what I can -- but I've hit a dead end when it comes to resizing the splitters programmatically. This requested property node doesn't exist and pre-sized splitters can't be copied from template VIs using scripting because the "Move" method throws an error.

String containers have display format Glyphs to help us understand the String Values when the display is Normal, Hex or Escape codes.

 

Integer numeric type containers have them too to show radix.

 

Timestamp containers must be inspected for their display format property to determine if they are in System Time or UTC.   

 

Timestamp Controls, Indicators and Contsants SHOULD have a similar optional display format glyph.

 

This could be used instead of the misleading coercion dots when operations are performed on Timestamps as Seen Here 

So the request here is to have an option wherein the array indicators (or controls) can be linked with another array indicator (or control) and can be scrolled synchronously.

 

There could be a property node for that (or invoke node or may be some other way), which will accept the reference of another array and thats all.

 

Sync Scroll

In above example, Array 1 is linked with Array 2, so whenever the user will scroll (or change the index of Array 2, the same change will reflect in Array 1.

When I have an array of clusters and I want to locate the array index where a specific element of the cluster has a certain value, I need to first build an array from the Array of Clusters and then search that to find the Array element I want:

 code example.jpg

 

It would be nice (and cleaner and likely faster) if I could wire the Array of Clusters into an unbundle by name function and select String[] to get the array of string element to search.

In addition, if the cluster contained nested clusters, I could access them the same way, using the dot notation alrerady supported.  For example, the unbundle would let me select 'cluster1.subcluster2.String[]' to access the subarray of an element.

 

code example2.jpg 

 

The venerable Array Size function should just resize itself for the number of dimensions on the input array and have each dimension's size emitted with a separate scalar output.

The Find Items With No Callers dialog box in Project could use a Remove From Project feature. Currenly, we have to use Go To, then remove the VI, then generate a new  Find Items With No Callers list.

 

remove_from_project.png

(This is different and less controversional than  this related old idea)

 

Arrays of timestamps only contain legal values and can even be sorted. We can use "search 1D array" to find a matching timestamp just fine.

 

As with DBLs, there might be a very close array value that is a few LSBs off, but well within the error of the experiment so it is sufficient to find the closest value. We can always test later if that value is close enough for what we need or if "not found" would be a better result.

 

If we have a sorted input array, the easiest solution is "threshold 1D array" giving us a fractional index of a linearly interpolated value. For some unknown reason, timestamps are not allowed for the inputs, limiting the usefulness of the tool. One possible workaround is to convert to DBLs, with the disadvantage that quite a few bits are dropped (timestamp: 128 bits, DBL: 64bits).

 

Similarly, "Interpolate 1D array" also does not allow timestamp inputs. In both cases, there is an obvious and unique result that is IMHO in no way confusing or controversial.

 

 

 

IDEA Summary: "Threshold 1D Array" and "Interpolate 1D Array" should work with timestamps.

 

 

 

One of the fiddly things I seem to do more than I'd like is adjust the bottom of block diagram comments to the right height. At a minimum there, but also for similar text boxes on the front panel or subdiagram labels, etc. I'd like to have a snap feature that sizes to a multiple of the text height. Examples:

 

With the three nodes aligned to Top, we get that annoying bend in the error wires. The spacing between Refnum and Error wires is one pixel greater on the Property Node than on To More Specific Class.

 

Align TMSC Terminals.png

 

Thoughts?

Have VI in the project windows behave like SubVI so when you CTRL + Double Click on one the SubVI FP and BD opens (and the BD is frontmost).

The last decade has seen a fantastic development in using neuronal networks to analyze data. Neuronal networks allow doing nowadays feasts that were unimaginable a decade ago at a tremendous speed. The latest release of LabVIEW allows using old TensorFlow versions <= 1.14. About four years ago, TensorFlow switched to TensorFlow 2.0, which resulted in a significant transformation in how models are organized. Models developed in TensorFlow <2.0 are incompatible with current standards. If National Instruments aims to maintain its position in the market of computer vision, developing tools to use current TensorFlow versions is a MUST.

 

                             

                              lvproj : be able to rename a folder

 

SR2.png

 

             SR3.png

Since LabVIEW 2017, the default build specification check the option "Allow future versions of the LabVIEW Runtime to run this application".

 

I don't undertstand why it is checked by default when you read the help that said "Disabling this option prevents any changes to the performance profiles and helps you avoid unexpected problems resulting from compiler upgrades."

 

Because NI and me doesn't know the future, NI will never garanty that your application will run with all the future LV Runtime. I know it well because I'm facing this issue with le LV RT 2020 that breaks the execution of different application that I have buid in 2017 SP1.

 

The behavior is a little strange too, because, even if you have the run-time use to build the application on your computer, the application will use the highest one present one the computer.

 

This options can be very usefull, and I appreciate it, but maybe we need to have more control on it. So my idea is to give to the user the ability to list supported or unsupported LabVIEW Runtime.

 

This will help us managing future LV Runtime without having to recompile the application which seems to be the goal of this option.

 

for Example :

 

[MyApplication]

ExcludeLVRTe = "2020,2021"

 

Regards

Problem: There are currently two ways of moving or transferring a VI or CTL from one lvlib or lvclass to another (or from having no owner to having an owner).

Method 1: Drag-and-Drop. Select the VI or CTL, drag to desired lvlib or lvclass, drop in desired location.

  • This works well in small to medium projects, and will probably remain the quickest and most common way of moving a item from one owner to another.
  • This does not work well in large projects, which can have dozens or hundreds of classes and libraries, sometimes deeply nested within complex virtual folder structures (for good reason). In this situation dragging the item can take a long time, especially when the destination lvlib or lvclass is "off the screen" and slowly scrolling through the project becomes necessary. This is done by hovering the mouse in just the right region in the top or bottom of the Project Explorer. This can be tedious and wastes time.

Method 2: Remove and Re-add. Select the VI or CTL, remove it from its current owner by right-click >> "Remove from Library" or by pressing Delete key, select destination lvlib or lvclass, right-click >> Add >> File... , then select the file on disk (this last step can take time in large projects with large folder structures).

  • This can be quicker than method 1 in large projects.
  • This method is technically a workaround. It does not directly transfer ownership of the item.

Note: Both methods usually now also require moving the file on disk to the new owner's folder location. This can be done using the Project Explorer Files view, but is an additional step that takes time. It's also possible to simply forget to move the file on disk (best practice is not encouraged). 

 

Solution:

  • Right-click the VI or CTL, select Move to... option
  • A window appears. This window may use a Combo Box or a Listbox to display a list of all the lvlibs and lvclasses available in the project.
  • The user selects the desired destination owner and presses OK.
  • The item is moved to the destination library.

1 (edited).png

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Nice-to-haves

  • The Move to... window may contain buttons or other mechanisms (e.g. clickable datagrid headers) that enable the user to sort the libraries and classes in the project by name in alphabetical or inverse alphabetical order, and to filter: display only lvlibs, display only classes.
  • When the selected VI or CTL is initially part of a lvlib or lvclass, that owner will be omitted from the list displayed in the Move to... window.
  • The Move to... window could contain a tickbox named "Also move on disk to new owner's folder?" or similar. When this is ticked, the item is also moved on disk to the folder of the new owner. This would remove the need to use the Files view to move the item on disk (would save time) and would help encourage a best practice (member items should be located in the same folder as their owner).
  • The Move to... option should be available when multiple items are selected. The items may initially be owned by the same owner, or may be owned by different initial owners.

This idea is inspired by the "Project Explorer "Move to Owner Folder" option in Files view" idea.