open TELEMAC-MASCARET The mathematically superior suite of solvers

BOTTOM seems ok now.


NOTE: The attribute in "New Mesh" is of type "double". Therefore, the attribute in BOTTOM will be type "double". HOWEVER, the Selafin options say that it should NOT be saved to a file as type "double". Therefore, saving and loading BOTTOM may result in some loss of precision. In order to have BOTTOM be "float" then the source mesh's attribute must be type "float". Currently Edit > Add Attribute... only provides "double" and "integer". I will now add "float" before marching on

This is what I mean about the attribute changing type and a possible loss of precision. On the left are the properties of BOTTOM before the slf file is saved. On the right are the properties after the slf file is loaded. Notice that "double" has changed to "float" due to the options on the slf object that indicated that this should be done.

To delete 1 attribute (e.g. of type double) from a mesh:

1. Make sure that the mesh is selected in the WorkSpace
2. Make the current attribute the attribute that you want to delete
3. Edit > Delete Attribute

To add 1 attribute (e.g. of type float) to a mesh:

1. Make sure that the mesh is selected in the WorkSpace
2. Edit > Add Attribute...
3. Set the new attribute's definition as you want; for example, change the selection combo box to "float"

This should solve all of the issues that I'm currently aware of. The next Blue Kenue will have these fixes.

Happy Kenueing!

Kindly,
Alan

Hi Alan,

I tried this again on a different machine and am getting this behaviour now. Every time I go back to the Data tab, it wipes out the interpolated bathymetry. This is not even on the SELAFIN BOTTOM, just on a regular mesh.
To be honest, I haven't followed 100% of your explanations, but it sounds like this is something linked to statistics and double/float that will be fixed?

Thanks,
André Renault

No need to follow all of my explanations! Just try the next version of Blue Kenue after I prepare it and then let me know if it works or not...

Kindly,
Alan

Hi Alan,

Thank you so much for your time spent on debugging this! I'm looking forward to the next release, as this is holding back my meshing somewhat (I'm a student) - not to pressure you, but do you have an idea of when this will be released?

In the meantime, I'm tempted to go back to 3.3.4, use Shift in X/Y to circumvent the truncating issue, and do bathymetry from there. I really don't want to do this because it could be problematic in the future, but it might be the best stopgap for me?

Thanks again!
André Renault

I've also improved the New Attribute dialog box a little; may help you (and others) more clearly see what's going on...



As for when is the next Blue Kenue ready... I need to complete some work on Green Kenue for a client and then release both Green Kenue and Blue Kenue at that time. I'm hoping for late this week or early next week.

BK 3.3.4 is the released version and 3.12.6 is an alpha version.
- Have a look on CHyMS for what it means to be an alpha version if you're not quite clear...

If you're not sure which version to use and you are under time pressure and can't wait for possible fixes in the alpha version, then I'd definitely suggest using BK 3.3.4 and trying out the alpha version if/when you have time to do so and can provide feedback about problems encountered and can wait for their fixes. My 2 cents...

Good luck with your work and thank you for taking the time to share your experiences.

Kindly,
Alan

BK 3.12.7-alpha is now ready for you to try out (if you want)

Kindly,
Alan

Hi Alan,

Thanks so much for the update! I tried out 3.12.7 and found that I had to save the SELAFIN with double precision in order to avoid XY-value truncation (curiously, to the nearest 0.5 m) in the mesh. However, it seems to work in double and the Z-values are not overwritten! Thank you thank you thank you! I can now move forward with my project.

Curiously, every time I open the SELAFIN object and show the BOTTOM variable in a 2D plot, the colour scale defaults to only one colour, red, and this for the minimum elevation value. I have to click Reset and Apply every time I want to show it. Not a showstopper, simply a curious behaviour that may want some looking at.

Incidentally, I ran into another bug in all of this. Whenever I create a 2D Interpolator with an .xyz file without a header (I haven't tested it with one, because of the aforementioned header bugs), and with no ConstraintOutline, if I save and reopen the 2D Interpolator, it gives me the following error:The dialog box only has an OK button, and then the 2D Interpolator is not displayed.

Again, I can happily open tickets for these issues in 3.12.7, as they are far less critical than the broken SELAFIN generation that you have magically fixed! Please let me know if I should indeed do so.

Kind regards,
André

I'm on vacation at the moment so can't create an account on CHyMS for you. Would be easier to keep track of things that way... meh

When I:

1. Load BOTTOM.t3s (dragging to a 2D View shows the correct colours)
2. File > New > SELAFIN Object...
3. Check "Save as Double Precision"
4. Drag BOTTOM.t3s to "new Selafin"
5. Click on "new Selafin"
6. File > Save Copy As
7. BOTTOM2.slf (Big Endian)
8. Load BOTTOM2.slf (I see Min: 280.26 and Max: 330.5 in the dialog box)
9. Drag BOTTOM2's BOTTOM mesh to a 2D View
10. The colour scale is correct

Looks like I'm doing something different that you're doing

I'll have a look at the second issue you mentioned...

Kindly,
Alan